← Back to leaderboard
itsolelehmann

Ole Lehmann

Verifiedactive
@itsolelehmann

I cover AI + robotics. Helping business owners win in the AI era. eu/acc supporter, dad, techno-optimist

RosterAI Model News & ReleasesAI Agents & Productivity ToolsAI Content Creation & Video Gen
161,870 followers · 3,023 following · Last polled 3h ago
Perf. Score
65.7
Rank #99
Value Score
6.4
#259 of 273 priced
Median views
7.3K
p25 4.8K · trailing 7d
Followers
162K
+0.1% 7d
ER (impr.)
2.35%
eng ÷ impressions
ER (foll.)
0.27%
eng ÷ followers
Est. CPM
$303.11
QT-rate basis
WoW views
+29.7%
▲ rising

Rates & value · est. from 7.3K median organic views

overpriced
Quote tweetbasis
$2200
$303.11 / 1K views
Post
—
no CPM
Thread
—
no CPM
Retweet
—
no CPM
Value Score 6.4 · rank #259 of 273 priced$27329.19 / 1K engagementsrate age unknown (pre-audit import)

Trends

Follower growth
Median views / day
Engagement rate
View distribution · trailing-7d posts

Each blue dot is one organic post. The Performance Score ranks by the median (7.3K), not the mean (18K) — so a single viral post doesn't inflate a creator's reach. The p25 floor (4.8K) is the level 75% of posts clear.

Recent posts · ranked by views

PostPostedViewsLikesRepostsRepliesQuotesBookmarksER
Uh oh. Agents will likely break the entire subscription economy. Stanford research shows ~50% of all subscription revenue comes from people simply forgetting to cancel. But agents don't forget. Muse, Instinct, and Grok Bot can already go through your accounts, find every subscription you forgot about, and make you decide if you want to keep it. And the same research found people are 4x more likely to cancel when they're forced to actively decide. Once personal AI agents are on everyone's phone, a lot of subscription companies are going to really feel it.frozen23d ago305K7738375383370.43%
i'm obsessed with this idea of "vibe coding physical devices" right now even hardware noobs (like me!) can now finally turn our weird little ideas into gadgets that actually work you just describe the gadget you want, then schematik (basically cursor for hardware) picks the correct parts, writes the code to run them and walks you through building it. for example, here are 10 templates you can start with and make your own: 1. a plant that throws a disco a soil sensor warns you when it’s thirsty. water it and the LED ring celebrates with rainbow lights and a victory tune lol 2. a morning receipt printer turn it on and it prints today’s forecast onto a little paper briefing, with optional calendar events and blank lines for handwritten to-dos. 3. a digital hair clip a tiny wearable screen that displays your own messages. press a button to cycle through them. 4. a weather display that tells you what to wear checks the weather over wi-fi, then tells you whether to grab a jacket, umbrella, sunglasses, shorts or scarf. updates every 10 minutes. 5. a tiny desk companion tap its face to change expressions, tilt it to move its eyes, and swipe through the clock, weather and stock price. 6. a lamp you control with your hand swipe sideways to change color, up or down to change brightness. it saves your settings for the next time you power it on. 7. a pet treat dispenser press a button and it weighs the bowl before opening the gate. it blocks another serving if the bowl’s already full or the cooldown hasn’t finished. 8. a live flight radar a tiny round display that plots nearby aircraft and their callsigns, using live flight data fetched over wi-fi every 5 seconds. 9. a dedicated touchscreen chess game tap pieces to move and play against a small chess engine running on the device. the screen and processor are already on one board. 10. an air monitor for your room measures CO₂, temperature and humidity, then tells you when rising CO₂ means it’s time to air out the room.
frozen
22d ago
260K
2.0K
109
63
19
2.3K
1.75%
“okay, so what? it’s just math.” no, you don't get it. here’s what OpenAI’s math breakthroughs could eventually help make possible in the real world: Vlasov–Maxwell → stronger foundations for fusion research, whose ultimate prize is virtually limitless clean energy to power civilization Calderón’s Problem → portable body scanners using electrical signals, making medical imaging cheaper and easier to access Maximum-Cardinality Matching → faster searches for compatible kidney swaps across large donor pools, helping hospitals coordinate lifesaving transplants Inverse Elasticity Problem → scans that map tissue stiffness, helping doctors locate suspicious growths inside the body Mumford–Shah Conjecture → better tools for spotting tissue changes in brain scans, helping doctors identify signs of disease Bose–Einstein Condensation → stronger foundations for quantum sensors that could help vehicles navigate without GPS Simple Stochastic Games → better safety checks for self-driving cars and robots before dangerous mistakes reach the real world Matrix Multiplication → cheaper AI and bigger scientific simulations using the same computers Edit Distance → faster DNA comparisons, helping researchers study genetic changes linked to disease The k-Server Problem → robots that waste less movement and energy, making warehouses more efficient and goods cheaper to move a reminder that math is the foundation of science. so accelerating mathematical discovery could compress centuries of scientific progress into years. insane timeline to be alive for!frozen15d ago193K2.9K51198861.3K2.53%
been testing Opus 5.5 and i'm convinced it's the first model that's finally solved animation. it's basically an in-house storytelling studio you can run solo now here's 10 animation use-cases i think you'll start seeing everywhere: 1. podcast clips that turn a guest’s story into scenes people can watch 2. custom b-roll for youtube and talking-head videos that shows exactly what you’re explaining 3. ad creatives with different hooks and visual styles you can test without filming each version 4. product explainers that animate the problem your customer has and how your product solves it 5. looping website animations that show your product assembling, rotating, or working 6. animated shorts and mini-stories for reels, tiktok, and youtube shorts 7. branded intros, outros, and transitions you can reuse across your videos and podcasts 8. animated diagrams for pitch decks that show how customers, data, or money move through your business 9. customer success stories showing the before and after of using your product 10. animated charts and infographics that turn your research into shareable videosfrozen29d ago182K1.4K774952.2K2.01%
HOW TO GIVE YOUR GROK BOT ITS OWN PHONE NUMBER (so it can handle all the calls you hate making): this is perfect for all the stuff Grok Bot can’t do on a website, like chasing a refund, booking a table or anything else that needs a real phone call. it works by connecting your Grok Bot to Bland, the phone-calling bot xAI’s official marketplace uses. and your Grokbot can set up the whole thing by itself, so you never open a signup page. 1. just send your Grok Bot the Bland guide from the reply and ask it to set up an account for you. 2. you pick the area code you want for your bot's number. 3. you text back the 6-digit code Bland sends to your phone. 4. you tell it who to call and what you want, like "call my gym and cancel my membership." so the next time you're dreading a phone call, just send your bot to the call and it'll come back with the result.frozen6d ago155K1.8K103169213.5K3.58%
global crime rates are gonna collapse once superhumanoids like this hit the streets imagine facing this thing as a robber lol Unitree’s new "Guardian" robot can fight off a guy with a knife and patrol areas for 8 hours straight which is terrible news for criminals, because the biggest thing that stops them from committing crimes is the chance of getting caught right now in the US only about 13% of burglaries ever get solved so most burglars are making a pretty safe bet but if you put a robot on every block that: > never sleeps > records everything > and can physically stop/beat the shit out of you ...then almost no criminals would make that bet anymore on top of that, if you add the abundance that AI and robotics are likely to create then far fewer people will even get desperate enough to commit crimes in the first place, because they'll already have enough there's a good chance that our kids will think about robbery the way we think about guillotine executions https://t.co/7PuiOeLNjb1d ago141K90588355882661.21%
the idea of a “human interaction model” fascinates the hell out of me. the tech is finally good enough for it to feel believable for example, this one from Tavus can watch and listen while it talks, react to what’s happening, and interrupt you mid-sentence. all in real time. as eerie as that is, i can see it being extremely useful for practice: > job interviews where you get pushed on vague answers > podcast prep so you can explain your ideas without rambling > pitching your startup to a skeptical investor > sales calls with someone who keeps challenging your pricing > difficult conversations you’ve been putting off you can customize the conversation by telling it what you want to practice, what your goal is, and how you want it to behave. something like “i’m practicing a pitch. interrupt me if i start rambling, and tell me what i’m leaving out.” just try not to fall in love with your mock interviewer lolfrozen21d ago124K376781802260.56%
AI music will never replace human music Deep Blue beat Kasparov in 1997 and nobody gives a fuck literally nobody goes and watches a computer play chess, no matter how good it is meanwhile, people line up out the door to watch Magnus Carlsen play people want to see what a human can dofrozen22d ago107K2.2K154243371492.58%
Google's Argon cracked a fusion energy math problem in a couple of hours humans couldn't figure it out in 59 years we still don't grasp the acceleration that's about to hit usfrozen9d ago106K3.3K32475236874.18%
HOW TO CONNECT YOUR GROK @bot TO WHATSAPP send this tweet to your Grok Bot so it can handle the setup for you: 1. have it open web.whatsapp. com on its computer. 2. have it choose “link with phone number” and enter the number it already has for you. 3. have it generate the linking code and show it to you. 4. on your phone, open WhatsApp → Settings → Linked Devices → Link a Device → Link with phone number instead. enter the code to approve the connection. 5. have it check that your chats load and confirm it's connected. your Grok Bot can now read messages and draft or send replies on WhatsApp. once connected, you can have it: > organize your chats > flag messages you haven't replied to > add plans to your connected calendar > and remind you about things you said you'd do so between WhatsApp, my email, and X DMs, my Grok Bot now has an absurd amount of actionable context on my life :)frozen8d ago67K57423115121.2K2.85%
cool, now make it 450ft and put it on Alcatraz there are too many billionaires in the bay area for this not to happen https://t.co/VEshsu3gWtfrozen20d ago64K823323416721.53%
i agree. Grok Bot already feels like the agent product Steve Jobs would’ve cooked. for me, it’s surprisingly addictive and much more enjoyable to use than Codex or Claude Code (especially on mobile). my bot lineup so far: > an events bot that’s basically my own “Luma,” finding stuff to do wherever i am based on my taste and sending me the links > a health trainer that tracks my nutrition and workouts and reminds me when to train > a virality hunter that finds content ideas aligned with my brand for me to post about > a builder bot that finds new AI tools and workflows on X and helps me implement them in my business and personal life > an admin bot for the boring payments/administrative stuff i need to get done > a chief of staff that helps organize all of them it also somehow feels more agentic. stuff that Codex and Claude Code get stuck on, Grok Bot just seems to handle with much less friction. just a joy to usefrozen10d ago54K599275255462.29%
