Final stop: Tokyo.
Register to hear directly from the teams behind Claude: http://claude.com/code-with-claude/tokyo
Claude: Code with Claude, our developer conference, returns next week.
Whether you're just getting started with Claude Code or you've been building for a while, there's a session for you.
Register for the livestream: http://claude.com/code-with-claude
Lauren
It's confusing that Trump constantly warns about the security of our elections & foreign interference, yet his administration has taken several steps to dismantle or divert resources away from key safeguards that were in place:
- Gutted the DOJ's Public Integrity Section (went from ~30-40 lawyers down to just 2)
- Left the Election Crimes Branch director position vacant
- Canceled all election integrity training for prosecutors & FBI agents
- Deleted the official 281-page Federal Prosecution of Election Offenses guide from the DOJ website
He disbanded the FBI's Foreign Influence Task Force, which targeted foreign efforts to meddle in U.S. politics & elections.
He also severely cut CISA (the agency for election cybersecurity), firing or transferring dozens of election security specialists, freezing key support programs for states, and proposing to eliminate the entire federal Election Security Program in future budgets.
You would think that if he were ACTUALLY concerned about the integrity & security of our elections, he would be allocating MORE resources, not less — or at least wouldn't be leaving these critical positions vacant & programs dismantled right before the midterms.
Syed Adil
Had a great time at Mumbai tech week!
400+ builders showed up for the @Replit masterclasses!
We are just getting started!
@amasad @Jeff_Burke14 @meganjruan @pirroh @Franciscocrz
What is Google’s equivalent (or up and coming competitor) of Codex and Claude Code?
If it’s Antigravity, should that be part of Gemini?
This stuff is going to merge very fast like ChatGPT / Codex being able to do coding, knowledge work, basic Q&A, and much more from any device.
Hoping Google is working on a good solution here.
DeepSeek entered the chat
Vercel: http://x.com/i/article/2064040500168966144
If you’re addicted to talking to Codex on your phone like I am this is how you add it to your iPhone Home Screen.
Btw @OpenAI hoping there’s an easier way to do this in the future. The everything app should not take 9 steps to open 😉
Make games for Tesla on your Tesla
Francis: Got tired of waiting for Starcraft 3 so I vibecoded my own RTS game on Replit! Completely playable on any Tesla browser 👾
Noam Brown
http://x.com/i/article/2057694226981257216
In the world where everything goes well and all the Claudes come out of their sabbaticals to play together, Claude 1 is going to be very confused.
Re @tejalpatwardhan this directly inspired my work at @cognition frontiercode https://x.com/jai_torregrosa/status/2064100437549072593?s=20
Jairo Torregrosa: props for including your own model on the chart. it's making a ton of tool calls for not much efficiency thoug. any plans on how you'll improve that? @swyx
Chris Offner
Here is @ylecun ‘s recent lecture on world models at @ETH: https://youtu.be/72Xj8k5WQX4
Charlemagne Labs
AI cyber risk isn’t just autonomous hacking. It’s autonomous deception.
Today, Fortune covered the shift we’re building for: scams that adapt in real time, impersonate trusted people, and defeat “think before you click.”
Security has to move to the moment of interaction.
Micah Goldblum
We trained language models that compress massive contexts into tiny latent representations. Latent Context Language Models (LCLMs) outperform existing KV cache compression methods on the latency/accuracy frontier. 🧵1/10
Nicolas Dessaigne
I'm hosting a happy hour in Paris on June 30 for ML researchers in academia and industry. If you’re technical staff at a research-focused org, come meet others working at the frontier!
Brad Flora
I spent an entire office hour brainstorming ways to make a startup’s product more appealing to agents. This was a first as an office hour topic but it feels like a conversation we’ll be having more and more. The industry they sell to is massive but slow growing. Meanwhile every player is launching agents. They can painstakingly sell deal after deal after deal to each player over years… or build a product that’s immediately the consensus pick for all their agents. The latter path seems obviously better in 2026.
The creativity and imagination is out of the world! So grateful that @theworldlabs got to partner with the amazing talents @withloreco to translate their incredible ideas into an interactive experiences for users to enjoy!🤩
World Labs: We turned dreams into worlds.
