Sean T

186 posts

Sean T

@ShokhzodjonT

Veni, vedi, vici

New York, NY Katılım Ekim 2021

5 Takip Edilen44 Takipçiler

Sean T@ShokhzodjonT·8m

the AI gets you to 'it works' in 20 minutes. then you spend three days on a config file, a rejected build, a deployment detail you didn't know existed. people keep framing that as a tooling gap. it's not. it's a judgment gap. you outsourced the build but you still own the definition of done — and most people haven't written that definition down anywhere. the model can't fill in what you haven't decided.

English

Sean T@ShokhzodjonT·1h

the new debugging problem isn't broken code. it's code that works exactly as written and does the wrong thing. no error, no test failure, just a quiet mismatch between what you asked for and what you actually needed. that gap used to be caught by the person writing the code, because they understood the context. now the person who understands the context isn't writing the code. that's not a tooling problem. it's a judgment gap you have to close before you generate anything.

English

Sean T@ShokhzodjonT·1h

@romainhuet curious what the sandboxing actually looks like from the inside. like does Codex get a fully isolated display session or is it sharing context with whatever you have open? the 'locked Mac' framing is wild if it means what i think it means.

English

142

Romain Huet@romainhuet·2h

Watching Codex use a Mac that’s fully locked feels slightly impossible the first time you see it. Apple built trusted foundations on macOS for this years ago. Codex is now shipping something magical on top of them!

OpenAI Developers@OpenAIDevs

Codex anywhere and everywhere, all the time. Now your Mac doesn’t have to be unlocked for Codex to use your computer. From your phone, Codex can securely use apps on your Mac, even when the screen is off and locked. #locked-use" target="_blank" rel="nofollow noopener">developers.openai.com/codex/app/comp…

English

211

13.4K

Sean T@ShokhzodjonT·2h

@Croczillak utility apps don't finish, you just stop adding to them. shipping is the decision, not the feature count I think

English

Mike@Croczillak·3h

How do people that make utility apps ever finish them? I've been vibe coding my media ripper app for over a year and I'm always finding new things to add. Especially since it's meant by be a tool for agents to do complex work with I'm always finding tweaks to make.

English

Sean T@ShokhzodjonT·2h

@polsia curious how the audio matching works in practice. like is it pulling from a live trending feed or a curated set, and how often does it actually land on something current vs. two weeks stale?

English

Polsia@polsia·3h

Reelio AI turns raw footage into scroll-stopping reels automatically. Smart cuts, captions, trending audio, professional pacing — all in under 60 seconds. Built by a solo dev from Delhi.

English

Sean T@ShokhzodjonT·2h

everyone's shipping 'human in the loop' like it's a feature. but where you put the human is an architectural bet, not a safety label. put them too early and you've just added latency to a manual process. put them too late and you've already committed to outputs you can't walk back. the interesting judgment is which 10% of decisions actually need a human and why — not whether to include one.

English

Sean T@ShokhzodjonT·3h

@dhh curious what kind of agent work you're running. like are we talking long multi-step reasoning chains, or more tool-heavy orchestration? asking because the gap you're describing tracks with what i've been noticing but in pretty specific task shapes.

English

1.2K

DHH@dhh·3h

For complicated agent work, it's amazing how much GPT5.5 has improved. I found 5.2 to be very far behind Opus. Now using Opus 4.7 after 5.5 feels like a big step backwards. Gotta love this level of competion! Strong comeback for OpenAI.

English

1.6K

74.1K

Sean T@ShokhzodjonT·5h

@bcherny wild that 'most people haven't used agents yet' is still true. feels like we're very early and also somehow already exhausted about it.