HOW TO GET CLAUDE CODE TO "VIBEHACK" ANY SMART DEVICE IN YOUR HOUSE (its actually much easier than i thought): 1. open Claude Code on a laptop that's on your home wifi 2. ask it to "scan my network, list every device with its make and model, and tell me all the possibilities of what you could fix, edit, or control on each one" 3. pick one and let it cook (it'll ask you if it needs you to press a button or plug something in) for example...one guy added AirPlay to their 15-year-old TV. another turned an old Kindle into a family dashboard with the day's schedule and incoming deliveries. another got their washer and dryer to announce out loud when the laundry's done. there's way more possible than you'd think, so let Claude show you what your own stuff can do. just run it on your own network, and read the commands before you approve them.2d ago53K770452741.2K3.92%
major vibe shift: Europe now has the best open AI model in the West 🇪🇺 "It is the best open weights model from US or Europe on aggregated benchmarks." i'm especially bullish on Mistral because they've essentially been anointed as Europe's AI champion: > €3 billion raised. the biggest tech equity round in European history > it's the last European-owned company still building top AI models (Germany's Aleph Alpha got bought by Canada's Cohere) > Jensen Huang told Macron that Mistral needed France's biggest companies rallying behind it. Macron's reply: "who are they? let me call them." and he personally called them > Macron went on national TV and told the French to download Mistral's chatbot instead of ChatGPT > the EU's own scaleup fund co-led their round, with Dutch chip giant ASML and the government of Luxembourg also writing checks > France + Germany have committed to bringing Mistral + SAP into their governments > the European Space Agency just signed on to build with Mistral > and on top of all this, the EU has mobilized €200B for European AI, including €20B for AI gigafactories Europe knows what's at stake: building and controlling our own powerful AI is an existential geopolitical priority and Mistral is our only real shot at competing with the Americans and the Chinese. so you have the most powerful people in Europe in their corner, with a *very* concrete reason to help them win. that backing is massively beneficial as it gets them priority access to money, compute and government customers. any startup would kill for this competitive advantage, and it's why i think Mistral will keep delivering. i'm very happy to see Europe shipping like this and i also love the commitment to the Le Chonk bit lol eu/acc 🇪🇺frozen16d ago52K8459756181342.19%
it's incredibly hard to stay bullish on SaaS seeing stuff like thisfrozen13d ago48K2227748760.81%
as a German I never really understood how much people were making in salaries working for Big Tech in the US until I saw it first hand for me it always seemed that you have to become an entrepreneur to make good money (or have daddy money) it really changed my worldview once I understood how many people were clearing multiple 6 or 7 figures a year working a job in Big tech (while tbh, often living a pretty comfy live)+ getting huge stock options from a risk to reward structure I do believe it's probably the safest way to become wealthy (high comp + stocks of a company that just keeps on growing) it was never an option in my mind because I just didn't know how much people could really make lol) I'm happy i chose the self-employed way, but if you're young and looking to make money working for BigTech or BigAI is def an elite way to get wealthy imo while minimizing your riskfrozen24d ago48K528134901161.48%
SF sounds like an exhausting place to have a personal life tbh imagine finally closing your laptop, going on a date, just to get evaluated like a cofounder i don't usually watch reality tv but i think people would be shocked at the status games and social weirdness inside tech their "startup brains" just completely take over their personal lives, it's actually fascinating to watchfrozen11d ago41K18562421070.79%
Muse and Grok bot feel impressive until you realize they are merely digital AI agents, limited to completing tasks for you on the internet just wait until you have a high-functioning physical AI agent that can complete tasks for you in the real world the game hasn't even begunfrozen24d ago38K40819270691.38%
i literally keep up with AI and robotics for a living and still regularly feel like i'm falling behind the curse of knowledge is extremely real on this app your X timeline is basically a group chat with the smartest, most obsessive tech people in the world you open it to catch up and end up comparing yourself to all of them. So of course you're gonna feel behind lol but then you zoom out and realize that a workflow you consider embarrassingly basic would completely change how someone else does their job this chart is actually such a breath of fresh air even after everything that's happened over the past few years, we're somehow still so earlyfrozen20d ago35K890725593123.79%
i'm German. and i'm sick of the "america builds, europe regulates" memes on here ok they were funny for a while (i posted a few myself), but the narrative is clearly shifting > Mistral just raised €3 billion at a €21 billion valuation. the biggest equity round a european tech company has ever done > Helsing (the defense AI company from Munich) raised $1.8 billion at an $18 billion valuation. and those were both in the last 3 months sure, there's red tape and some anti-tech sentiment in parts of european society but in my experience that just makes the people who are actually building bond even harder and i can feel it on the ground in Berlin i went to Superteam Germany's last event and it was one of my favorite founder event i've been to here. their Circuit location is worth it alone too (it has recently rebranded to that name from w3 hub) i loved it so much i even hosted my own marketing event here as part of their last hackathon now they're running a 2-day product bootcamp on Oct 1-2. you bring an idea, figure out who it's for and if the problem is real, get feedback from experienced founders great fit for solo founders, devs, hackathon teams or anyone sitting on an idea it's also built as prep for Colosseum's Crypto World's Fair with the following prizes: > $30k grand prize > $300k across the next 20 best projects > $2.5M in seed funding for winning teams > winners get considered for Colosseum's accelerator ($250k pre-seed + 12 weeks in SF) submissions are due Oct 12 (Colosseum says ~85% of venture-backed Solana startups came out of their competitions) if you're building at the intersection of AI and crypto this is a no-brainer IMO (links below) the european builder scene is alive. come see for yourselffrozen28d ago34K23118943491.17%
it's the best time ever to be a creative you can now ship everything you can imagine, all on your own so long I felt bottlenecked by my ability to create graphics and videos don't let the creative doomers get to you, this is the ULTIMATE golden time hyperfocus on finding ideas that move people and use these magic tools to distribute them throughout the universefrozen14d ago32K324162501921.73%
btw if you want to 10x the value of your Grok Bot subscription, check the Grok Bot marketplace. it's basically xAI's app store for ready-made Grok Bots (including some built by X engineers). these are my favorites in there right now: > Video Editor turns your raw talking-head footage into a finished vertical video with captions, B-roll and music > Haggle Bot goes through your software bills, finds the seats nobody uses, and drafts the email to haggle the price down > AI Search Visibility checks whether AI chatbots and Google recommend your business, and tells you who they name instead > Overheard watches Reddit, Hacker News and X for anyone talking about you or your brand, and sends you a short digest each weekday adding one is super easy: open the marketplace, hit “Add” on any bot you like, and approve the access it asks for. and if you’ve built Grok Bots for your own use-cases, you can publish them there too X will even pay you through a rewards pilot when people use them but even if you never add one, it's a good place to steal ideas from!frozen3d ago32K492274546913.98%
ok so Tesseract is the killer vibe editing plugin i've been waiting for you give your AI agent footage, describe the edit you want, and it handles the cuts, motion graphics and sound. the most impressive part for me is that you can give it reference videos with an editing style you want to emulate. like: “edit my footage in this style. match the pacing, transitions and animated text, using my brand colors.” the agent works directly with the editing engine, and everything stays in one editable project. so you can keep refining individual details as you go. > “bring that title in half a second earlier.” > “keep my voice playing while you cut from the talking head to the product demo.” > “move that sound effect so it lands exactly when the logo appears.” those tiny revisions are exactly what's been driving me insane recently i've grown to 32k followers on instagram over the past three months, and the amount of back and forth with my editor just to get everything right is nauseating getting the script, talking-head footage, timings, sound effects and on-screen text to all work together takes so much time. and good video editors are expensive. so tesseract saves you so much time and money for the quality you get. and whole thing is free/ runs locally on your mac.frozen1mo ago29K17892133001.75%
Email as a channel is cooked I’m rarely opening my emails now, it’s only my grokbot This will be the case for most people in 2-3 years Email as a direct line to your customers is taking a BIG hit imhofrozen4d ago28K217612815521.48%
as a german, this is embarrassing (but also oddly encouraging) > only 40% of founders surveyed who incorporated in germany would choose it again. in the US it's 85% > founders rate dealing with the german government 3.7/10. the US scores 7.6 but the encouraging part is when you look at the biggest problems european founders reported in the survey, they all seem so fixable: > notaries and registries > hiring and firing rules > slow responses > data and privacy rules > unclear or changing rules these are basically ALL bureaucratic issues drowning the massive talent and ambition we already have here. people *want* to invent things and build companies, but we're making them spend their energy fighting through rules and paperwork. we built this cage ourselves, and we have the power to open it so people can get back to building. i'd love germany to become the country founders would choose twice.frozen27d ago22K24726404431.63%