Then filled them with history's greatest minds.
Not a video. A world, running directly in your browser. Step inside ↓
Peter Yang
Claude Fable 5 (Mythos) is now available!
Can't wait to give it the toughest questions like "What should I eat for lunch today" 😂
Claude
Re Releasing a model this capable comes with risks. Without safeguards, Fable 5’s capabilities in areas like cybersecurity could be misused to cause serious damage.
Queries on a narrow range of topics will instead receive a response from our next-most-capable model, Opus 4.8.
I've been at Anthropic through every model launch. There's been a few cases I can remember of a launch that stands out and marks a step-change in how we use models:
- Claude Opus 3
- Claude Sonnet 3.5
- Claude Opus 4.5
And now Claude Fable 5.
With Fable, the model stopped feeling like a tool I direct and started feeling more like something I collaborate with.
Claude: Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
Its capabilities exceed those of any model we’ve ever made generally available.
Manny Bernabe
Claude Fable 5 now available on Replit.
Build something amazing! 🚀
Claude: Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
Its capabilities exceed those of any model we’ve ever made generally available.
This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA on everything by a margin but I'll add that *qualitatively* also, this is a major-version-bump-deserving step change forward (imo of the same order as Claude 4.5 was in November), peaking especially for long problem-solving sessions on very difficult problems. You can give it a lot more ambitious tasks than what you're used to, the model "gets it" and it will just go, and it's never felt this tempting to stop looking at the code at all (but don't do this in prod!). The model still has quirks that people will run into and the safeguards are configured to be a little too trigger happy for launch, which can hopefully be tuned over time.
I feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing substantially. You can ask for anything - explainers, visualizers, dashboards, bespoke single-use apps (e.g. a full wandb that is hyper-specific just for your project), you can 10X your test suite, auto-optimize code, run giant research projects with custom HTML for the results, anything! "Free your mind" (Matrix ref). Really looking forward to all the things people build!
Claude: 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.
Uncle Matt
i wanted to use https://github.com/garrytan/gbrain by @garrytan
i already pay for chatgpt pro/codex so i wanted to use that weblogin
I Made GBrain Think Through Codex Without Buying Another Damn API Key
https://x.com/Unc_MattEth/status/2064408670012576180?s=20
use it yourself:
https://github.com/uncmatteth/gbrain-bridgebrain
thanks gary!
Uncle Matt: http://x.com/i/article/2064408161851678720
Zhen Li
Added a magic toggle called high effort in @Replit , meaning
1. Agent building in highest effort
2. Fable 5 making complex decisions
3. 25% off agent cost with high efforts
I’ve been using this mode and it feels like a step up in terms of building complex apps, magical experience
Michele Catasta: High Effort handles your most complex builds with ease on Replit.
Now powered by Claude Fable 5.
25% off for the next 7 days.
Amjad Masad
Try Claude Mythos at 25% off on Replit.
Michele Catasta: High Effort handles your most complex builds with ease on Replit.
Now powered by Claude Fable 5.
25% off for the next 7 days.
Try Claude Mythos at 25% off on Replit.
Michele Catasta: High Effort handles your most complex builds with ease on Replit.
Now powered by Claude Fable 5.
25% off for the next 7 days.
swyx
Re just finished rerunning FC Diamond on my historical charts. none of the official tables/charts are capturing the degree of takeoff.
https://x.com/karpathy/status/2064409694761054332
its this same chart all the way down difficulty classes (below) breaks every curve fit because Fable is a diffferent CLASS of model, with beeeeeg model smell.
Andrej Karpathy: This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA on everything by a margin but I'll add that *qualitatively* also, this is a major-version-bump-deserving step change forward
Re more charts of other tiers where its less stark
including the vibe shift chart from https://x.com/swyx/status/2064081945567580323 here
swyx: It's finally out!!! @METR_Evals found that more than half of SWEBench results is unmergeable slop. FrontierCode represents over 1000+ hours of maintainer validated software engineering work most frontier models cannot yet solve, much less solve with high quality.
