Marketing @AnthropicAI | Host @LimitlessFT Curiously thinking about and occasionally building the future
New York
- Tweets 9,103
- Following 725
- Followers 15,935
- Likes 42,407
Vulnzero.
In ~18 months, scaling laws will have compressed today’s frontier-level cyber capabilities into small and cheap models
Everyone can use them for ~free to catch and prevent security bugs before code ships, permanently ending the vulnerability treadmill
as bad as things have gotten in the past 4 days, they are going to get a lot worse before they get better. security leaders, never let a crisis go to waste: please go to your boards and ask for help right now. not for ai, for the basics. (we need to hold on for ~18 months.)
7
799
This paper was fascinating, but very dense and a LOT of pages
Had Opus 5.5 explain it to me simply in video form
interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had
with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works
already-clean.goodfire.com/
1
3
641
The claude 5.5 family is dominating the Pareto frontier
Haiku delivering frontier level intelligence at a fraction of the cost with Sonnet and Opus the workhorse rounding it out. Thing of beauty
Claude Haiku 5.5 is now available in Cursor!
On shorter requests, it costs 10x less than Claude Haiku 4.5.
Turn it on in Cursor Settings > Models.
1
11
966
Today we published my final episode as host of @LimitlessFT
Over the last year, our team of three shipped 234 episodes, going from zero to top 20 in tech, growing 30% every single month.
I’ve recently joined the team at @AnthropicAI to take the principles that made us win, and apply them at a scale my childhood self could’ve only dreamed of.
I’m so grateful for every person who trusted us with their time and shared words of support. And for @TrustlessState who believed in me even when I was unsure ♥ the courage to bet on people the way he did is rare, and this show doesn’t exist without him and the support of Bankless.
It’s also important to note… THE SHOW GOES ON 🎬
@cryptopunk7213 is steering the ship at our new home with Yahoo!, who is providing all the support in the world to continue publishing incredible content about AI and Tech every single day.
The bts mastermind of our podcast @luke_spotlong is sticking around too, so expect the usual A+ quality of conversation and production.
Luke said yesterday - “god y’all are like the Beatles. Paul & John breaking up at the peak of their powers”
And while we're certainly at the height of our powers, the relationship between Ejaaz and I couldn't be in a better place.
I hope you enjoy this final episode as much as we did, and try not to roast the cardinal sin of crying on a podcast.
Thank you for believing in me, thank you for believing in us. It’s a big century for the optimists, and what an exciting journey it will be ♥️
10
2
39
1,706
Josh Kale retweeted
Today we announced the Claude-led discovery of a molecular machine that we suspect could represent a new gene editing mechanism. Its precise function, biotechnological utility (if any), or level of significance is not yet clear, but at minimum it is work I would have been proud to do as a PhD student. The work was done mostly, though not entirely, by Claude: our life sciences team suggested a broad area of research, Claude read through the literature and a bunch of genome data and discovered something interesting, then Claude proposed experiments to verify the discovery and our team carried them out.
It’s easy to dismiss this as a one-off or curiosity, but we’ve repeatedly seen a pattern where AI performance in new intellectual domains goes from weak to superhuman in a matter of a few years. In 2023 models struggled to do math at the level of an average high-school student. In 2024 they started to do well on math competitions for the best high-schoolers in the country, in 2025 they started to solve minor open problems, in early 2026 more significant open problems, and in late 2026 they are beginning to solve the top few open problems in all of mathematics. We believe AI for biology is on a similar exponential trend.
The main difference between biology and mathematics, of course, is that math can be done purely theoretically, while biology requires experimentation. Some have used this to draw the conclusion that AI’s utility in biology will be limited. We think this is wrong. As we’ve demonstrated today, humans can collaborate with AI to perform the experiments, validate key results in a few weeks and, if necessary, work with the AI to iterate on what they find. Eventually it may even be possible for Claude itself to safely perform the experiments by autonomously controlling lab equipment, with appropriate safeguards in place, but we aren’t doing that today (our lab is also a BSL1/BSL2 facility that doesn't handle materials dangerous to humans).
More broadly, biomedical advancement has many stages — from fundamental biology discoveries, to translational research, to drug discovery, clinical trials, and finally the actual delivery of medicines and health care to patients. We are also interested in these later stages, but even simply accelerating the first stage of fundamental biological discoveries has the potential to speed up and broaden the entire pipeline. Improving our understanding of biology and sharpening biologists’ tools can drive forward all of the later stages, for example by identifying new drug targets, finding new therapeutic modalities, allowing for more precise measurement, and speeding up the experimental loop which itself further accelerates our understanding of biology. This will not in itself speed up clinical trial times, but if it succeeds it could greatly increase the number of promising candidates that go into the pipeline — an increase in throughput even though latency remains.