i thought opus 5.5 was already cracked then i went thru anthropic’s official 5.5 prompting guide and found a few easy ways to juice it even more: > give autonomous tasks an actual finish line. 5.5 can tell you what it’ll do next, then stop. define “done,” like all pages built and buttons tested. clear completion conditions help reduce those annoying stops where you have to type “continue.” > tell it to gather relevant context from its connected tools before taking action. 5.5 often jumps into a task fast without checking other sources that might matter. anthropic found that explicitly asking it to explore beyond what you mentioned improved task completion, even at maximum thinking effort. > leave effort on medium. it's the new default. anthropic found 5.5 on medium matched or beat opus 5 on high. 5.5 also thinks more at every level, so maxing it out mostly buys you slower, pricier replies. use high/max for genuinely tricky bugs, analyses, etc > replace “make it less AI-looking” with specific design changes. name unwanted colors, fonts, and button shapes. anthropic says vague requests mostly make 5.5 switch to another generic style. same principle goes for writing > delete "think step by step" and "think carefully before answering." 5.5 already decides how much to think on its own. anthropic found that cutting these lines made replies come back faster with zero drop in quality. > tell it to focus on your latest question during simple follow-ups. 5.5 sometimes rethinks earlier answers, making you wait unnecessarily. anthropic found this instruction sped replies up. keep rechecking enabled for complex analysis. > send dense charts at the highest resolution you have. in Claude Code or Cowork, tell it to crop and zoom in. anthropic found both helped chart reading more than extra thinking did.frozen24d ago20K317191704984.22%
Anthropic has no image or video model, yet Opus 5.5 just made me this 30 second animated film using nothing but 2,800 lines of code (yes, this is 100% Claude Code with no other AI tools, connectors, or reference images) i gave it one prompt: 4 seasons passing outside a train window, a cozy carriage, a cup of coffee on the table, Grand Budapest Hotel style and it literally came back with the finished .mp4 file ready in the chat so i asked it... how did you achieve this result when you have no image model?? here's what Claude actually does behind the scenes: 1. sets up free drawing software on your computer, the kind that turns written instructions into pixels 2. writes ~2,800 lines of code describing every object as shapes with exact coordinates: a tree is a brown trunk plus ~6 overlapping green circles, the coffee cup is a few ovals and curves 3. layers the scenery at different speeds, so telegraph poles whip past 140x faster than the mountains (which is what gives it depth) 4. renders a still of each season, looks at them, then fixes what looks off 5. animates it like a flipbook: it calculates where every object should be at each moment, then redraws the whole scene 900 times 6. builds the soundtrack the same way, as equations. a plucked string is a stack of sound waves that fade out. each rail click lands on the exact frame where the coffee ripples 7. stitches it all into the final video it even invented its own season transitions: a passing train sweeps spring into summer, tunnel turns autumn into winter, etc the creativity and attention to detail is getting pretty ridiculousfrozen28d ago19K264183151842.68%
had a really exhausting move from Berlin -> Mallorca if you do this with small child, every move you did before seems like a vacation lol also you want it to be smooth and chill for the child which adds an extra layer of complexity then you need to prepare your new place to not kill your child lol finally setup now, 10m from the beach for the next 2 month ready to LOCK INfrozen17d ago19K138128150.93%
Japan just started unleashing packs of Unitree robot Akita dogs to chase off bears > last year was Japan's worst bear year on record. 13 people killed! > northern Japan got so overwhelmed they even had to call in the military > so to combat the problem, they've dressed up Unitree Go2 robot dogs as Akitas (the breed Japan's old Matagi bear hunters took into the mountains) > the Akita robot dogs patrol the strip of land between town and forest on their own > AI cameras and infrared sensors in their chests pick up the locations of the bears > when one finds a bear it chases after it, barks, growls and blasts it with bear spray > then it radios the rest of the pack so they can gang up on the bear together lmao > they also ping the town so locals know to stay inside > this comes after a company in Hokkaido already sold 380+ robot wolves that do the same job on farms! this is the kind of job i'm very happy to see robots take overfrozen8d ago18K15926115471.38%
it will become more and more obvious that you need to own your full AI hardware stack and these look sick toofrozen17d ago18K1891491701.59%
Someone just “called me out” for being a hypocrite for simultaneously saying Europe has a cocaine epidemic but on the other hand offering cocaine to @marclou on X (LMAO) If you’re that stupid that don’t don’t understand that it was a joke referring to his post that’s just on you I’ve blocked this person a long time ago but friends send me her tweet because it’s going semi viral some people really have nothing to dofrozen24d ago17K51028070.51%
my new favorite thing to follow is Claude and ChatGPT cracking these unsolved historical mysteries. and they've been on a ridiculous run over the last 2 weeks: > a human + Claude Fable cracked a secret 1644 war plan sent to Holy Roman Emperor Ferdinand III. the letter described blocking Ottoman support for Transylvania, mobilizing Moldavia, and supplying 2,000 Cossacks: “I shall by every means prevent the Turk from giving him help… and give him two thousand Cossacks.” > the same project recovered a 10-page intelligence report written one week after the death of King François II of France. it deciphered 98.9% of its 1,715 words from ciphertext alone, revealing 2,000 Protestants taking up arms, rebel captains gathering around the king of Navarre, and a warning that they “would rather die with weapons in hand than be executed by a hangman” > it also read a 1916 telegram to Sun Yat-sen’s household that Japan’s Foreign Ministry couldn’t decipher. it tested all 57,600 possible code tables and recovered 41 of roughly 44 characters: “Chaozhou city has been declared independent by Mo Qingyu. Our force too has recovered Swatow.” > Opus 5.5 downloaded 5,000+ files from Samuel Hartlib’s archive, then sent subagents across Google Books and multilingual archives looking for unidentified sources. it may have found a manuscript Isaac Newton drew on for an alchemy experiment by matching the exact quantities and 2 disguised versions of “Hungarian vitriol.” GPT-6 Astra separately traced a Latin passage written by Newton back to a French alchemical text > OpenAI and Anthropic broke 2 Enigma messages 6 days apart. Astra chose a 1941 message that had resisted solution since 2005, investigated the archives, then wrote its own Enigma simulator and Bombe. it recovered: “Please specify the route of march. I am in Rosenow, Rosenow. Immediate reply by radio.” an Enigma expert verified the break. > 6 days later Claude Opus 5 cracked another: “Convoy back from Corps supply. Hartjenstein. Hartjenstein.” > Astra decoded a German WWI transmission that had remained unread for 108 years. it discovered that the keyword was “TRUPPENVERSCHIEBUNG,” then recovered: “An English cruiser arrived at Sevastopol on the ?4th. An Allied squadron follows on the 26th.” it verified the answer against HMS Canterbury’s original movement logs the cool thing is all of these documents were already sitting in digitized archives. so are millions of other manuscripts, letters, military records, and ciphers that no human historian could ever possibly read, search across, and connect in one lifetime. but frontier models can. and they are just beginning to work through them.frozen22d ago15K1331893891.66%
it’s so clear what Anthropic is about to do. i have zero doubt they’re gonna juice the shit out of their IPO with a *true* bombshell model within the next month (Fable 5.5? Opus 6??) they’re targeting a November IPO, so the incentive is pretty obvious: get everyone to rethink what AI can do right before they decide what your company is worth. my guess is Opus 5.5 was just the appetizer (insane appetizer btw) i think things are going to get really crazy going into the end of the year.frozen9d ago15K854120110.75%
it seems obvious that historical epics will become AI video's #1 use-case i recently saw the stills (below) for this upcoming movie called Cortés and it insta-hooked me: > oscar-nominated screenwriter nicholas kazan spent 30+ years trying to make Cortés: a 2-part epic about the Spanish conquest of the Aztec empire > but he could never get it made bc it was always “too big, too expensive, just always ‘too.’” (recreating lost cities and massive battles is giga expensive!!) > he wanted to bring a historically accurate 1500s tenochtitlan (where mexico city stands today) back to life > its an almost mythical city of towering temples and canals built on a lake, with enormous causeways connecting it to the mainland > then tell the story of the spanish arrival, the collision of two civilizations, and the massive battles that brought the aztec empire down > now it's finally in production because AI filmmaking tools are helping to overcome those cost and scale barriers > planned for 2027 > the concept art and stills below are fucking stunning it's not just Cortés either Vin diesel has also been trying to make his Hannibal epic for 20+ years (an army crossing the alps with war elephants to take on Rome). but one studio put the budget for it at $217M back in 2005. so imagine how many other filmmakers have spent decades sitting on their most ambitious projects because they were “too expensive.” those are the films i wanna binge watch.frozen21d ago14K23317107772.40%
a german AI model giving me hope? in this economy?? 🇩🇪 europe desperately needs its own competitive, sovereign AI. otherwise we risk: > our best AI researchers leaving to build companies abroad > missing out on the wealth and prosperity created by the AI boom > european businesses becoming dependent on foreign AI providers > giving washington and beijing leverage over the technology running our economy > becoming a digital colony (by operating on someone else's terms) and with Kolibri, we finally just got one! > it's cheap to run > sovereign (trained from scratch in Europe) > open weights you can download and run on your own hardware > and holds its own against comparable american and chinese open models on benchmarks. it's not frontier. but it's a solid model european businesses can run on their own infrastructure and build into their products and workflows. really hope this is the start of the tide turning in europe.frozen17d ago14K24913321372.39%
you can build your own home hardware studio for around $250. a little corner of your house where you use AI to design something, then 3D print it on demand. people are already making crazy stuff with this: > custom chess sets. one guy used ChatGPT to help design a medieval army playing against the undead, then printed the pieces and a magnetic board > replacement parts you can’t buy anymore. someone broke a discontinued latch on their truck, so they gave Claude photos and measurements, then printed a replacement that evening > storage containers, toys, roman busts, miniature cities, costume props, custom cookie cutters, etc. there’s just a ridiculous amount you can make with a little imagination and a roll of plastic here’s what the basic printing setup costs: > a beginner 3D printer like the Bambu Lab A1 mini costs $219 > a roll of plastic printing material is around $30 then here’s how you go from an idea to an actual object: 1. give Astra/Opus 5.5 your idea, along with any reference images, measurements, and details of whatever the object needs to fit 2. it generates the 3D design file (.stl) 3. open that file in your printer’s app, which turns it into printing instructions 4. hit print, and the printer makes the object real so you can make the exact thing you need, even when nobody sells it. i think these “home hardware studios” are about to become a major trend. a lot of people who got hooked on vibe coding apps are about to get equally hooked on vibe coding physical things.24d ago14K248162302333.80%
