max | entrepreneur | ∞/21M ⚡ @freinode

This is becoming my favorite way to read Twitter. birdclaw.sh

Welcome to the actual video. This is the opening keynote of Compile 2026. This is by Michael the CEO of Cursor. Kevin comes in at approximately 13 minutes in so make sure you check the full video out youtu.be/fWa7uxyhVDE?si…

Spacex just bought Cursor for $60 billion. the same day, Kevin Niparko ( Product Lead at Cursor) showed exactly what Elon bet on: "we're moving towards an era of loops and agent systems where agents are going to continue to be working on our behalf at all times" 13 minutes that laid out the next five years of how software gets built: agent swarms + loops. i wrote the full guide on building agent swarms two weeks ago. bookmark it before you need it.

Avid @Av1dlive

3 weeks ago

x.com/i/article/2062…

26 81 439 338K 1K

I think this is bigger than most people realise. The first era was humans using software. The second era was humans using AI. The third era is humans working alongside AI. Not as a tool. Not as a website. Not even as an app. As a persistent teammate with memory, context, tools and agency. The organisation chart of the future may include humans, AI agents and teams of both working together.

@karpathy Reminds me of this sketch I drew up a while back

After 2 days between conferences I finally figured out some sort of autonomous flow for my coding flow. I'm a business guy, so my agent autonomy for biz stuff is pretty cool for engineering... meh still stuck on baby sitting codex/cursor. So here we go... I'm running this across 5 projects simultaneously right now: Orchestrator is Openclaw/Hermes agent on discord. 1) Spec generation - Use Grill with docs skill to create a detailed spec - Chain a few skills into one to convert the spec into discrete issues - Load them into Linear 2) Autonomous execution - Queue a cursor/codex agent via CLI to pick up the issues and work through all of them in goal mode - Quality gates (every pass) - Run Peter's autoreview - Run Cursors thermonuclear review - Call the orchestrator with original session injected via CLI for final testing. 3) The Orchestrator Loop - Runs every 30 min checking on the coding CLI - Ingests the orchestrator session as context - Watches Cursor: if there's a prompt → drive it; if there's an error → solve it - Keeps the work moving without manual babysitting - Don't bother me at all

Henry Mascot @iAmHenryMascot

3 days ago

My move when I'm about to run out of time to use my cursor credits! connect all my projects from other harnesses EXTRA HIGH FAST! scorch earth 🤣 😂😂😂 Thanks to @mattpocockuk's grill with docs spec-ed out a couple of weeks of roadmap then @SuperAda___ and co

0 0 2 501 1

Maxxxx em tokens 🔥🔥🔥

Henry Mascot @iAmHenryMascot

2 days ago

After 2 days between conferences I finally figured out some sort of autonomous flow for my coding flow. I'm a business guy, so my agent autonomy for biz stuff is pretty cool for engineering... meh still stuck on baby sitting codex/cursor. So here we go... I'm running this

0 0 4 914 0

Introducing Claude Tag, a new way for teams to work with Claude. In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.

This is how my team at @curacel has been using @openclaw since Jan. Except that they aren’t called Claude, they have personality and different names for different teams depending on their roles. The one for engineering has a different personality, skills, access, crons for their team. This is the future of agent use, not on cowork or some ui. But talking to them in your workspace. With some of the latest OC features they can even join calls and you talk to them like humans. OpenAI will launch theirs too, Google will add this feature to Spark, everyone will. MS already launched theirs 3 weeks ago called MS Scouts powered by OC. @steipete had incredible foresight in creating OC. I don’t think you should use these if you want your sovereignty. I encourage you and your team to use an opensource version like OC or Hermes might be a bit more technical to manage, but you own your data and can use any model you want and not locked into one platform, the switch cost will be 10x more expensive than switching SaaS.

Claude @claudeai

23 hours ago

Introducing Claude Tag, a new way for teams to work with Claude. In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.

1K 2K 24K 13.8M 11K

This is a new paradigm for interacting with Claude that is significantly more "inline" with all the other human activity org-wide. Once you do all of the under the hood engineering work to make this "just work" (e.g. across tools, integrations, compute environments, memory, security, etc.), Claude basically joins the team in a seamless way - you can talk to it as you would talk to a person and it can help with a very large variety of workloads. Imo this is the 3rd major redesign of LLM UIUX. The first paradigm was that the LLM is a website you go to, the second was that it is an app you download to your computer. This third one is that it is a self-contained, persistent, asynchronous entity with org-wide tools and context, working alongside teams of humans. It really takes a while to wrap your head around it, but it works and it is awesome.

Claude @claudeai

23 hours ago

Introducing Claude Tag, a new way for teams to work with Claude. In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.

