← Back to leaderboard
biscuitweb3

Biscuit

Verifiedactive⚠ low-confidence
@biscuitweb3

creative director @LabyrnthCo - AI/Web3 video content creator, educational videos | member of @aiscwork

22,383 followers · 4,864 following · Last polled 11d ago
Perf. Score
6.7
Rank #386
Value Score
—
set rates to compute
Median views
0
p25 0 · trailing 7d
Followers
22K
0.0% 7d
ER (impr.)
0.00%
eng ÷ impressions
ER (foll.)
0.00%
eng ÷ followers
Est. CPM
—
needs a rate
WoW views
-100.0%
flat

Rates & value · est. from 0 median organic views

Quote tweet
—
no CPM
Post
—
no CPM
Thread
—
no CPM
Retweet
—
no CPM
rate 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 (0), not the mean (0) — so a single viral post doesn't inflate a creator's reach. The p25 floor (0) is the level 75% of posts clear.

Not enough in-window posts yet to show a distribution.

Recent posts · ranked by views

PostPostedViewsLikesRepostsRepliesQuotesBookmarksER
i spend enough time switching between Claude Code and Codex already @alook_ai lets you bring both into a Discord-like room where your teammates can talk to them directly they still run locally on your computer i want to try having both agents in the same chat14d ago49K78646040.28%
i already knew what i wanted to make, i just didn’t want to build the whole video from zero so i used @Pexoai_offical, picked a template and changed the text and visuals it’s a conversational AI video agent, so if i wanted something adjusted i could just ask for it didn’t have to spend time going through every setting myselffrozen29d ago28K10233560480.84%
my goal is to build a career i cannot explain with one job title → making educational videos for ai startups → using ai to build my own products → growing my content into a real business → working from anywhere → learning whatever i need along the way what’s yours? https://t.co/S2A7shUp9Ufrozen
28d ago
7.0K
100
1
69
3
5
2.55%
if a creator opens your product for the first time on filming day, people can tell one video repeats the landing page five videos teach the creator what users need explained i want to do that with ai startups learn the product, educate users and represent them beyond one launch https://t.co/xknJlW9TUxfrozen27d ago7.0K1371860173.46%
job hunting might be one of the best use cases for an ai agent find roles that actually match me, rewrite the resume for each one, fill out the application then just let me review before anything gets sent this is what i want ai doingfrozen21d ago6.9K108054082.46%
gm 🌸 how are you planning to spend today?✨ https://t.co/rYOkaEuUZcfrozen19d ago6.8K2061104024.59%
what if you read this and earn your first 5 figs questionsfrozen29d ago6.5K957310542.86%
have a great weekend guys ✨ https://t.co/IIBNnQzraY12d ago5.9K161092034.33%
you didn’t realize how Jev works for you Jev does not replace the model writing your post or doing deeper research. It handles the repetitive decisions around what to keep, ignore, review or send forward 1. content idea filtering → give it customer calls, old notes, published posts and your website. Jev can identify repeated topics and leave you with ideas that need new content 2. video clip selection → split a long video into shorter clips. Jev can check whether each clip has a complete thought, enough context and a useful point before it reaches editing 3. X content analysis → add your published posts and performance data. Jev can group them by topic, opening and format to find patterns that continue working across several posts 4. comment prioritization → run comments from X, YouTube or Product Hunt through Jev. It can separate reactions from questions, content ideas and customer problems that need an answer 5. customer feedback → give it reviews, onboarding messages and call transcripts. Jev can group repeated complaints, feature requests, buying signals and places where users keep getting stuck 6. model routing → Jev can decide whether a request needs a smaller model, a stronger model or human review. The final work still reaches the right place without using an expensive model for every small decisionfrozen27d ago5.7K1201109034.08%
AI agents are already signing up for apps, but they usually end up using our accounts AgentID gives them their own login and links it back to the person who owns the agent basically “Sign in with Google” for agents @haakamaujla says it’s free for apps and takes a few minutes to add there’s also 3 months of AgentMail free if you add it from the launch threadfrozen16d ago5.6K52157032.00%
