← Back to leaderboard
rohit4verse

Rohit

Verifieddormant⚠ low-confidence
@rohit4verse

Engineer who builds, solves, and ships | FullStack + Applied AI | Agentic AI |

RosterAI Agents & Productivity ToolsAI Model News & Releases
24,330 followers · 538 following · Last polled 2h ago
Perf. Score
48
Rank #249
Value Score
6
#261 of 273 priced
Median views
1.4K
p25 1.4K · trailing 7d
Followers
24K
-0.1% 7d
ER (impr.)
3.36%
eng ÷ impressions
ER (foll.)
0.19%
eng ÷ followers
Est. CPM
$583.94
QT-rate basis
WoW views
-12.8%
flat

Rates & value · est. from 1.4K median organic views

overpriced
Quote tweetbasis
$800
$583.94 / 1K views
Post
—
no CPM
Thread
—
no CPM
Retweet
—
no CPM
Value Score 6 · rank #261 of 273 priced$17391.3 / 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 (1.4K), not the mean (1.4K) — so a single viral post doesn't inflate a creator's reach. The p25 floor (1.4K) is the level 75% of posts clear.

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

Recent posts · ranked by views

PostPostedViewsLikesRepostsRepliesQuotesBookmarksER
if you're still sitting in front of your laptop waiting for your agent to finish, then you're the bottleneck. for an extra $20 a month, you get a night shift engineer that ships while you sleep. here's mine: → devin pro, $20/mo. runs on its own cloud computer, browser included → sign in with chatgpt. gpt-6 sol bills to the plan you already pay for → swe-2 is free on it right now → kimi k3 (2.8 trillion params) runs locally off my ssd as the qa agent devin builds in the cloud. kimi checks it locally. i wake up to a reviewed pr. that's how a 100x engineer moves.frozen21d ago2.5K324130262.96%
nobody taught this robot football. it played itself for 140 years. every opponent was a previous version of itself, each one tougher than the last. then skild took it out of the sim and put it on a real pitch against a human. AlphaGo did this to a board game. now it has legs.frozen29d ago2.2K
29
2
11
0
13
2.53%
spend $100k on the higgsfield api, get $100k back... i don't get why every ai startup isn't building on this yet. 100% of your paid usage comes back as api credits, every model. up to $100k per business, from a $20m pool, first come first served. what i'd build on it: - ugc ads on seedance 2.5 + face references (public api, us) - a short-series studio on cinema studio - an agent that renders 20 genjutsu variants, keeps the best - one product photo in, a 10s video out at $0.163/sec that clip costs $1.63. the cashback pays for the next one. credits must be earned and used by sept 30. price sheet below:frozen28d ago1.9K261120102.58%
your decision model is making confident calls on data it never saw. most decision models chunk anything long. i keep running into it on jev. give it a 100-page contract and the liability clause on page 2 can be gone by page 50. the dangerous part isn't that the model forgot. it's that it still answers like it didn't. subq snap takes 384k tokens in one call, no chunking. 99.8 on decision ruler at 256k. jev, chunked, gets 90.7. and it's $0.05 per million input tokens. output is free. let snap handle the structured calls. let your llm handle the open-ended ones. stop paying frontier prices to ask if a ticket is urgent.frozen10d ago1.8K25480163.01%
ai voices sound pretty damn good now. but editing them is still a pain. you get a perfect take, except for one word. fix that word, and suddenly the whole line sounds different. now you're regenerating it again, trying to get back to the original take. fish audio's drama 3 lets you fix that one word without touching the rest. and instead of messing around with [whispers] and other tags, you can just tell it how you want something said. like you're talking to an actual voice actor. that's the kind of control ai voice has been missing.frozen9d ago1.7K28490113.11%
every department in your company wants ai. none of them need a frontier model. i turned @av1dlive's inference guide into a 30-second film. one support ticket, from enter to the last word. it shares a gpu with strangers and survives a dead worker before you see a single word. watch it once and you'll know what inference engineers actually do all day. 2026 is the first year the world spends more running models than training them. $23.3b vs $19b, per gartner, driven by fine-tuned models running inside company systems. nobody wants to pay a billion dollars in tokens for work a model made for their stack can do. but that model is only cheap if someone knows how to serve it, and anthropic lists those roles at up to $850k. av1d's guide is the path to becoming that person. it starts with a 0.6b model on your mac: → a gpu kernel you wrote yourself, by month two → killing a worker mid-answer, on purpose → a benchmark that shows where your own change lost the model is free now. serving it is the job.frozen13d ago1.5K23280163.35%
the prompt box had a good run. sabi just raised $50m to replace it with a baseball cap. you think it. the cap writes it. no keyboard, no voice, no surgery. → the world's first chip built for consumer bci, designed in-house, fabbed by tsmc → non-contact sensors that read your brain through your hair → its own brain foundation model, trained on neural data it collects → already predicts your next 3-4 keystrokes before you type them your agent can ship an app overnight. it can't tape out a chip.frozen13d ago1.4K254110113.67%
you know exactly what you want to build. you can picture every detail in your head. > then you open a coding agent, spend hours trying to explain it. > you rewrite the prompt, tweak it, try again. > and somehow it's still not what you imagined. everyone says taste is the most important skill in the age of ai. but what good is taste if you can't get it out of your head? sabi is building a cap reads what you mean straight from your brain signals. no typing required. it's still early, but imagine where this could go. what if the hardest problem of building wasn't explaining what you wanted?frozen5d ago1.4K2837083.36%