1K 2K 24K 13.8M 11K

PRÄSIDENT PUTINS BOTSCHAFT AN DIE DEUTSCHE REGIERUNG UND DAS DEUTSCHE VOLK So wird der Gnadenstoß versetzt – ohne Waffen, verbal… 🥳 „Wir wollen euch nicht angreifen! Warum sollten wir? Diese Zeiten sind längst vorbei! Jeder, der noch bei Verstand ist und klar denken kann, versteht das. Erstens: Ihr habt bereits Staatsschulden in Höhe von 2,5 Billionen Euro, und kein seriöser Wirtschaftswissenschaftler hat eine Ahnung, wie ihr die jemals zurückzahlen wollt. Und jetzt wollt ihr weitere 1 Billion Euro aufnehmen, um euch gegen uns zu bewaffnen. Wollt ihr, dass das russische Volk diese Schulden bezahlt? Niemals! Zweitens: Euer Land ist voller Millionen von Migranten, die euch 50 Milliarden Euro im Jahr kosten. Sollte das russische Volk dafür zur Rechenschaft gezogen werden? Drittens: Ein beträchtlicher Teil eurer Bevölkerung ist so verrückt, dass er glaubt, das Klima durch Fahrradfahren und Insektenessen beeinflussen zu können. Vielleicht ließe sich dieser massive Hirnschaden beheben, aber das würde uns auch etwas kosten.“ viel. Viertens: – Euer Bildungssystem war einst vorbildlich. Jetzt findet in vielen Klassen praktisch kein Unterricht mehr statt, weil fast niemand mehr Deutsch spricht. Fünftens: – Eure Infrastruktur verfällt, und ihr kommt mit den Reparaturen nicht voran. Sechstens: – Eure Eisenbahnen waren einst der ganze Stolz der Welt. Jetzt fahren eure Züge wie in Indien. Siebtens: – Wir brauchen eure berühmten Ingenieure nicht. Während der Sanktionen haben wir gelernt, dass wir ohne sie auskommen können. Sollten wir sie aber doch brauchen, wenden wir uns an China. Dort sind sie nicht nur billiger, sondern auch besser. Achtens: – Ihr habt weder Rohstoffe noch Energiequellen. Warum sollten wir also euer Land erobern? Um Probleme zu lösen, die wir sonst gar nicht hätten? Realistisch betrachtet: Selbst wenn ihr uns rufen, kapitulieren und weiße Fahnen hissen würdet, würden wir trotzdem nicht kommen!

x.com/i/article/2064…

A senior Google engineer just dropped a 19-page PDF on "Loop Engineering" for LLM and agentic systems. Act → Observe → Learn → Repeat • Act: the LLM proposes a code transformation (tile this loop, parallelize that one). • Observe: a compiler runs it and reports back - is it valid? faster? slower? by how much? • Learn: the LLM reads that feedback and adjusts its next move. • Repeat until it stops finding improvements. The agent gets smarter purely from grounded feedback inside its own context window. This 19-page PDF totally changed the way I’m building agentic systems today. Read it now, then explore the article below.

Codez @0xCodez

2 weeks ago

x.com/i/article/2064…

100 832 5K 5.0M 17K

I've dictated almost everything for 6 months with Wispr Flow. 44,414 words, 161 wpm, top 0.1% of users. Last week I tried FluidVoice. Open source, runs local on my Mac, corrects as I speak with no API key, and handles slang better than I expected. Cancelled my paid plan. If you're on a Mac, this one's for you: altic.dev/fluid @ALTIC_DEV

@Raytar the recordings are literally published m.youtube.com/playlist?list=…

Met a guy making $1.6 million a year. Three days ago he was at a Meta conference. Told me he saw the best AI talk of his life. Boris Cherny was on stage. Showed how the Anthropic team actually uses Claude day to day. Boris deleted his IDE eight months ago. Now he codes from his phone. I watched it last night. Had to pause it twice. Not because it was hard. Because I realized I've been using Claude like a toy. He sent me the recording. It was never published. Posting it below.

Codez @0xCodez

2 weeks ago

x.com/i/article/2064…

100 832 5K 5.0M 17K

Anthropic engineer: "At Anthropic, 90% of our engineers use loops and ‘dreaming’ to build self-improving agentic systems. сlose the agent loop. give an agent a way to verify its own output." in a 30-minute session, an Anthropic team member explains how to build an agent that improves itself. Claude + loops + dreaming + CLAUDE.md - that’s the secret. Watch the talk, then save the playbook below.

Movez @0xMovez

2 days ago

Everyone's still optimizing prompts. The real moat is loop engineering. A prompt gets one good answer. A loop holds a 40-minute call together while the human wanders and changes their mind. @usebland just raised $100M building exactly that. The difference: • most voice AI

17 5 118 465K 106

APPLE JUST SHOWED HOW TO RUN 10 AI AGENTS LOCALLY ON MAC - NO CLOUD, NO API KEYS, ZERO COSTS 00:10 Apple engineer says: "your data stays on your device, AI available anywhere at any time, zero usage costs" agent reads your code, checks GitHub, finds what needs attention and writes a report - all on your Mac, nothing goes to the internet 10 agents work simultaneously - one writes code, another tests, third fixes bugs - parallel with no queue built a full iPad app from scratch in 2 minutes, fixed its own mistakes and compiled without issues takes 5 minutes to set up - and you never pay again for a bot that runs 24/7

Noisy @noisyb0y1

5 days ago

x.com/i/article/2067…

32 61 399 1.8M 1K