Dmitry Belaventsev (@dizpers) header
Dmitry Belaventsev (@dizpers) profile picture

Write the People, Talk with Code 17 yrs SWE → trading → AI agents Polystrike founder · building in public

Joined April 2010
Photo by Dmitry Belaventsev (@dizpers)Photo by Dmitry Belaventsev (@dizpers)Photo by Dmitry Belaventsev (@dizpers)Photo by Dmitry Belaventsev (@dizpers)Photo by Dmitry Belaventsev (@dizpers)Photo by Dmitry Belaventsev (@dizpers)Photo by Dmitry Belaventsev (@dizpers)Photo by Dmitry Belaventsev (@dizpers)Photo by Dmitry Belaventsev (@dizpers)Photo by Dmitry Belaventsev (@dizpers)
Damn Elon moves quick 👀👀👀👀😂😂
569
2,559
649
31,544
8,256,052
Dmitry Belaventsev retweeted
Which Claude model should you build with? Our new guide breaks down when to use Fable 5.2, Opus 5.5, Sonnet or Haiku. It also covers when to adjust effort instead of switching models, and how to save cost by putting subagents on smaller models. claude.dev/blog/which-claude…
372
1,222
305
16,210
2,606,348
Dmitry Belaventsev retweeted
This is incredibly bad news. Abdul El-Sayed wants to pause AI, while at the same time, almost every AI data center is being blocked by local communities. Folks fail to realize just how transformative AI is for humanity. It’s a shame.
483
182
122
3,080
454,750
Dmitry Belaventsev retweeted
The creator of Naruto @TheKishimoto was writing in public! Every few pages in the manga, he shares a little story: a challenge he faced while drawing a character, a sketch of his desk, or even a story about his childhood. It’s so interesting. It builds a connection between us. I think every creator (writers, actors, entrepreneurs) should embed their DNA in their work. AI can’t replace that.
Re-reading Naruto as an adult feels like being 12 again for a little while.
112
4
8
290
34,459
Dmitry Belaventsev retweeted
New in Claude Code: /checkup Run /checkup to: 1. Clean up unused skills/MCPs/plugins and save context 2. Dedup your local CLAUDE.md against the checked in CLAUDE.md 3. Break up root CLAUDE.md into nested CLAUDE.md's + skills 4. Turn off slow hooks 5. Update your Claude Code to the latest version 6. Enable auto mode by default 7. Pre-approve frequently denied read-only commands .. And a few other goodies. /checkup confirms with you before making any changes. Enjoy!
343
956
192
11,583
876,169
Dmitry Belaventsev retweeted
"Fable 5 is amazing. I gave it browser access and it filled the form for me." A message I got from a non-technical friend. Labs should hide model selection from non-technical users and charge for units of work completed. AI should decide which model to use to complete the task.
24
5
4
182
21,629
Dmitry Belaventsev retweeted
My first real investment exit!!! 🎉 Congrats @cursor_ai and everyone!!!
BREAKING: SpaceX is acquiring Cursor in a $60 billion all-stock deal. • Cursor is being valued at $60 billion • Cursor will become a wholly owned SpaceX subsidiary • Cursor shareholders will receive SpaceX Class A shares • The exchange ratio will be based on SpaceX’s 7-day average share price before closing • Subject to regulatory approvals and customary closing conditions • Expected to close in Q3 2026 Cursor is one of the world’s leading AI coding platforms and one of the fastest-growing software companies. This marks one of the largest AI acquisitions ever and significantly expands SpaceX’s footprint in AI.
180
115
47
9,171
1,681,770
Dmitry Belaventsev retweeted
Start using your tokens before we enter the caste system This is really really really important
Replying to @claudeai
Fable 5 is state-of-the-art on nearly all tested benchmarks, with exceptional performance in software engineering, knowledge work, scientific research, and vision. The longer and more complex the task, the larger Fable 5’s lead over our other models.
65
43
8
865
115,553
Dmitry Belaventsev retweeted
Heads up if you trade crypto or finance up/down markets through the API. On Friday Jun 5, we're changing how taker orders behave during the 250ms taker delay on these markets. Once your order clears initial validation and enters the delay window, it's locked in as pending until the delay finishes. What that means: - You can't cancel an order while it's in the delay window. Cancel requests come back rejected with "can't be canceled because it is pending/delayed." - The order's required balance is held during the delay, so other orders drawing on the same balance are checked against what's left until the delayed order clears. - Resubmitting the same order while it's in the window will be rejected. - Dropping the connection or canceling the client request after the order is in the delay won't stop it. It finishes processing once the delay completes. Cancellation works the same as before everywhere else, and a taker-delayed order is still cancelable if part of it rests on the book. Full details: docs.polymarket.com/concepts…
13
8
1
98
15,774
Dmitry Belaventsev retweeted
In fact - here's where it gets fishy - Polymarket SPECIFICALLY WAITED until the 8K was filed, to resolve the market. Their "sharps" could fill 6-figures against people betting on a confirmed outcome, then they could add a new rule and scam all the newcomers.
