Joined July 2021
128 Photos and videos
I need Fable back @AnthropicAI, please guys
1
32
this is a primitive
Jun 16
Today we're introducing the world's first influencer Agent. Tell it what you want to promote and it finds creators, reaches out, manages campaigns, handles payments, and gets content live. Try it now at okara.ai/influencer
1
16
talking his book, but its not unreasonable BUT the labor contract needs a first-principles rewrite
JEFF BEZOS AT VIVATECH: I THINK AI IS GOING TO CREATE A LABOUR SHORTAGE
9
Daniel retweeted
cursor just launched their GitHub replacement called origin
104
112
3,329
301,940
Truescan basically lives in my group chats now someone News comes up, type one command, get a quick prediction market chart. Polymarket and Kalshi are banned in most countries; the bot makes it much easier
1
30
Elon wont be happy if $SPCX doesn’t become the most valuable company
1
47
i hope @karpathy still publishes his coding experiments; the LLM-wiki he came up with is super underrated!
21
"private evals" Basically entering the hive-mind era of agents study up on your harness engineering; it's going to be a major discipline
28
Daniel retweeted
Building AI-Native Startups [002] x.com/i/broadcasts/1PJqrrRbQ…
3
14
58
2,902
Daniel retweeted
🚨 @Karpathy predicted the power of the "LLM Wiki." Google just formalized it. Meet Open Knowledge Format (OKF): a vendor-neutral standard for giving foundation models the curated context they need. I can genuinely see this replacing Notion, Obsidian, or traditional wikis for developer teams, and the reason comes down to bookkeeping. Traditional wikis fail because humans inevitably abandon the tedious work of updating them. As Andrej Karpathy pointed out recently, LLMs don't get bored. They don't forget to update a cross-reference, and they can touch 15 files in a single pass. OKF standardizes the interoperability layer so agents can actually do that heavy lifting autonomously. Because the format is minimally opinionated, it doesn't dictate what you write, it just dictates how it's structured. You get: → Human-readable documents that live right alongside your code in version control → Cross-links that map out complex entity relationships without needing a graph database → A system that survives moving between different tools and organizations There is no complex compression scheme. No central registry. If you can cat a file, you can read it. If you can git clone a repo, you can deploy it. This is how we stop rebuilding context pipelines from scratch every time a new model drops. Announcement spec file in 🧵↓
64
308
2,254
203,476
Daniel retweeted
As a result of a US government directive, we are suspending access to Claude Fable 5 for all users. You can continue to use all other Claude models. Here’s what this means for you: Across Claude products, new sessions will run on your selected default model or Opus 4.8, and existing Fable 5 sessions will end with an error. On the Claude Platform, requests to Fable 5 will also return an error. Please update your integrations to other Claude models. We know this is a disruption to your workflows; we appreciate your patience and support.
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…
3,677
7,274
44,664
12,929,714
Daniel retweeted
Someone asked me to recommend a good book on parenting. I would guess that most books about parenting are tedious. So my recommendation is to read autobiographies, the early parts of which are usually implicitly about parenting.
177
88
2,135
378,024
Daniel retweeted
Fable 5 says poker will change your life
117
278
4,950
415,860
Insane if true
Replying to @SemiAnalysis_
Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until we exhausted the weekly limit. It's widely believed that a $200/month plan maxes out at ~$2000/month worth of tokens (assuming API pricing). However, we found that the subscriptions are actually far more generous. (2/4)
41
going band for band
OPENAI PREPARING TO RELEASE A NEW AI MODEL, CODENAMED 5.6 THIS MONTH -THE INFORMATION
12
No, the biology is being gated because you are moving to a different business model: pay-for-results instead of pay-for-tokens, as Dario himself said on the Dwarkesh pod.
Replying to @claudeai
Fable 5’s safeguards detect requests related to cybersecurity, biology and chemistry, and distillation. Users are informed whenever a fallback occurs—on average in less than 5% of sessions. We’ll keep refining the safeguards to reduce false positives.
34
So I implemented a whole feature yesterday, which was about 5h with Codex xHigh /goal. Ran a Claude Fable high audit on it today, and it figured out all the issues with it, plus a bunch of improvements, which were the same as the ones I found manually. GIVE ANTHROPIC MORE COMPUTE @elonmusk
41
I could see the AI bubble ending this week or going for another 4 years.
15
I just ran a Claude Fable 5 audit on the product I am building and got some critical fixes. But it kind of reminds me of devs worrying about security when the product has 0 traffic… The honey pot has to get big enough for these issues to become problematic, which is a problem for incumbents, not "AI slop"/new products (including me). But it's good to see models improving.
52