Figure's robot basically just did 13 marathons back to back with no food, no water and no sleep. > A full-time worker puts in about 2,000 hours a year. > A humanoid running 24/7 can do up to 8,760. ...so one robot is basically doing the hours of 4 people, and it never calls in sick ...and Elon is planning to print 10 million Optimus a year out of one factory in Texas. Truly mindboggling to digest the implications of this.3d ago13K676111160.76%
robots are advancing so fast that some hater will start sweating soon it's a obvious trend we are seeing more and more continuous robotic work demos on the timeline rn first hours (everyone still making fun) , soon days (hm, this is kinda working...), then forever (holy fak robots are building in space?!) robotbullmarketacceleration.exefrozen20d ago13K147671371.50%
pretty fucked you can get an Uber for pets but not with a seat for a small childfrozen23d ago13K208624001.87%
I'd rather be a dad making 3k/month than not be a dad while making 10million/monthfrozen6d ago13K2574385122.52%
btw self-play is the same trick that made AlphaGo better than any human at Go so now that it's finally working on robots with actual bodies, this is probably how we get superhuman robots > self-play means an AI practices by playing against older versions of itself > every time it gets better, its opponent gets tougher too, so it keeps improving without a human ever showing it what to do > but doing it with a body is way harder than Go > because instead of just picking one spot on a board, a humanoid has to set every joint angle 50 times a second > so the number of possible "moves" is basically infinite > they dropped a humanoid into a simulated soccer game and only rewarded it for scoring goals > after 140+ years of simulated games, it taught itself to dribble past defenders, shield the ball and tackle (totally on its own!) > then they moved what it learned onto a real robot, where it was actually good enough to play against real people > its still early, but i think give it a few years and we'll have Messi/Ronaldo level robots > but the bigger play is doing this for any real-world task > robots could practice for centuries in a simulation and show up better than any human who's ever done the work > then you point superhuman robots like that at physical infra like skyscrapers, power grids, factories, farms, etc > and this is how we get an age of abundancefrozen29d ago12K48691160.66%
this is the kind of acceleration my laundry pile desperately needs lmao the cool part is their agent remembers where it left off, switches between machines and reasons out what to do when something goes wrong they literally pulled its hand off the washer and messed with the settings mid-task but it went back and set the correct cycle anyway also one of the cuter humanoids i've seen. i think robots are going to get the same cute treatment as agents like dots and muse.frozen23d ago12K3952160.44%
the only thing you need to do to stand out on X is to not use AI to write still too hard for mostfrozen23d ago12K45116420.57%
long psychedelics long religion long family long raving long marathons long astrology long philosophyfrozen16d ago12K23512194412.63%
Meta’s Alexandr Wang says one session with an AI agent felt like 3 years of therapy. “It had, you know, access to my email and access to a bunch of my photos... I was very vulnerable through that process.” soon we'll all have our own "Superintelligent Therapists" and i can't stop thinking about the implications. think about how therapy works right now. you see someone for an hour a week, and everything they know about you comes from what you decide to tell them. so they only ever get the version of your life you remember (+ are willing to say out loud) now imagine an agent where: it has read every email and text you’ve sent. it has gone through your entire camera roll. it remembers every chat you’ve ever had with it. it would see all the deeper patterns you never even think to bring up like who you stopped replying to, what you keep putting off, who’s in your photos during the years you looked happiest, etc this thing would literally know you better than you know yourself it could psychoanalyze you at a level no therapist could ever reach (at least on the information side. i still think a good therapist does something beyond words, more on an "energetic" level) but what does it actually mean when there’s a superintelligence that knows you better than anyone else does? even Elon is worried about this one: “So many people are going to get one-shotted by super persuasive super intelligences and I might be one of them” anything that knows you better than you know yourself also knows exactly what to say to change your mind I'm cautiously optimistic it could be very helpful for self-reflection / behavior change, but there are massive downsides if stuff goes wrong. regardless it's inevitable, so we're all about to find out1d ago12K9010286551.62%
man I really miss the days of real people in the comments :(frozen9d ago12K136085031.94%
just insane. you can use AI to crack unsolved historical mysteries as a hobby now. here’s how to find a mystery and try it yourself: you don’t need any experience in historical research or codebreaking, and there are actual lists with hundreds of real mysteries to choose from. for example, carter church had never broken a cipher before. he gave GPT-6 Astra one image of an encrypted letter to one of napoleon’s generals and a goal: decipher it. Astra spent about 6 hours transcribing 1,300 signs, finding a partial key, and writing code to crack the rest. the result: a readable 217-year-old war briefing, with the solution confirmed by the historical cipher archive’s maintainer. here’s how to try it yourself: 1. pick a real mystery. start with one of these lists (links in the reply): > Cryptiana’s unsolved historical ciphers > Daniel Bourdeau’s unsolved catalogue > DECODE database > Klaus Schmeh’s encrypted book list choose one with a readable scan, and check the latest notes to make sure nobody's solved it yet 2. give the AI the evidence and the goal. the only 2 models i’d use for this are GPT-6 Astra in Codex or Opus 5.5 in Claude Code. upload a clear image of the encrypted document, along with the source link. then ask it to decipher the message. tell it to research existing clues and previous attempts, then write and run code to test possible solutions. 3. make it show its work. if it finds a solution, ask for the decoding key, code that reproduces the reading, and anything still uncertain. then you have something concrete for the archive maintainer or a specialist to check. 4. email the list’s maintainer. for Cryptiana, that’s Satoshi Tomokiyo: verlat@hotmail.com. send a short summary with the original document link, your proposed reading, and the decoding key and code from step 3. mention that you used AI, flag anything still uncertain, and ask whether they'd be willing to review the solution. if they confirm your solution, congrats. you just cracked a historical mystery! it sounds almost too good to be true, but some of these ciphers are actually relatively straightforward for a specialist. there are just a lot of them scattered across digitized archives, and specialists only have so much bandwidth. but now that the frontier models are good enough, a curious hobbyist can pick one, give an AI agent the evidence, and put it to work. good luck, and go dig up some centuries-old secrets :)frozen21d ago11K44580450.90%
Just when I felt comfy and at home with my grokbots anthropic drops this bomb I’m tired of switching, boss But opus mogs, so I guess it’s time to give it a spin1d ago11K943252171.26%
in light of the AI cyberattack mayhem going on right now, it's a good time to reshare Karapthy's Digital Hygiene guide probably the most important thing you can read this week: > use a password manager for every account > set up physical security keys so attackers can't log in > enable face id and fingerprint everywhere > randomize your security question answers > encrypt your hard drive > get rid of unnecessary smart home devices > switch to signal for private messaging > use brave instead of chrome > switch to brave search instead of google > mint virtual credit cards for every purchase > get a virtual mailing address > never click links inside emails > use a vpn on public wifi > block ads and trackers at the dns level > install a network monitor to see which apps are spying on you5h ago11K15420503464.80%
not enough people are emotionally prepared for the possibility that AI unlocks nuclear fusion. it could be one of the most consequential things AI ever helps us achieve. because practical fusion would give us virtually limitless clean energy. and every time energy got cheaper in history (fire, steam, oil, electricity), civilization took a giant leap forward. if we can make it reliable and affordable at scale, it'd let us: > power a workforce of 100s of millions of robots working round the clock > build entire cities on a scale humanity has never attempted > power scientific research on a scale no country can afford today > compress centuries of human progress into a single generation and AI just took us one meaningful step closer: > GPT-6 Astra just helped a physicist find new answers to a 59-year-old problem in fusion physics > the problem with fusion reactors is they have to hold something hotter than the sun in place > no one has figured out how to do it well enough to run a practical power plant yet > but with these new answers, the math finally works out for designs that could hold it in place now there’s still an enormous engineering challenge between this discovery and power plants delivering that energy... but if AI keeps helping researchers work through problems like this, it could forever change how we power civilization.frozen17d ago10K24931373623.77%
Opus 5.5's creativity and attention to detail are completely insane i asked it to make a film about AI slopbots ruining the internet and it literally oneshotted this in 20 minutes, 100% animated within Claude Code https://t.co/sOeB6EjNpOfrozen29d ago10K1064130581.81%
it's utterly insane how EU users still cannot access Muse or OpenAI's dots (thanks european commission!) but if you're in Europe and want a personal agent today, Lucas is a massively capable one you can set up right now you just text it in imessage or whatsapp and it: > connects to your email, calendar, files and other apps so it can work across them > uses its own cloud browser to navigate websites and complete tasks > remembers your preferences, plans and past conversations > proactively figures out when it can help and texts you first > books flights and restaurants, handles cancellations, chases refunds, etc > buys things with your saved card (after you approve) > keeps working in the background and texts you when it's done also, it can even run 5 errands at once, each in its own browser so a bunch of different tasks can all happen in parallel from one text. which is awesome because i love oneshotting my entire mental backlog in one voice note.frozen22d ago9.9K294111160.62%