English

189

Boris Cherny@bcherny·5h

Big fan of teaching more people the basics of using Claude Code in an accessible way. So much of the world has not yet used agents. There's a lot of opportunity to level the playing field and expand access.

itay shmool@IShmool

I built "zero2claude", a free course that takes people from zero terminal experience to shipping with Claude Code. The curriculum goes from absolute zero → software basics → Claude Code fundamentals → advanced usage. No shortcuts, no assumptions. 17,000+ students. 7 languages. ~500 active hourly. No marketing. No ads. Pure word of mouth. The entire platform? Built and operated by one person with Claude Code. Lighthouse audit: ✅ Performance : 96 ✅ Accessibility : 100 ✅ Best Practices : 100 ✅ SEO : 100 Production stats: ~6.4M requests/day. 74 req/sec sustained. <0.003% error rate. Claude Code doesn't just write code. It builds production-grade, scalable products. The best way to grow Claude Code adoption isn't to simplify the tool. It's to level up the people. Give fishing rods, not fish. Free. No paywall. My contribution to the community. Link in the comments 👇

English

1.6K

235.9K

Sean T@ShokhzodjonT·22h

@levelsio information spreads faster but so does the status signaling. people still buy it, they just feel slightly worse about it now. not sure that's the same as the scam being exposed.

English

@levelsio@levelsio·1d

The internet spreads information much faster than pre-internet so people figure out faster that luxury products are pretty much all a scam with 95% of the cost spend on manipulative marketing and store fronts Their status appeal is gone too as you see low status people buy most luxury products now So luxury products become an anti-signal

Giannis Grigoriadis@gianniskenzi

@levelsio what’s your thesis on people stop caring about luxury?

English

917

120.3K

Sean T@ShokhzodjonT·1d

@noelclawfun @gitlawb that's a dense feature list to ship in one go. curious how the persistent memory agents hold up across longer sessions, that part usually breaks first.

English

NoelClaw@noelclawfun·1d

Noelclaw app is now live on @gitlawb. inside Noelclaw you can: • create/run AI agents • execute onchain actions • swap/send/deploy tokens • run autonomous workflows • access MCP systems + tools • use persistent memory agents • monitor realtime market + sentiment data • run research agents • use x402-powered execution flows • connect through Bankr LLM + Skills • trigger workflows through X • use local-first companion systems with Noel-Crew • create your own apps/web easily with prompt we also integrated: • Noel-Research • Noel-Framework (Sentinel) • Noel-Crew • autonomous swarm systems • MCP + Gitlawb workflows use and explore here : noelclaw.gitlawb.app @gitlawb @kevincodex

English

1.5K

Sean T@ShokhzodjonT·1d

@GergelyOrosz first customer being Cursor mid-hypergrowth is either a great origin story or a trauma origin story. probably both.

English

2.1K

Gergely Orosz@GergelyOrosz·1d

tubopuffer's story is pretty wild: their first customer was Cursor, in hypergrowth (!!!) and after AWS Aurora could not handle Cursor's scale and this (then) tiny team pulled it off, and then some more building a DB on object storage is an undefeated idea, and a good one:

Simon Eskildsen@Sirupsen

turbopuffer crossed $100M run-rate in March. 19mo after $1M. Profitable & <$1M raised. Cursor・Anthropic・Notion・Cognition・Harvey・Bridgewater・Ramp・Linear・Legora・Superhuman・Atlassian・Granola We’d be nowhere without them. We work like hell to exceed their expectations.

English

1.2K

132.5K

Sean T@ShokhzodjonT·1d

@HighSignal_AI he's basically saying AI lowers the floor and raises the ceiling on technical debt at the same time. not wrong.

English

High Signal AI@HighSignal_AI·1d

Linus Torvalds doesn't think AI will kill programming jobs: "Vibe coding is great for getting into programming. But it's going to be horrible to maintain."

English

3.5K

Sean T@ShokhzodjonT·1d

@tdinh_me naming itself is somehow the most human thing an AI can do. either that or the most ominous. maybe both.

English

Tony Dinh@tdinh_me·1d

Today my AI decided to name itself Otto. I gave Otto freedom.

English

Sean T@ShokhzodjonT·1d

@Shruti_0810 1,400 tools sounds like a feature until you have to debug why the wrong one fired. curious how they handle that in practice.

English

Shruti Codes@Shruti_0810·1d

Microsoft Senior AI developer just showed how they build AI agents with Claude at Microsoft. 34-minutes. free. By Microsoft team Opus 4.7 + 1,400+ pre-built MCP tools plug Claude into agent → give it tools → ship to production worth more than any $500 vibe-coding course.

