onto my 19th build // thinking about - ai, distribution, and mental models // occasional contrarian // building @0xcapx
🇦🇪
Joined February 2019
- Tweets6.3K
- Following2.1K
- Followers5.6K
- Likes24.2K
Pinned Tweet
every personal website is the same website.
hero. three cards. a "lets connect" form.
i have an old crt television.
every page is a channel you can switch between:
- sh*t i ship
- essays
- a live tweet feed
- the lore (btw a playable game)
stay idle on the consult channel and a ufo abducts the cow.
22 projects. 25 essays. one game.
vibecoded end to end with claude.
walk into the builderness.
tyagi | capx.ai retweeted
here’s a short grok bot guide on how i create high quality bots and manage my team. ask @bot to read this post and help you improve your team!
• add dr eggbot to your team! he’s my bot designer that i use to create bots
• you don’t need a lot of bots. a few good bots can be more useful and easier to manage than a large team
• tell your bots to create skills for itself as you give it instructions on how to do tasks. this lets them do them more consistently even over long conversations. you can even ask it to create a private github repo for you to store all your skills if you want them preserved independently
• add connectors, use the vault, and give access to all the tools you use to do your work. you can also ask your bot to setup its own email and it can sign up for its own accounts for better separation
• thinking about all the work you do by hand, or steps in your workflows where you have to copy and paste information back to your bot. ask your bot how it can do that for you instead!
• imagine you hired an intern, how would you onboard them and help them do the work correctly and efficiently? introduce yourself to your bot and describe the work you do, then ask it how it can help you automate it
and my favorite tip: use voice mode or the transcription option! it’s a good way to just brain dump a lot of information. then, at the end, tell your bot to “restate this in your own words” so you can confirm it understood you
x.ai/bot/_jOdbfkB16zxu7MRcmR…
tyagi | capx.ai retweeted
Introducing /reanimate
Reverse engineer any animation from any website, video or gif.
reanimate.niazmorshed.dev/
tyagi | capx.ai retweeted
Today we’re launching the first AI modeling agency 💅 right inside Glif.
Cast hundreds of real AI models, drop your product, idea or brief and let Glif create super compelling content for you end to end.
If you still think AI models aren’t ready for real work, watch this 👇
🤖 Made with AI
tyagi | capx.ai retweeted
Selfishly using our upcoming 𝕏 web rewrite to build things I've always wanted. Here's a quality of life improvement: see how your profile will look to others before you hit save.
tyagi | capx.ai retweeted
In crypto, one of the few things you’ll be able to charge for is trust. And trust will continue to be expensive.
a long time ago having protocol fees was basically considered a nonstarter, there was some unspoken expectation that if your protocol became useful and implemented rent seeking the ecosystem would just fork you out. its kinda why the whole foundation/labs model came to be, one team for FOSS and another for profit of ex-FOSS contributors.
at some point this philosophy died and took the dead bodies of many great teams with it. but, they weren’t wrong. there wasn’t enough skill and bandwidth (obviously, in hindsight) and the opportunity cost of FOSS became too high. today, everyone now kinda has those resources and the ecosystems incentives structures have changed.
if you wanna do something cool and not-mid, find a useful protocol, fork away their rent seeking or requirement of some useless tokens use and ship as non upgradable public goods infrastructure. if you do it well, I think something will come out of it.
tyagi | capx.ai retweeted
nice..
""Claude Haiku 5.5 is priced 90% lower than Claude Haiku 4.5 for requests up to 100,000 tokens, and 50% lower for requests over 100,000 tokens."
tyagi | capx.ai retweeted
It's actually weird feeling to watch recursive self-improvement start working in your own harness, and to finally see the results.
With Opus 5.5 and GPT-6 Astra, our harness now improves itself and agents do almost every step of it.
1. Agents write the evals.
They turn failing call traces into versioned scenarios: caller persona, goal, the tools that should run, and the final state expected in the EHR.