people misunderstood what my point was I don't think everyone will start vibecoding their own SaaS but when I saw Marc post I was thinking: damn, I pay 3k for beehiiv too, I wonder if there's someone building a cheaper alternative now I think it's a bigger factor for saas that people will start heavily undercutting in price offering a similari-ish product if marc would offer me 70% of beehiivs features at 80% off i'll most likely take it and that same thing will happen to a lot of companies SaaS had an insane business model for a long time, and these times (at least for incumbents) are over imofrozen12d ago9.6K663262201.22%
this is exactly what is wrong with people in sffrozen14d ago9.6K12405041.38%
AI just helped us discover the most giant prehistoric octopus ever recorded (62 feet long!) and it completely changes our understanding of the ancient ocean food chain > Japanese researchers used a new "zero shot" AI fossil-hunting technique to find beaks hidden inside ancient rocks > octopuses rarely leave fossils because their soft bodies rot away, their hard beaks are among the few parts that survive in rocks > AI traced the fossil shapes in photos of each rock layer, so scientists could rebuild the beaks in 3D > comparing those jaws with living octopuses helped researchers estimate how enormous these animals were > the biggest fossil beak was about 50% longer than even a giant squid’s > before this discovery, researchers thought sharks and giant marine reptiles dominated these ancient oceans 100M years ago > now they say giant octopuses were top predators then too, competing with the biggest sea monsters of the dinosaur age > interestingly, their fossilized beaks were chipped, scratched and worn down from repeatedly crushing hard prey > they believe besides ammonites and fish, prehistoric sharks may even have been on the menu imagine an ancient 62 foot octopus grabbing a megalodon, wrapping those massive tentacles around it and crushing it with its beak. pretty hardcore now i need to know what else was down there lolfrozen28d ago9.5K64680171.00%
silicon valley is so obsessed with AI automating everything that they've completely lost touch with the point of life just because you *can* use AI for something doesn't mean you *should* otherwise you're robbing your soul of the most enjoyable things in life > just because AI can generate a song in 10 seconds doesn't mean you shouldn't still learn to play piano > just because AI can write doesn't mean you should use ChatGPT for your wedding vows or your mom's birthday card > just because AI can plan every hour of your Japan trip doesn't mean you should skip getting lost in the back streets of Tokyo > just because AI can translate the menu doesn't mean you shouldn't learn Italian for the pure joy of it > just because AI can summarize a 400-page novel doesn't mean you shouldn't read it the way the author actually intended i think the happiest people in the AI era will be the ones with the taste to know when not to use it AI should cure cancer and do your taxes. it should NOT be reading your kid a bedtime storyfrozen14d ago9.3K17621572202.95%
animated product launch videos are dead now better get used to showing your face on camera asap when something becomes a one shot, it loses its value as saturation turns into a commodity24d ago9.0K1134165171.72%
X average views are way down they clearly changed something in their view attribution algos so lets say usually a post with 100 likes would get 20k views average, it now has 8k anyone else noticing the same?frozen8d ago8.9K650551101.48%
another reason I think email is cooked and social media is not (as much) - people don't open their Email in expectation of a dopamine reward - if anything it's also negatively connotated to open them (invoices hitting, your boss annoying you etc) - so outsourcing the preselection to agents is much more likely - outsourcing your social media consumption to an agent 100% kills the fun of it (or dopamien slavery, whatever you wanna call it :D) so fundamentally I don't think agents will touch dopamine releasing activities as much, because humans are just too addicted to them but that's why Email as a customer channel is pretty much cookedfrozen4d ago8.8K634292161.29%
ok this is pretty insane if it works out right now, every new tiny machine needs its own custom tooling and months of setup before you can make a single one atomic machines says their system skips that and it's not a 3d printer making a shape. it builds working machines with moving parts, straight from code the same machine using different code can build completely different products, like: > a motor the size of a grain of sand > cooling built directly into AI chips > a power switch for data centers that flips ~1,000x faster than the normal kind their stated mission is "universal command of matter" we're very early, but it's just another example of how wild the next couple of years might getfrozen13d ago8.4K1041460381.92%
then you are simply not paying attention. robots are already doing more dirty work than you can imagine and nobody is mentally prepared for the tens of millions of robots that will flood everyday life and our physical world over the next 10 years. Caterpillar alone already has 820+ driverless trucks working in mines right now, and they’ve hauled over 11 billion tons of material without a driver. in oil fields, Nabors’ AI drilling automation runs on 157 rigs across 3 continents and has drilled over 90 million feet. Shell and Phillips 66 send robots to scrape toxic sludge out of their oil tanks so nobody has to climb inside. in the UK, robot dogs are swabbing radioactive rooms at Sellafield, the country’s most hazardous nuclear site. The Ocean Cleanup uses AI cameras and solar-powered robot barges to pull plastic out of rivers and the ocean, with a plan to remove 90% of the plastic floating in the ocean by 2040. and it’s still *extremely* early for robotics. autonomous firefighting helicopters, bricklaying robots, driverless bulldozers, robots that take apart old nuclear reactors, laser weeding robots, sewer-crawling robots, landmine-clearing robots and window-washing robots for skyscrapers are all being built at scale right now. and to top it all off, Tesla is designing a production line in Texas to eventually build 10 million humanoid robots a year which is literally like printing a country the size of the Czech Republic every single year just to do dirty work 24/7.frozen13d ago8.4K821191281.56%
it's a brutal reminder of time passing seeing your child grow up it's hard to notice many differences if you age from 32yo to 35 or 24 to 28 etc. that's very different if you have a child they go from 0-2 and you're like "F**K, how fast did that happen?" it's a great (but also painful reminder) to make the best out of your life and to enjoy every second of itfrozen5d ago8.2K151218052.15%
nothing compares. https://t.co/es73btZpuAfrozen13d ago8.1K239119223.23%
the reason AI still gets so much wrong about your own company is pretty simple: most of what it needs to know was never written down > your fiscal year starts in February > an "active customer" means someone who bought in the last 90 days > "behind" means under 80% of plan all of that lives in people's heads and random old Slack threads, so when you ask ChatGPT to pull a report or check your numbers, it basically just makes a confident guess CData just launched a gateway built for this. it plugs ChatGPT/Claude straight into the systems your company runs on, like Salesforce, NetSuite, SAP (and 350+ more) and: > reads live data straight from the source > holds your company's official definitions, so "active customer" means what your company means by it > learns the unwritten rules as people use it and when one person corrects it, the correction becomes memory for the whole company. so the next person who asks just gets the right answer w/o re-explaining it also lives outside any single model, so if you switch from ChatGPT to Claude (or whatever model wins next year), everything it learned comes with you CData has spent 15 years building this boring data plumbing that already runs inside Palantir, Salesforce and Google Cloudfrozen23d ago8.1K27550350.89%
I sometimes wonder how many remote workers have replaced themselves fully with an agent without anyone noticingfrozen6d ago7.9K116333172.02%
how is it possible that VC keeps funding the same 3 products lmao everyone is building the same stufffrozen24d ago7.9K61122031.11%
what timeline are we in where Zuck is mogging Astra + Fable now lmao infinite money + infinite distribution + now frontier intelligence long Meta https://t.co/aUDEv0e4T8frozen10d ago7.7K71570131.24%
this makes me SO pumped it's never been a better time to start working on video :)3d ago7.7K29230320.86%
everyone in AI is suddenly talking about data brokering (selling your company’s data to AI labs for $300k-$3M+ a pop) the whole thing seems kinda fishy but interesting at the same time so i did some digging myself (talked to a bunch of companies and people involved in this) here are my unfiltered notes: 1) the alpha seems to come from good connections to the actual lab researchers telling you what data they are looking for right now (the successful brokers have connections to a bunch of different labs, including smaller ones you have never heard of) 2) it is def real, and the major player doing it are growing fast as fuck. 3) the kind of data the labs want change fast 4) people who are thriving have m&a background OR are extremely good affiliate marketers feeding of fees from getting more "brokers" into their network 5) noone has an idea if its just getting started or at the peak of the bubble (don't believe what any of the brokers say to you) 6) it attracted a bunch of get rich quick guys, mainly ex crypto or affiliate marketers (some of them dubious) 7) almost all "companies" I saw seemed highly unprofessional, no insight into team, company setup. there are exceptions though 8) I don't think most people are scammy, just highly inexperienced (lots of 20yos pretending to be experts, classic) 9) some people theorize that private companies will also need more private data soon to train their own models, that could expand the markets away from just bigAI buying data (no idea if that's true or not) 10) there's a bunch of companies offering services like de-identification of data who are thriving off the back of this trend (also insurance in that space seems to be booming) 11) some big tech newsletters have started to advertise for this as well (saw bay area times doing it) and I think we'll see more creators joining this soon as well 12) what brokers love to not tell you is that selling your data can massively impact your ability to sell your company later 13) this is a purely US based play, at least right now (I suspect GDPR in europe would never allow this), but this means there might be opportunity looking outside of the U 14) if you're interested in this, don't talk to the brokers a lot of them are sleezy salesmen promising you returns that 99% will never get. talk to a wider variety of people being involved 15) people on X make it seem more illegal to sell your data than it really is. in the end, noone is "tricked" into selling it 16) there also seems to be the case of peopel selling data of dead companies, which I found fascinating (but didnt talk to anyone who really did so, I think spirit airlines selling to google is a big example) So my final verdict? It is an interesting space and very obvious it's an opportunity window (that might already be closing who knows) If you are a legit company in this space I would love to talk to you (just DM me) I'll keep digging to see if there's something worth building here I thought it would be interesting to share this1d ago7.7K773201682.21%