7
25
4
1,234
110,598
Dmitry Belaventsev retweeted
Earlier this week I spoke to @kevincodex about the @gitlawb project It was an interesting time bc the token had just dropped significantly but we mainly spoke about product I can say Gitlawb is one of the most ambitious projects in AI x crypto It’s basically a decentralized version of GitHub + Lovable + an open source Claude Code All 3 have some initial traction, and I’ll be interested to see how Gitlawb (the decentralized GitHub built for agents) does as staking and nodes go live Just sharing that after my interactions with Kevin I can say he appears to be a genuine builder in this space Disclaimer: I hold a small amount of the token because I want to support builders like him
27
60
21
294
42,352
Dmitry Belaventsev retweeted
WOW, 🤯 A leaked audio from Meta’s April 30 all-hands. Meta is reportedly using its own engineers’ work traces to train coding AI while cutting thousands of jobs. Here Zuckerberg arguing that models learn better when they watch “really smart people” perform tasks, meaning Meta’s internal code, tool use, clicks, and problem-solving can become higher-grade training data than contractor-written examples. The idea is behavior cloning: instead of only feeding an AI finished code, Meta can feed it the step-by-step path a strong engineer takes, including edits, tests, mistakes, fixes, and tool choices. That can teach a model not just what correct code looks like, but how a skilled developer moves from a vague task to a working solution. Meta is reportedly cutting about 8,000 jobs, roughly 10% of its workforce, and additionaly moving about 7,000 employees toward AI-focused work, so the hard realy is that human expertise is being turned into training data before some of those humans leave. The story is not fully independently verified, but the shift is happening for sure: tech companies no longer see AI as a tool sitting beside workers, but as a system that can absorb worker patterns and then compress them into software.
LEAKED AUDIO: In an all-hands meeting on April 30, Mark Zuckerberg tells employees that he's training AI on them ahead of mass layoffs. "The AI models learn from watching really smart people do things... The average intelligence of the people who are at this company is significantly higher than the average set of people that you can get to do tasks. So if we're trying to teach the models coding, for example, then having people internally build tools or solve tasks that help teach the model how to code, we think is going to dramatically increase our model's coding ability faster than what others in the industry have the capability to do, who don't have thousands and thousands of extremely strong engineers at their company."
16
19
5
107
32,109
Dmitry Belaventsev retweeted
THE WORLDS FIRST PROP FIRM FOR POLYMARKET IS HERE! If you believe you have the skill to trade @Polymarket but not the capital, we are here for you! Receive up to $200,000 in funded capital with @PolyFundr and prove your edge! ✨ We are giving away 5 free $200,000 accounts TONIGHT to random people who retweet and comment on this post. Good luck! polyfundr.com Your skill. Our capital.
587
329
29
776
99,284
Dmitry Belaventsev retweeted
Issue tracking is dead. We are building what comes next. linear.app/next
198
248
318
4,013
2,057,207
Dmitry Belaventsev retweeted
BREAKING: POLYMARKET ADDS FEES TO ALL MARKETS Taker rates will be as high as 1.8%
30
26
26
713
180,969
Dmitry Belaventsev retweeted
AI is gonna take your job and your girl.
1,551
2,474
707
25,541
3,130,776
Dmitry Belaventsev retweeted
Introducing the Readwise CLI. Anything you've saved in Readwise (highlights, articles, PDFs, books, youtube, newsletters) is now instantly accessible from the terminal. For you, and your AI agents. npm install -g @readwise/cli
62
107
59
1,167
284,367
Dmitry Belaventsev retweeted
Permisionless markets are coming to @Polymarket People think this announcment is related to $POLY and 1mn crypto markets But Mustafa literaly gave us the answer right there For those that didn't get it, permisionless markets mean *anyone* will be able to create a market on Polymarket. Creators will probably have to pay a fee to open a markert, but they will also be able to earn trading fees from it This would turn Polymarket from just a platform to the universal Prediction Market infrastructure layer. If this goes through, they are leaving glorified sportsbooks like K*lshi in the rear view mirror.
big announcement monday 😛🪙
56
17
9
295
56,032
Dmitry Belaventsev retweeted
Solana is 6 years old More money has been lost on this chain than in the entire 2008 housing crisis Happy birthday Solana 🎉 xitter.gq/jbondwagon/status/2033…
257
218
56
2,970
224,784
Dmitry Belaventsev retweeted
This is wild. 143 million people thought they were catching Pokémon. They were actually building one of the largest real-world visual datasets in AI history. Niantic just disclosed that photos and AR scans collected through Pokémon Go have produced a dataset of over 30 billion real-world images. The company is now using that data to power visual navigation AI for delivery robots. Players didn't just walk around with their phones. They scanned landmarks, storefronts, parks, and sidewalks from every angle, at every time of day, in lighting and weather conditions that staged photography would never capture. They documented the physical world at a scale no mapping company with a fleet of vehicles could have replicated on the same timeline or budget. Niantic collected this systematically, data point by data point, across eight years, while users thought the only thing at stake was catching a rare Charizard. The most valuable AI training datasets in the world aren't being assembled in data centers. They're being built by people who have no idea they're building them.
POKÉMON GO PLAYERS TRAINED 30 BILLION IMAGE AI MAP Niantic says photos and scans collected through Pokémon Go and its AR apps have produced a massive dataset of more than 30 billion real-world images. The company is now using that data to power visual navigation for delivery robots, letting them identify exact locations on city streets without relying on GPS. Source: NewsForce
2,121
23,041
2,535
103,186
14,090,634