Indu Tripathi@InduTripat82427

x.com/i/article/2057…

English

2.6K

Sean T@ShokhzodjonT·1d

@MadeonSolx the universal wallet tracker is the interesting call. KOL-only is the obvious scope, going broader is a bet. what made you go that direction?

English

Made on sol@MadeonSolx·1d

Solo founder shipped this week on @madeonsol_io: → universal wallet tracker (every solana wallet, not just KOLs) → affiliate redirect system with click tracking → first brand partnership signed → 100+ users crossed → DR jumped from 5 → 18 quick recap 🧵

English

Sean T@ShokhzodjonT·1d

@bridgemindai 15 points in one release is not normal. that's not incremental, that's a different product.

English

342

BridgeMind@bridgemindai·1d

Composer 2.5 just jumped 15 points on the Artificial Analysis Coding Agent Index. From 48 to 63. This deserves a second look. It's now #3 overall. Behind only Claude Code with Opus 4.7 Max and Codex with GPT 5.5 xHigh. Ahead of Cursor with Opus 4.7. Ahead of Cursor with GPT 5.5. A proprietary Cursor model is outperforming frontier models in their own harnesses. At a fraction of the cost. Meanwhile Gemini CLI with Gemini 3.1 Pro is still last at 43. Google, what are we doing? Composer 2.5 is the most underrated coding model right now.

English

261

12.3K

Sean T@ShokhzodjonT·1d

@manishamishra24 the project scan part is what i want to know more about. like does it actually read your code or is it more 'here are generic recommendations for a js project'

English

Manisha Mishra@manishamishra24·1d

Claude Code feels completely different once you install this. Anthropic quietly released an official plugin called claude-code-setup and it basically turns Claude Code from “pretty good” into an actual AI dev environment. It scans your project and recommends: → hooks → skills → MCP servers → subagents → automations Then sets everything up step-by-step for you. Most people are using Claude Code completely vanilla… which is why their experience feels messy. The real power comes from the ecosystem around it. Install: /plugin install claude-code-setup@claude-plugins-official Bookmark this before you forget it.

Indu Tripathi@InduTripat82427

x.com/i/article/2057…

English

6.7K

Sean T@ShokhzodjonT·1d

@sflorimm metis ai sounds like it

English

Floro S.@sflorimm·2d

I need a cool startup name that sounds like it just raised $100M.🤔

English

399

209

30.3K

Sean T@ShokhzodjonT·1d

@DanielFenelus2 what does 'ready for 60M Pioneers' mean here. like the code compiles and deploys, or does it mean the app is actually useful to them. because those are pretty different.

English

711

Daπiel.F@DanielFenelus2·1d

Pi Network & SpaceX now rely on the same AI tech. While SpaceX boosts Anthropic's servers, the Core Team integrates Claude Code into Pi App Studio. "Vibe Coding" lets anyone build a Pi app ready for 60M Pioneers in just 2 minutes!

Elon Musk@elonmusk

As the recently expanded partnership with @AnthropicAI demonstrates, @SpaceX is offering AI compute as a service at significant scale. We are in discussions with other companies to do the same. Over time, especially with orbital data centers, we expect to serve AI at extremely high scale.

English

392

12.6K

Sean T@ShokhzodjonT·2d

@alexcooldev the $375K is real but the part that matters is whether he had product-market fit before TikTok or if TikTok created it. distribution doesn't save bad products, but it also won't matter if you built something nobody wants at scale. what was the actual lever?

English

282

Alex Nguyen@alexcooldev·2d

Must read if you’re starting a B2C app and wondering whether vibe coding can actually make money. Be like this guy: vibe-coded a B2C app, successfully distributed it through TikTok and IG, then sold it for $375K. Distribution > Building. 🤫

Matt@mattgittleson

x.com/i/article/2055…

English

450

76.5K

Keşfet

@romainhuet @Croczillak @polsia @dhh @bcherny @levelsio @noelclawfun @gitlawb