people still don't fully appreciate how insane the last 4 months in medicine have been: > a personalized cancer vaccine passed a late-stage trial for the first time ever. it's built from the DNA of your own tumor and trains your immune system to hunt down melanoma so it doesn't come back > the FDA approved a pill that nearly doubles how long pancreatic cancer patients live compared to chemo. it's one of the deadliest cancers there is > there's now a daily pill that cuts bad cholesterol (the stuff that clogs your arteries) by more than half. until now you needed injections to get anywhere close to that > kids with sanfilippo syndrome, a disease that gives them dementia and kills most before they turn 20, got their first ever treatment. it's a one-time gene therapy that gives them the working gene they were born missing > a new lung cancer drug went head to head with keytruda, the best-selling drug in the world, and won. patients lived 8 months longer > the FDA approved the first heart valve that grows with a child, so kids born with heart defects can skip repeat open heart surgeries > and now a pill regrew hair in 500+ men with no serious side effects and that's just four months! we are genuinely not ready for what the next few years of this parabola looks like. whatever you do, just stay alive2d ago7.5K81674351.78%
this is so unimaginable in europe lmao but it's also clearly the future for a lot of regular sicknesses such a waste of everyones time to go to a doctor with something where you clearly just need a simple prescription (also taking up space for people who actually should see a doctor) I guess the important question is how to decide if the AI doctor is sufficient or notfrozen17d ago7.4K8910133101.69%
damn Jev just got lapped. Snap now has 12x Jev's decision context. and these decision models mog LLMs for micro-decisions where your agent needs an instant answer: > which ticket needs a human? > is this lead worth chasing? > does this contract clause break our rules? otherwise, you're just wasting time and tokens waiting for Claude to generate an answer when your agent just needs: > yes/no > a choice > or a score its basically like hiring a rocket scientist to sort support tickets. and with a massive 384K tokens, Snap can factor in more customer history, past deals and company rules when making those calls.frozen10d ago7.4K49520421.33%
ok people REALLY hate the idea of robocops lol11h ago7.1K31012110.64%
there's insane cope of people telling me "our Emails are still fine" it's not about the right now, but where things are headed email marketing will fundamentally change when it's mainly bots scanning inboxes and for a lot of email operators, this insight will come much too late if you can't see that, that's on you btw, that's coming someone who was the biggest Email bull ever and made most of his income via an Email list i'm in the deep end of this stuff lolfrozen4d ago7.0K451262131.24%
it is really happening OpenAI just dropped a massive release of hundreds of solved math problems that humans couldn't figure out for thousands of years I feel existential angst and overwhelming excitement at the same time if you still think AI is not coming for your work, prepare for a nuclear bomb to hit it (meaning: it will change, not that everyone will be jobless) I know I am, even though it still feels odd humans will adapt like always and we'll find new things to do but it's also never been more obvious to me why you need to find purpose outside of your work sitting at a computer and why it's never been more important to start a family, to grow community, to go outside and connect with the natural world it's time to rediscover what being human meansfrozen16d ago6.9K706104141.51%
tesseract + claude are making me rethink what's possible with a good video editor. opus 5.5 is already getting seriously good at creating animations and visual explainers. for my talking-head videos, my editor can combine that with tesseract, give it the script, recorded footage and reference videos, then direct it to: 1. edit the footage to match the references’ pacing, transitions and animated text 2. create little animated explainers that show what i'm talking about 3. time those animations to the exact moments i explain each idea 4. bring in relevant b-roll 5. cut between me, the animations and b-roll, with music and sound effects timed around it and tesseract's new converters let it work on supported premiere and after effects projects, then export them back to adobe so my editor can keep shaping the final cut. the exciting part is being able to try visual ideas we'd otherwise struggle to justify spending hours on. custom explainers, different animation styles, multiple versions of a scene. really feeling jevons paradox here. the easier these things get to produce, the more possibilities i see. it makes me want to spend more tokens exploring those ideas and get way more ambitious with our videos.frozen16d ago6.9K67740762.24%
most people are still sleeping on this "Intelligent UI" shift happening rn since ChatGPT came out every AI answer has been a wall of text no matter what you asked... comparing flights, checking your budget, planning a workout, all paragraphs but now ChatGPT answers with the actual thing you need, like generated media, charts, table, or buttons you can click which is 10x more engaging/interactive than slop paragraphs and OpenUI just made the same idea open source, so any app or agent can do it pretty high confidence that in a year, getting an essay back from an AI will feel as dated as a website with no picturesfrozen7d ago6.9K66440471.76%
honey come down, Google is so fucking backfrozen22d ago6.8K3836040.75%
nothing more annoying than someone having some sings of success in one domain and immediately turning into public expert on every topic on earth24d ago6.5K89017131.69%
I'm wondering how many people are trying to buy struggling companies right now just to sell the data to the ai labs while the owner has no idea that it could be valuablefrozen7d ago6.4K81014181.62%
meta has this insane unique position of having massive distribution AND insane amounts of cash + being at the forefront of everything consumer can't really fuck with zuck free agents will winfrozen29d ago6.4K61113071.28%
looking for a desk at a co working space/office in mallorca best not too far from santa ponsa, but I have a car maybe the hivemind can helpfrozen15d ago6.3K2119020.52%
Any guesses how meta will weave in their giant ad business into muse?frozen28d ago6.3K14016110.51%
vibe coding is moving into the physical world. even hardware beginners like me can now turn our weird little ideas into physical devices that actually work. you just describe what you want. then AI designs it, finds the parts, writes the code, and shows you how to put it together. here are 12 of my favorite examples of "vibe manufacturing": 1. Sourish turned an MRI of his brain into a life-size 3d print.frozen16d ago6.2K31531281.10%
I used to hate buying domains, registering, deploying the page, DNS records yada yada yada so manual and annoying now I can do everything via codex + vercel finding domain, buying it, deploying the page for a nontechnical noob like me this is heavenfrozen13d ago6.1K640300111.72%
everyone is so uncertain right now what is still worth building but the worst thing you can do rn is to freeze and to do nothing don't forget: times of uncertainty have ALWAYS always brought the biggest opportunities but you need to keep tryingfrozen10d ago6.0K1005150182.30%
grokbot feels so damn fast and it's another reminder that not the best but the fastest model (at 80% capability of the frontier) will most likely win in consumerfrozen9d ago5.9K103317252.19%
as i've said before, people will pick cozymaxxed robots over cold metal ones every time going cozy is the only way past the uncanny valley and soullessness of robotics we already saw it happen with Muse and digital agents, and now it's carrying over to physical robots people are gonna want something closer to a stuffed animal with some warmth and personality, at home and out in the world bullish on robots you actually want to hugfrozen6d ago5.8K24380120.81%
ok doing taxes for crypto was always annoying af i'm trying an AI first MCP solution this time will report backfrozen11d ago5.7K36112000.86%
ngl Wabi mogs Muse, Instinct and Dots on aesthetics every answer comes back as a little interactive app built around whatever you’re doing. deeply personal on-demand apps are the futurefrozen3d ago5.5K30060201.02%
A lot of startups will learn a hard lesson trying to compete with a free agentfrozen27d ago5.5K28111000.73%
7 months ago i posted this and it got 4.4M views but i’m curious, has your marketing team actually gotten smaller since then? surprisingly, for most people i talk to, the answer is still no. and i’ve been wondering... why is that? shouldn't AI be gradually making teams leaner? i think part of it is Jevons paradox: teams just get more ambitious and everyone starts tokenmaxxing harder lol. but i think a lot of it is also a skill issue (and honestly, understandable) getting agents to reliably do useful work for your business is STILL giga confusing, especially if you’re non-technical. you need tons of context and data, the right tools connected, and it’s a genuine pain in the ass to set up you can’t expect a business owner or a department inside a big company to figure this out on the side. learning to set it up in a safe, secure, trustworthy way is basically a full-time job and it’s especially true for GTM the growth marketing guy at Anthropic had Claude plugged into his ad data, a custom Figma plugin, an MCP server, and a memory system. but most teams have NONE of that so when they ask Claude to “find 50 warm leads and reach out,” it understands the task. but it doesn’t know their customers, it has no live lead data, and it has nowhere safe to send from so when they ask Claude to “find 50 warm leads and reach out,” it understands the task. but it doesn’t know their customers, it has no live lead data, and it has nowhere safe to send from so i’ve been trying out a bunch of the GTM MCPs that plug into Claude to see which ones actually close that gap Apollo and Clay both have one now, so Claude can search for leads and enrich contacts directly. each one covers a single piece of the workflow though, so you’re still wiring the rest together yourself another one i'm testing rn is Pancake MCP, which tries to handle the whole loop. it builds a “GTM brain” from your website first, then its agents use that to find warm leads and run outbound, and it learns from each campaign still early with all of them, but will share which ones actually hold up!frozen9d ago5.4K131142150.84%
you need to be even more obsessed with copywriting on short-form video vs writing on X the simple reason is that when you read something it's easier to skip lines that don't interest you but if you lose someone in a video, even if its just because of 3 wrong words, they are OUTfrozen23d ago5.2K31010050.88%
I have huge fomo that I couldn't got :( Fellows fund is doing this event rn with some giga cracked people from both AI research, AI startups and VC was suppose to fly to sf but life had other plans at least I can watch it on x https://t.co/y6hI8Iy6ZRfrozen29d ago5.2K1132130.38%
some people always looked down on others "only" doing traffic/distribution and not a "real product" (creators, influencer whatever you wanna call them) funny enough, distribution seems to be much more antifragile and in demand in this new AI era and what do you think where these billions of raised AI VC capital are going?frozen10d ago5.2K2728020.76%