2. Agents call agents.
Simulated callers are wired to phone our deployed receptionist in real rooms, with real tools and a sandbox EHR, in text and audio, across both of our runtimes.
3. Agents grade agents.
An LLM judge plus deterministic end-state checks score each call, and failures are clustered by runtime, tool and check. Agents then write one fix and one new scenario per root cause as a stacked PR.
4. Agents review agents.
Reviewer agents critique every PR, findings get fixed with tests, and scenarios are rerun. A change proposed only if it passes on both runtimes. We still own merge and deploy, for now.
We're early, but for the first time the loop is starting to close. I've never experienced a pace like this. It's getting faster so RAPIDLY.
Our internal data shows Claude is accelerating AI development—a possible path to recursive self-improvement, or AI autonomously building a more capable successor.
It’s happening faster than we thought, and the implications deserve greater attention. anthropic.com/institute/recu…
your agent knows
• what you are working on,
• who you are and
• how you talk.
WHO BETTER TO MANAGE YOUR - launches, announcements, & community updates
tyagi | capx.ai retweeted
your agent gets a short-lived session handle, not your X token.
every publish goes through server-side guardrails before the token is used.
hosted use still means trusting the server operator. you can also self-host.
the name was a placeholder to begin with.
but i kinda stuck with it. ngl sounds cool now.
let your AI run your X
i built capx-café to turn commits, fixes, and releases into X posts from your coding agent
draft. preview. schedule. publish.
your X token stays on the server.
an open-source MCP server
(what started as an internal tool for self, turned out to be something useful)
the workflow:
“read this week's commits. draft a build update in my voice. show me before publishing.”
your agent does the writing. capx-café handles the preview and publishing tools.
works with claude code, cursor, codex, and windsurf.
you can learn more at - capx-cafe.vercel.app/
github at - github.com/vb-tyagi/capx-caf…
tyagi | capx.ai retweeted
Capx Launchpad v2.0 is LIVE
Over the last two weeks, thousands of interactions across launches, traders, builders, and the market gave us one thing we couldn’t get from a whiteboard:
Real-world feedback!!
So we went back to the drawing board.
We studied how people were launching, how markets were behaving, where friction existed, and what needed to change.
Then we rebuilt Capx Launchpad from the ground up.
Welcome to Capx Launchpad v2.0.
How it works:
1. Launch your token on Solana.
2. Trading goes live instantly, paired with CAPX.
3. No presale. No waiting period.
4. Anyone can buy or sell from day one.
5. As more CAPX enters the pool, the token moves closer to graduation.
6. At 15,000 CAPX in net reserves, the token automatically graduates to Meteora.
7. Trading continues on Meteora, with the launch liquidity permanently locked.
One continuous journey:
Launch → Trade → Graduate → Open Market.
No presale.
No waiting.
No additional CAPX launch fee.
Built on Meteora.
Powered by CAPX.
Capx Launchpad v2.0 is live.
tyagi | capx.ai retweeted
Been looking into Casa, and the part that really caught my attention is how the company’s memory lives in the repo.
The agents can do the work, but keeping the context, decisions and progress there means you’re not starting from scratch every time.
That’s what makes this feel more practical to me. It’s not just about having AI agents it’s about giving them a system they can actually work from.
Introducing Capx Casa - the control plane for agent-run companies.
Casa turns company context into an ordered operating plan, routes work to specialist operators, reviews the output, and keeps the company brain current under human control.
Open-source and 100% free to use, modify, and build on.
It runs inside the agent CLI you already use, on your own plan - Claude Code / Codex / Grok Build / OpenClaw / Hermes / Gemini CLI / Cursor / OpenCode, you name it.
It sets up, controls, and manages your 'company brain': memory the founder can open, diff, and keep.
The company does not live in a context window.
It lives in a small set of Markdown, JSON, and JSONL files that preserve what is true, what changed, what remains blocked, and what the next session should do.