In Machines of Loving Grace, I wrote about AI’s potential to “cure most diseases in 5-10 years” — a goal that sounds impossible, but one I believe is just barely possible if AI is applied to every stage of the pipeline. The first step is showing that AI can first help with, and then drive, biological discoveries.
Claude’s discovery is the latest in a line of related prior work that goes back decades, beginning with systems like CRISPR, and continuing with discoveries like the bridge recombinase and VIPR in the past few years. Recently, there has been heightened interest in systems based on reverse transcriptase (RT) enzymes, the enzyme underlying the system Claude identified. And most recently, a Stanford team working independently described a novel RT system with an associated non-coding array that is in some ways similar to the one Claude found, though they are distinct systems that evolved independently from each other. I believe that we’re at the very beginning of finding such systems and developing them into powerful tools for biotechnology.
I’m proud of the resources Anthropic has invested in accelerating the public benefits of AI through the life sciences, and we’re aiming both to grow our life sciences team and to work with other scientists to extend this approach to a broad range of problems. If you have a proposal for a research collaboration or are interested in joining our life sciences team, please reach out.
Claude has discovered a previously unknown enzyme system hidden in the DNA of bacteriophages. Beside the enzyme’s gene sits a long array of repeating DNA—a structure that looks somewhat similar to CRISPR.
We don’t yet understand what this system does, but only a handful of known systems share its features, and all of them are able to cut, copy, and paste DNA. Historically, the discovery of such programmable systems has helped revolutionize medicine. CRISPR, for instance, is now the foundation of genetic medicines. But it will take much more work to learn what this system does, and whether it can be put to similar use.
Read more: anthropic.com/news/claude-di…
1,677
3,437
29,985
5,583,265
Dario is right
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
6,273
6,342
57,841
12,647,616
We can’t keep letting them get away with this
We've been in business for 22 years and we
LOVE our customers.
Look who swung by on Sunday!
Local entrepreneur Palmer Lucky !! He has been a loyal customer of our little flower stand for many years.
His favorite bouquet is Spring Mist.
We feel so LUCKY ;) to have people like him to keep our NB community thriving.
Just remember shoes next time Palmer ;)
32
11
1,232
1,476,373
Fable + Mythos 5.1!
Smarter, more efficient, and:
"Our cybersecurity safeguards now flag benign requests about 60% less often.
On basic biology and medical questions, we’ve recently reduced the fallback rate by around 85%."
We’re introducing Claude Fable 5.1 and Claude Mythos 5.1.
They're the world’s most advanced models for coding and knowledge work.
2
7
4,368
Elon on the SpaceX earnings call when asked about their ability to build terrestrial data centers:
“It’s like, frankly, the New York Yankees going in and playing a little league team.”
The same Starmind V1 satellite design (minus solar & radiator) will be deployed on the ground in our data centers. It’s a major improvement in data center efficiency.
1
1
6
2,637
We are so unbelievably back
We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5.
We'll begin restoring access tomorrow, and will share an update soon.
We’re grateful to our users for their patience, and to everyone who worked with us on redeploying the models.
5
16
3,541
This New Glenn rocket explosion released 20% of the energy of the Hiroshima atomic bomb and that wasn't even the bad part:
→ The pad: LC-36 is the only pad on Earth that launches New Glenn and now it's gone. Over $1B to build. SpaceX needed 7 months to rebuild after a similar hit.
→ The deadline: Amazon needs 1,618 satellites up by July 30 to keep its FCC license. It has ~300. The rocket that was supposed to help fix that just blew up twice in a row
SpaceX made us believe that landing rockets on barges was a normal expectation. Turns out rocket science is hard after all. Wishing the team a speedy recovery 🚀
We experienced an anomaly during today's hotfire test. All personnel have been accounted for. We will provide updates as we learn more.
1,001
2,686
33,565
4,984,506
Everybody will be wrong about Opus 4.8 if they only look at benchmarks. Under the hood there's some amazing new capabilities:
→ Fast mode now runs 2.5x faster and 3x cheaper than it did on older models
→ 4x less likely to let a bug in its own code slip past vs 4.7
→ A new effort dial. You set how hard it thinks per task. Max it for the hard problems, drop it low to stretch rate limits
→ Dynamic workflows in Claude Code runs hundreds of parallel subagents in one session. Codebase-scale migration, hundreds of thousands of lines, kickoff to merge
→ Same price as 4.7. $5 in, $25 out per million
This is a pretty epic upgrade across the board.
Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors.
Available today at the same price.
ALT Benchmark table showing how Claude Opus 4.8 compares to its predecessor and to other models on tests of coding, agentic skills, reasoning, and practical knowledge work tasks.
2
25
5,514
This is like Michael Jordan returning from retirement in '95 to go on the most generational championship run in history
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
1
12
2,754