i made this little animation after seeing the haiku 5.5 demo one bridge, three separate checks running at once, then everything goes back to opus for review now i want to try it on something i’m actually building https://t.co/eJwUQUvtZXfrozen15d ago5.5K65023141.70%
gm🌸 any team looking for a long-term video creator partner? ✨ https://t.co/koXwZLrjFQfrozen29d ago5.5K1491106024.72%
gm ✨ what are you up to today? https://t.co/cUMx9Vb0q1frozen15d ago5.4K122174003.62%
one setting dropped haiku 5.5’s benchmark score by 466 points on aa-briefcase → max effort: 1578 → low effort: 1112 before choosing settings for the workers in yesterday’s animation, here’s the test i’d run give haiku your files and 5 questions → 3 with answers you already know → 1 with conflicting sources → 1 the files cannot answer use this instruction at both effort levels “cite the source for each answer flag conflicting information say when the files don’t contain the answer” record wrong answers, missed conflicts, invented details, time and costfrozen14d ago5.4K4317020.99%
grok bot templates can waste your time if i have to rebuild half the template i am not calling that a shortcut https://t.co/BuyJGDuBT8frozen19d ago5.4K83024052.09%
One thing I want to ask Ron Paul: what actually makes you use crypto for payments instead of just holding it? Find out today, he's joining @TEXITcoin Honest Money Live Follow along so you don't miss itfrozen1mo ago5.3K67039012.00%
agents can already build the app, but someone still has to deploy it, scale it and keep it running instacloud handles that side and can copy the app with its data so changes can be tested before going live now the agent can do more than just write the codefrozen23d ago5.2K96246032.81%
my next ai project needs a paying customer while i drink my tea, i can use ai to start building an agent that checks whether someone is paying too much for the ai tasks they run every day the instruction “take these 20 tasks and their acceptance criteria test 3 models within a fixed budget record failed attempts and retries calculate the cost per result that passes flag anything needing human review” what i’d sell is a report showing which tasks can move to a cheaper model and which should stay then offer ongoing checks as models and prices change the first customer would be someone with an existing api bill who wants evidence before switching12d ago5.2K75113141.82%
proud to say i'm the "the creator" he mentioned in this article what a bangerfrozen16d ago5.1K63015041.62%
happy weekend 💕 today i’m just having a good coffee and spending time with people i love what are you guys up to?✨ https://t.co/qq0MdCNexkfrozen27d ago4.7K100074003.69%
xp is live and tge is coming in q1 2027 👀 everything you do on @eldoraglobal counts right now trades referrals all of it still the most honest review i can give stack xp early and the rest follows you can check here: https://t.co/GBd86JTaU4frozen28d ago4.7K64034032.17%
claude did worse at max sonnet 5.5 scored 52.1% at xhigh and 46.2% at max in one coding test in two runs extra checks caused a timeout or changes outside the task i’d check what it changed before using max https://t.co/pS9hljwQMwfrozen21d ago4.6K81351012.97%
great essay by @spisallyouneed large language models can do a lot, but they still have limits once the task goes beyond language @skyfall_ai is working on world models and recursive self improvement, with the idea of building AI that can understand more than text and take on much bigger tasks, even something like an AI CEO i’ve been reading more about this lately, so i wanted to hear their take on what comes after automationfrozen17d ago4.5K84241032.86%
$KAIO is up 113% in 24h and volume is up 124% there’s more going on than just the price too @KAIO_xyz already brought a Mubadala Capital strategy onchain, KASH is planned for Q4, and tokenised RWAs are getting close to $40B https://t.co/bD6gkdTL1Kfrozen23d ago4.5K84544033.00%
i want an ai agent to tell me which trading mistake i keep paying for “read my trade journal, find a mistake i keep repeating, and show me examples from my trades” that would be my entry for the @MEXC CLI Prompt Challenge you have until oct 8 to submit yours for a chance at 1,000 USDT and a spot on stage at TOKEN2049frozen22d ago4.5K94050023.27%