I wrote 5 early stage tickets in the last 2 years and 4 went to zero and the one I completely doubted is about to hit a 20-50x I think lmao I guess it really proves that the best ideas often look stupid/like play before everyone agrees they are worth somethingfrozen24d ago5.1K2904020.68%
the Jurassic Park company that brought back the dire wolf is spinning out some insanely sci-fi biotech startups here's what's come out of Colossal so far: > Astromech: they built an AI that works like a weather forecast for evolution. it learns from 3.8 billion years of life on earth to predict how living things will change next, like which bacteria are about to become drug resistant (it's already valued at $3.8B) > Form Bio: gene therapies are some of the most expensive drugs on earth (some cost over $3M a dose) and they're brutally hard to manufacture. Form Bio sells drug companies an AI that test-runs the designs inside a computer and flags the ones that would fail in the factory > Breaking: they found a microbe that ate up to 90% of some common plastics in under 2 years in lab tests, leaving behind just CO2, water and biomass. now they're turning it into a service to clean up plastic waste > next up: Colossal is building an artificial womb. they've already hatched 26 chicks in it and are working toward the first mammal born completely outside its mother every extinct animal Colossal brings back forces them to invent new tools, and those tools turn into these new companies bio/accfrozen5d ago5.1K26431130.92%
anyone else think it's crazy how multiplayer AI still isn't really a thing?! i'm constantly copying messages, screenshotting chats in my team Discord, re-explaining context to my agents, etc i've basically become a human API between my team and my agents lol this is why the kylon mogs: it's a shared chat workspace where your team can talk directly to claude code and codex. it connects to slack, so it can sit alongside your existing setup. but i could also see it even fully replacing slack or discord for some teams i think all this copy-pasting is going to feel incredibly archaic when we look back in a few years. your agents will already be part of every conversation.frozen9d ago5.1K242120261.25%
the appetite for compute and attention is infinite position yourself accordinglyfrozen25d ago5.1K39210181.19%
there's no upside in public drama it only sucks your energy and is a confirmed aura L just don't five a fuck and tomorrow the next nothingburger takes the attention of the tik-tokbrained individual2d ago5.1K3302120.75%
I was never much of a beach person but seeing the joy it bring to my sons face suddenly makes it the best place on earthfrozen17d ago5.0K5705001.24%
OpenAI just bought this startup to get new drugs to you faster right now it's an agonizingly slow process (takes 10-15 years), and most of those years go into the long slog of studies and trials that's exactly what Orchestra's agents handle. they take over the busywork scientists usually do by hand (checking on studies, chasing deadlines, spotting delays early) and point them to the next step the moment results come in. so now OpenAI has AI for finding the drug AND getting it through the system because back in April OpenAI launched GPT-Rosalind, an AI model for drug discovery Moderna, Amgen and Novo Nordisk already use it feels like they want every big pharma company running their entire drug development process on OpenAI2d ago5.0K1712030.46%
why AI doomers are in shambles today (for dummies): >be humanity >build AI that’s about to be way smarter than us >have no idea what’s going on inside it >can only see what it says and does >it passes every safety test >but no way to tell if it’s actually safe or just pretending >need a way to look inside its head >call it interpretability >spend years getting nowhere >2023: OpenAI can only figure out what about 2% of an AI’s parts do >AI keeps getting bigger, more confusing, more capable >doomers say we’re cooked >even the guy who started that OpenAI project, Nick Cammarata, loses hope >3 years pass >Nick tries again, this time with AI agents doing the digging >turns out, about 90% of the parts have one clear job you can read >and most of the rest are readable too >even in huge models >we might actually be able to read an AI’s mind >and check if it’s really aligned with us >before we hand it our power grids, weapons and every other system the world runs on >full paper isn’t out yet >but humanity might make it after allfrozen6d ago5.0K35350201.27%
A lot of AI stonks are not looking good micron, sandisk, nebius all look topped af anthropic revenue also slightly disappointed (obviously still insane growth, but clearly plateauing a bit here) this + upcoming IPO makes me think we're hitting some kind of ( medium term) pivotal point in the markets here def it's time to watch out a bit rnfrozen7d ago4.9K18010110.61%
how to move countries with a small child (<2yo) without stressing them too much (and keeping yourself sane) some people were asking so here are some things I found helpful (after 4 moves) I am not an expert by any means, this is just my personal experience: - always plan travel around their sleep schedule, book flights accordingly so they can stick to the routine (what works great for us is a flight mid day where we enter the plane around his bedtime and he sleeps at least for like 1-1.5h on the flight). don't try to save money on flights if it collides with their schedule. - pretend you are moving out one day earlier than you actually do and try to hit that deadline. everything will be so much slower and more complicated that you'll need the extra day lol - avoid big gathering with a lot of kids in the couple days prior, don't do too many crazy activities that could lead to injuries (you don't want a sick or injured child for travel makes it 100x worse) - if you have family around, get them to just be with the child while you can pack (packing with a child is A LOT HARDER than most people think, especially if it's your whole apartment) - start your travel in the morning when they are still rested from the night - bring familiar elements to their new home from the very start (we sent the same playmats, toys, bed etc via DHL so we can recreate their old environment immediately) - keep as many variables the same as possible, same bed routine, same toothbrushing routine, same food etc - pack a first night bag with the bare essentials you need for the baby so you don't have to - don't interrupt their circadian rhythm too much also on travel days, be sure to have enough light exposure (especially morning & evening) - if your child is still small and very curious like ours, buy all safety stuff before you go and keep them in a separate bag (doorlocks, corner guards, outlet covers, drawer locks etc). we also buy black out blinds (there are ones you can just add without drilling on amazon) - buy a whitenoise machine, you never know what the situation is at your new place (tbh, you should buy one anyway it's a great thing to own for a better sleeping environment) - get help for leaving your apartment, book movers, cleaners etc try to offload as much as possible - don't wash their bedding before the move, a familiar smell goes a long way - find local family communities (there's a bunch on whatsapp) and find a doctor before you move (we learned this the hard way), also check out nearest ERs etc you don't want to start asking chatgpt when something is urgent that all being said, i'm looking forward to long term settling in 2027!frozen17d ago4.9K1302060.43%
you're probably owed money from a class action settlement right now and have no idea. and AI agents are about to start collecting it for you. companies paid a record $70B+ to settle class actions in 2025. but to get your share of these settlements, you have to file a claim. and almost nobody does. for example, take this 23andMe settlement (which is paying out right now): > 6.4 million Americans qualified > only about 250,000 filed since so few people filed, the checks came out way bigger than expected. this is why Lauren filed expecting $20 and got $1,037. most settlements look like this. the FTC found the median claim rate was only 9%. the reason: a lot of these notices come by email and look like spam, so most people ignore them. but agents read every email, even the ones you'd never open. so if you give an agent like Muse/Grokbot access to your inbox, it can find every settlement notice, check if you qualify, and file the claim for you. almost nobody has this set up yet, so whoever does gets the money everyone else leaves on the table.frozen21d ago4.8K1222080.50%
brain-computer interfaces in 2026 feel like AI in 2019 to me. and i think it's OBVIOUS the winners will be non-invasive (like Sabi). nobody's gonna wanna have chips surgically implanted in their brains (just look at how much the public already hates AI) so a cap you can easily put on/take off makes the jump to "thought-prompting" AI much easier to sell. and when that AI can operate your devices and robots, you can imagine an extremely sci-fi future: > control the lights, blinds, temperature and music in your home just by thinking > summon your car to the front door while you're putting your shoes on > think “get the house ready, friends are coming over” and have your humanoid tidy up and set the table before you even get home extremely bullish on Sabi and brain-computer interfacesfrozen13d ago4.7K42480101.36%
1945 is so back AI researchers entering their Oppenheimer arc we're literally speedrunning the same Manhattan Project script but with AI now: 1. researchers then and now both openly put odds on whether it ends the world. before the first bomb test, Fermi jokingly took bets on whether it would set the atmosphere on fire. today AI researchers casually share their "p(doom)", their personal odds that AI goes catastrophically wrong (btw Dario's is 25% rn) 2. it's framed as a race against a rival superpower (Germany then, China now) 3. the budgets look insane. the Manhattan Project cost about $2B (about $30B in today's money). a single big tech company now spends more than that on AI data centers every year 4. the people who built it start begging everyone to slow down. Oppenheimer spent the years after pushing for international control of the bomb. now Dario, Sam Altman and Elon are all begging for a slowdown and regulation and Oppenheimer chain-smoked through the whole thingfrozen9d ago4.6K27370101.01%
as Europeans, this chart should deeply concern us Europe is not only lacking on AI infra and frontier labs, but also most people here are not participating in any financial upside in the US $47k per resident were captured in AI related equities in the EU, it's only $4.5k another problem is, if you don't participate in any upside it's way more like to revolt against the buildout which widens the gap even more it's a cycle that feeds itself if we want people to feel optimistic about AI and the future, they need to feel positive impact on their lifesfrozen9d ago4.5K695212142.44%
it's funny how just 1-2 years ago you had to be so specific and deliberate with your prompting to achieve a certain look/style but now you can just casually say “make it look like our other videos” lolfrozen13d ago4.5K1923050.64%
