Yann

[Yann @kastelnik]

@AlfieJCarter Agents

[Alex Albert @alexalbert__]

We released Claude Opus 4.6 just two months ago. Today we're sharing some info on our new model, Claude Mythos Preview.

[Todd Saunders @toddsaunders]

The token cost to build a production feature is now lower than the meeting cost to discuss building that feature.

Let me rephrase.

It is literally cheaper to build the thing and see if it works than to have a 30 minute planning meeting about whether you should build it.

It’s wild when you think about it.

This completely inverts how you should run a software organization. The planning layer becomes the bottleneck because the building layer is essentially free. The cost of code has dropped to essentially 0.

The rational response is to eliminate planning for anything that can be tested empirically. Don’t debate whether a feature will work.

Just build it in 2 hours, measure it with a group of customers, and then decide to kill or keep it.

I saw a startup operating this way and their build velocity is up 20x. Decision quality is up because every decision is informed by a real prototype, not a slide deck and an expensive meeting.

We went from “move fast and break things” to “move fast and build everything.”

The planning industrial complex is dead.

Thank god.

[Yann @kastelnik]

This

[Romàn @romanbuildsaas]

The “problem” with AI agents is that I don’t enjoy doing anything anymore except working.

It feels like I’m inside a massive video game, and my character is leveling up at an exponential rate.

In 6 months, we’ve compressed what should have taken 4 years.

I wake up at 9am and I’m at my computer until 11pm.

I force myself to go to the gym. I force myself to be social.

But the only thing that truly excites me is progressing in the game.

Maybe it’s just a phase.

Isn’t the goal to build, exit, and then finally enjoy life?

But what if this is enjoying life?

What if building at full speed… is the reward?

[Adham Dannaway @AdhamDannaway]

Has this become the new design process? 🤖

What happens when AI improves?

Image credit Hang Xu

[Yann @kastelnik]

@JorgeCastilloPr Ok, need!

[Yann @kastelnik]

So Anthropic is building OpenClaw, pretty sure Sam saw that coming

[Claude @claudeai]

New in Cowork: scheduled tasks.

Claude can now complete recurring tasks at specific times automatically: a morning brief, weekly spreadsheet updates, Friday team presentations.

[Yann @kastelnik]

@noahzweben Damn, sounds magical if it can really manage multiple sessions seamlessly

[am.will @LLMJunky]

Anthropic just shipped Remote Control for Claude Code.

Start a session in your terminal. Pick it up from your phone.

This is the first official phone ↔ local machine integration from any major AI lab, but not the last. This will become widely adopted.

https://t.co/xujY5HGtX7

[Yann @kastelnik]

@steipete @OpenAI @openclaw Definition of speed-running, contrats!

[Matt Shumer @mattshumer_]

https://t.co/ivXRKXJvQg

[@levelsio @levelsio]

This doesn't work because a "product launch" isn't what it was 10 years ago anymore

Back then you'd launch on Product Hunt and you'd get thousands or tens of thousands of users overnight and journalists would pick it up after

A lot has changed since then

Firstly, nobody cares anymore, there's too many products and things launching and unless your product is completely groundbreaking and new, a launch won't get much attention anymore

Secondly, these days you essentially should be launching every day non-stop: you try get attention from potential new users, posting new features you build based on user requests, tapping into trends you see and then jumping on them

You even see it with AI companies now, they just add the new version like Grok 4.1 or ChatGPT 5.1 without a big presentation. Just roll it out and improve the product for users

So yes launching is dead I think

[Nikita Bier @nikitabier]

The only thing that matters in consumer software is velocity of shipping. You will regularly be wrong about what people want: no one is a perfect psychologist. But if you are out at the frontlines with concepts everyday, your team will have their finger on the pulse of users and know where to add fuel.

[Yann @kastelnik]

@calcsam Book 🙌

[Yann @kastelnik]

@civilianweb3 Yo :)

[TBPN @tbpn]

"You can break AI down into 5 tiers." - George Hotz

"Data centers - tier 1, fabs - tier 2, Nvidia/AMD - tier 3, OpenAI/Anthropic - tier 4, and completely worthless things like Cursor and Windsurf, which are tier 5."

"OpenAI and Anthropic will eat all the value from the Cursors and Windsurfs of the world. I argue that the tier 4s (OpenAI/Anthropic) aren't even going to have value."

From his June 2025 appearance, @realGeorgeHotz really raised eyebrows with this clip when it hit the timeline.

[Sam Altman @sama]

someday soon something smarter than the smartest person you know will be running on a device in your pocket, helping you with whatever you want.

this is a very remarkable thing.

[Jean de La Rochebrochard @2lr]

Until you're 10+ in a startup , the only roles that exist are head of get shit done & chief common fucking sense officer.

[Ash Lamb @ash_lmb]

One brick at a time they say...

[Elon Musk @elonmusk]

Wise words from a recent interview with @geoffreyhinton, one of the smartest people in the world regarding AI