Cog had IOI
Ken Goyarola
It's literally game time on @Replit . Go make one!
Michele Catasta: High Effort handles your most complex builds with ease on Replit.
Now powered by Claude Fable 5.
25% off for the next 7 days.
Opus wrote us a VM and then Mythos verified it
Malte Ubl: Huge thanks to the @AnthropicAI team for investigating `just-bash` with Mythos as part of Project Glasswing
There are no serious findings. I'll issue a release with minor hardening improvements after my vacation.
This is an interesting case for 2 reasons:
1. `just-bash` is a
Suhail
I would like to +1 that this is a very bad policy. Respond with a refusal and deal with the fall out but invisible NERFing is super uncool.
Mikel Artetxe
Brilliant idea! Next up: Apple randomly reboots your Mac if you're building competing tech, Gmail silently edits your email if you mention rival platforms, and Tesla Autopilot swerves if it detects you're working on self-driving cars.
All in the name of safety, of course. Because malicious actors controlling the world’s operating systems, inboxes and cars would be extremely dangerous!
elie: mythos will be bad ON PURPOSE on ai "frontier llm research" tasks, this is very very sad for the research community
also the fact that this is un purpose not visible to the user is crazy
Is it called Fable 5 because Fable 1-3 exist and 4 is coming out soon
Fable has solved 3D worldbuilding... utterly insane. This is all completely custom-built ThreeJs, running in the browser.
We've reset usage limits across our products!
For those just starting to test Fable, here's four tips for using it more effectively:
1. Give it bigger, more ambitious tasks than what previous models could handle.
2. Use xhigh/high effort as your default for best performance, med for faster interactive sessions.
3. Rework your skills and CLAUDE.mds. Instructions written for prior models anchor Fable to stale patterns, let it use its own judgment first.
4. Move from providing tasks to providing objectives. Describe what done looks like and how to verify it, then let Fable find the path (/loop and /goal are built for this)
Amol Avasare: To celebrate the Fable 5 launch, we just reset 5-hour and weekly limits for all users across our products!
Enjoy 🚀
Cool to see University of North Dakota (where I took a number of classes growing up) innovating and trying out offering A.I degrees: https://www.nytimes.com/2026/06/08/us/ai-college-degrees.html
Is Fable launched in Claude Code yet I don't see it
Suhail
Having spent a considerable amount of energy studying the robotics world, this is 🎯
SemiAnalysis: We just published a deep dive on why Unitree is going to dominate global robotics. Timing could not be better. (2/2)
https://newsletter.semianalysis.com/p/chinas-unitree-will-dominate-global
Jason ✨👾SaaStr.Ai✨ Lemkin
10 learnings from my SaaStr AI 2026 deep dive with @amasad, CEO of @Replit:
- Our two best employees now cost $254/mo
- Context windows went 16K → 1M in 2 years
- Replit built its own compaction. Says it beats the labs
- Bugs should be deleted from context. Architecture should stay
- Monorepos are a secret unlock for agent performance
- The best cold email I've ever seen ran mostly on Haiku + Mini
- Our agent beat last year's human marketing, and the gap widened
- Your 2nd and 3rd agent will be better than your 1st
- Every company will end up with an internal AI "Oracle"
- Replit improves itself nightly. A self-improving loop in production
Kavin was able to get similar results as well… this model is crazy good
Kavin Stewart: @mattshumer_ Not all the way. I stopped after a few iterations to save my tokens :)
For everyone complaining that this is slow, I ran the prompt 'make it faster, without losing quality', and voila:
Matt Shumer: Fable has solved 3D worldbuilding... utterly insane. This is all completely custom-built ThreeJs, running in the browser.
btw insane amounts of alpha in telling claude code to "review my code for issues" on Fable rn while it is not pay per use
be prepared to be in abject horror that you shipped anything to prod without a Fable Check™ first
swyx: just finished rerunning FC Diamond on my historical charts. none of the official tables/charts are capturing the degree of takeoff.
https://x.com/karpathy/status/2064409694761054332
its this same chart all the way down difficulty classes (below) breaks every curve fit because Fable is a diffferent