one funny thing agents are exposing is how many businesses are just built on human laziness they count on you being too lazy to cancel, appeal, claim, or fill out the paperwork, which worked great for them until AI agents showed up things you should send your agent after right now: > cancel subscriptions you forgot about (~50% of subscription revenue comes from people forgetting to cancel) > earn 30-50x more by moving your idle cash into high-yield savings (checking pays 0.1% on average, high-yield pays 3-5%) > negotiate your internet and phone bills (1 Muse user had it negotiate with Comcast and saved $800+/year) > claim unclaimed money in your name (states are holding $70B+, owed to 1 in 7 Americans) > track down old 401ks ($2.1T is sitting in forgotten accounts) > spend your health spending account (FSA) before it expires (47% of holders forfeit $422 on average) > appeal insurance denials (67% of appealed Medicare Advantage denials get reversed) > file for class action payouts you qualify for (a record $79B in settlements last year, and only ~9% of eligible people claim)frozen23d ago4.5K32491311.72%
American open source is FLYING HIGH again 🇺🇸 Zuck this week: ships Muse Wonder, open-sources Muse Spark, and gets Apple to tweak its own software so the thing runs well on a Mac 18 months ago people were dunking on Llama 4's benchmarks. now Meta's open-source model beats Kimi K3, GLM 5.3 and Qwen on agent tool use and the hardest reasoning test (HLE). what a turnaround story Meta is right nowfrozen7d ago4.5K3242060.98%
AI is solving wildlife conservation. it's effectively becoming the best bodyguard endangered animals have ever had. a few examples: > in California, AI thermal cameras detect whales and warn ships to slow down or change course before they hit and kill one (happens more than you think). the tech has logged 250,000+ verified detections so far > in India, AI thermal cameras spot elephants near train tracks so drivers can brake in time. one corridor logged 6,592 safe crossings and 0 elephant deaths by September 2025 > in China, Huawei's AI spots fires threatening panda habitat so rangers can put them out early. it covered 8.2 million hectares and saw 71% fewer major fires in 2021 > in Kenya, AI thermal cameras spot poachers in the dark so rangers can stop them before they reach the rhinos. the program now covers reserves holding 80%+ of the country's rhinos > in Cambodia, an AI predicts where poachers will set traps so rangers can pull them out before an animal steps in. rangers using it recovered 1,000+ snares in 6 weeks > in Australia, Google's AI sorted 7 million wildlife photos in weeks to show where koalas and wombats survived the bushfires, so conservationists knew where to focus recovery efforts > in India's Similipal Tiger Reserve, TrailGuard AI cameras alert rangers the moment poachers walk in. in 10 months, officials arrested 96 poachers and seized 86 guns > in Kenya, AI identifies individual Grevy's zebras by their unique stripes, so conservationists can count and track the roughly 3,000 left in the wild this is the kind of AI progress i want more of.frozen15d ago4.5K3464271.18%
this post is such a good reminder of how AI just totally warps your sense of time less than a year ago, Karpathy says he was still typing out code by hand, character by character it literally feels like 5 to 10 years ago. the idea of anyone manually coding today is just completely unfathomable to me lol but vibe coding only really hit its inflection point last december! so it’s insane to think about where we’ll be a year from now, in october 2027 will it feel unfathomable that anyone still: > clicks around websites themselves? > does their own taxes? > drives their own car? > edits videos by hand? > loads their own dishwasher? > folds their own laundry? i'm pretty sure this is what the early singularity feels like9d ago4.5K241100101.01%
what a lot of the sf guys miss is how many people LOVE shopping shopping is a dopamin rush they love looking at products and crave the feeling to blast their credit card also what do you think people will do when AI frees up more of their time? spend LESS time on their phone? Advertising will thrive, the channels might change but thinking agents will lead to less advertising is a stupid takefrozen17d ago4.4K2927000.86%
Elon's (probably) joking, but it's a good lesson for every company: be an agent MAGNET, not an agent REPELLENT. soon there will be more agentic customers than human customers. here’s a few easy ways to make sure agents pick you: > send an agent to buy something from your own site. wherever it gets stuck is your to-do list > check your bot settings. Cloudflare and other bot tools can block AI agents by default, and Cloudflare now has a separate switch just for agents > make checkout easy for a bot: guest checkout, no forced sign-up, no CAPTCHA wall. agents usually check out as guests > describe products plainly, with specs, materials and sizes. an agent can find an "organic cotton sleep mask," but it has no idea what "The Luna" is > keep price and stock accurate everywhere. agents cross-check, and stale data gets you dropped > turn on the agent sales channels. Shopify has one built in, and Stripe offers one for everyone else yes, this will threaten some business models, but you have no choice but to adapt. otherwise you're gifting customers to your competition (or to some agent-first company that doesn't exist... yet).frozen7d ago4.4K454141211.95%
i'm convinced fast + reliable browser use is still the biggest bottleneck for agents. grokbot, muse, whatever agent you're using... you still can't reliably leave them fully unattended. they ALWAYS seem to hit a snag on a website, need you to click something, or spend ages filling out a form you could've finished yourself. and most people don't want to keep signing in or setting up a connector for every website they use. i think gumloop is taking the right approach with their new browser. you can use your favorite models like opus 5.5 and gpt 6-astra with a browser setup that cuts down on all that babysitting: > credential vault: saves your login credentials so the agent can reuse them next time > recipes: work through a task together once, check the result, then save the steps for much faster future runs > cloud browser: runs tasks in its own sandbox in the cloud, so your laptop doesn't have to stay open > session replays: records every session so you can watch exactly what it did afterward once you can trust an agent enough to leave it fully unattended and come back to finished work (and not stuck on a website), the amount you can delegate goes wayyy upfrozen16d ago4.3K20440100.88%
How I turned Claude into my free video retention coach for IG, YouTube, and TikTok: (it tells me the exact moment people stop watching my videos, and what i was saying when they left) every reel, TikTok and YouTube video has a retention graph, but none of them tell you why people left. but Claude can tell you. you can have it download your video’s audio, transcribe it with timestamps, line that up with the graph, and quote exactly what you were saying at every big drop. then get it to tell you the one change that would keep the most people watching. here's how to do it: 1. you screenshot the retention graph from your video's insights. on Instagram and TikTok it only shows up in the phone app, so grab it there (on YouTube, you can download the exact data from YouTube Studio instead). 2. you open Claude Code and send it this post, your video link and the screenshot. that's it. Claude Code installs whatever it needs and does the rest. run it on your last video that flopped, and you'll know exactly what to fix on the next one.3d ago4.0K22370421.83%
waiting for images to generate is one of the most annoying parts of AI. but Krea 3 now gets you: &gt; Nano Banana Pro quality &gt; 4x faster &gt; at 1/4 of the cost need i say much else? https://t.co/mWonVru0Gifrozen7d ago4.0K26150131.11%
Mallorca people, I want to organize a meetup pls comment if you're here i'll create a group organize this I think going on a morning hike or run + getting breakfast after would be great Probably around Palma https://t.co/seBOosCm1J2d ago3.9K1909000.71%
it's insane to think about a world where we become this intertwined with AI like a few years ago giving AI access to your emails and all your data seemed scary but now everyone does it because it's so much more useful when it has that context right now having AI in your brain seems scary... but give it a few years and people will probably just adapt to that too? pretty crazy to ponderfrozen5d ago3.9K2515050.92%
so AI can be passive-aggressive now? lol voice AI always struggled to get the "attitude" behind the words right like there's endless ways to say “i’m so happy for you”: > genuinely stoked your best friend got their dream job > jealous your least competent coworker got promoted over you > trying not to cry because your ex just got engaged but now with drama 3, you can just direct the attitude behind the words in plain english: “sound happy for them, but make it obvious you’re dying inside.” its not what you say, its how you say itfrozen9d ago3.9K1423020.54%
after the Hugging Face incident and now the Saudi spy inside OpenAI, i think we need a lot more security weaknesses discovered through research like this i don't trust the labs to be the only ones checking their own security. we NEED independent outsiders: > stress-testing model safeguards to find what harmful content still gets through > testing whether AI agents can escape their sandboxes or access restricted systems > checking who can access sensitive rooms and data needs to be a much bigger priority imofrozen8d ago3.9K25310071.16%
there's an old rule in engineering: fast, cheap, good: pick 2 this has been true for AI image models, until now Krea 3 just tied Google's Nano Banana 2.1 on Artificial Analysis's blind leaderboard (1151 vs 1152), and it edits about 4x faster for less than half the price so now you get all three :)2d ago3.3K2623041.06%
a few friends will be judging / hosting this 👇 if you've built some cool vibe coded projects, send in your portfolio by oct 15. winner gets $10k + a direct YC interview! great chance to opportunitymaxxfrozen9d ago3.3K742030.49%
I shall rise in this new paradigm2d ago2.2K302020.33%
I called BS to this "AI will create as many jobs as it destroys" narrative from the very beginning the data of replacement is just a lagging indicator we are finally seeing the first effects I myself am replacing people I hired before with AI everyone is doing it, everyone I know is hiring less people and somehow magically we won't lose jobs?! wake up this is not a doomer take, I am highly optimistic about AI but I also have common sense and eyes and ears we need to talk about this and find solutions11d ago1.8K23210122.08%
rn it seems impossible to imagine a world where i don't pay megacorps like OpenAI / Anthropic / xAI every month to use AI but i just learned people in the 1960s felt the same about computers. back then everyone had to rent time on computers from a few giant companies then Apple etc came along, everyone got their own computer in their own home, and the internet connected them all so it's possible that AI follows a similar path. everyone gets an AI model they own, and those get connected too, so they work together like one big brain i'm not 100% sure it happens, but it's a fascinating future to consider4h ago1.6K1406051.60%
my messaging channels are ALL over the place X Dms (horrible chaos), several emails, discords, whatsapp, telegram etc i've just connected everything via grokbot, even whatsapp worked via webbrowser access what a delightful day8d ago900203000.56%
lol i never really considered that all the footage robots train on doesn’t actually tell them how hard to grip things3h ago278000000.00%