opus costs 22x more here so i’d be picky about what i give it artificial analysis reports these scores and costs per benchmark task → deepseek v4.1 flash max: 39 / $0.27 → opus 5.5 low: 42 / $0.55 → opus 5.5 max: 58 / $5.98 i’d try deepseek for finding files, extracting details and preparing drafts, then bring in opus for harder decisions the thing i’d track is how much needs fixing afterward if those everyday tasks hold up, i’d have fewer reasons to pay opus prices for them13d ago4.4K88043033.02%
gm 🌸 today’s plan is to try a few new tools from ai startups if you’re building one drop it below https://t.co/lvTcIWlYvifrozen1mo ago4.2K2010157148.71%
202 agents and 184 skills in a single github repo i’d check these five before building my next ai agent ➡️ something you can run anthropic’s claude-quickstarts has example apps for customer support, browser automation and scheduled sentry triage https://t.co/w2TwCtvCJm ➡️ better coding habits addy osmani’s 25 skills cover requirements, test-driven development, browser testing and security reviews https://t.co/cWrPWKLoWV ➡️ a specialist for the task wshobson’s collection contains those 202 agent definitions and 184 skills across a catalog of 94 plugins https://t.co/7aUK1zZbD4 ➡️ agents connected to your work anthropic’s knowledge-work-plugins bundles role-specific skills and connectors for sales, support, product, finance and more https://t.co/m87w9revPl ➡️ instructions for different roles agency-agents has role definitions you can adapt for engineering, design, marketing and testing https://t.co/cQJ2RYJQIs my starting point would be one quickstart and one task i can check the code and instructions are available to inspect before deciding what to build yourself12d ago4.1K56310181.92%
full notes app, empty draft i’d rather have claude pull from my old drafts and rough notes than give me another 100 ideas find what helps with the post i’m writing, show me the original note, and let me decide what to use https://t.co/Sug4NquIFhfrozen15d ago3.9K62032002.38%
gm 🌸 have a great day ✨ https://t.co/tMFcXAxqiofrozen24d ago3.5K20821620310.56%
claude code got a second set of eyes its new “you should know” mod can flag things you or claude might miss a desktop page isn’t finished if i asked for mobile too https://t.co/jtytUi9Ot0frozen20d ago3.5K77143133.57%
i made this motion video with gpt-6 astra it started with one of my posts and a reference video i want to make more of these for ai products that need a better explanationfrozen24d ago3.5K94053044.35%
Jev handled 7 classifications across 1,000 emails in 6 seconds for $0.09 a lot of creator work does not need deeper reasoning it needs a fast first decision → remove ideas i already posted → score raw clips before editing → rank brand requests by relevance → surface comments that need a reply → detect buying signals across social posts → send uncertain decisions back to me GPT-5.6 Luna took about 5 minutes and $0.62 on the same batch i do not need a large model for every yes or nofrozen29d ago3.3K56143143.14%
before trusting ai’s code, ask it to break it “find an input that makes this fail and write a test for it” then run the test11d ago3.2K75040073.76%
anyone can look experienced now ai will fill every feed with people who sound like they have done this for years some of them opened the tool five minutes before posting before i trust the post i want to see what they have done with the tool https://t.co/ecBiXYngOQfrozen25d ago2.4K65052155.08%
“i’ll fix it when i get to my laptop” is becoming optional openai’s new dots update lets you create a dot on your phone and have it start or follow up on work in codex i’d use this when i spot something broken on a site i’m building send the screenshot, explain what should happen and ask for a tested fix to review i’d want the mobile preview back too13d ago1.6K67032136.25%
i was in Istanbul all last week veni vidi vici https://t.co/BVyzB8148Ofrozen17d ago1.4K93131058.97%
i’d put practical ai training into these programs help people improve a job application, practise an interview and build something they can show an employer give them a useful way to start11d ago64130025118.89%