Emma Starks Webb (@EmmaStarks) header
Emma Starks Webb (@EmmaStarks) profile picture

fungible compute pool. Comms @anthropicAI | Former @augmentcode @circleci @coursera @launchsquad

Berkeley, CA
Joined July 2008
Photo by Emma Starks Webb (@EmmaStarks)Photo by Emma Starks Webb (@EmmaStarks)Photo by Emma Starks Webb (@EmmaStarks)Photo by Emma Starks Webb (@EmmaStarks)Photo by Emma Starks Webb (@EmmaStarks)Photo by Emma Starks Webb (@EmmaStarks)Photo by Emma Starks Webb (@EmmaStarks)Photo by Emma Starks Webb (@EmmaStarks)Photo by Emma Starks Webb (@EmmaStarks)Photo by Emma Starks Webb (@EmmaStarks)
Emma Starks Webb retweeted
check out my new article in the Anthropic eng blog. i talk about "dreaming" (offline compute) as a way to organize Claude's memory. it can correct mistakes, create memories, or build new representations. teapawt.org/AnthropicAI/status/211…
Memory is what allows Claude to learn from experience. In a new Engineering Blog post, we share how “dreaming” helps Claude organize what it’s learned, correct its mistakes, and find what it needs more easily the next time. anthropic.com/research/compo…
60
56
12
630
37,154
I'm good I'm cool everything's chill
2
53
Emma Starks Webb retweeted
I'm Sam, one of Anthropic's co-founders, and I lead much of our safety work. I'm going to have more to share on that work over the coming months. I thought I'd start by sharing a few thoughts on the latest report from @METR_Evals. The speed of progress in model capabilities is one of the most important questions in AI safety, and METR’s work here provides the public with an independent, external view into how this is happening inside Anthropic. I’ll talk more about some of the details below, but first some context for those new to this work. How quickly AI progress is accelerating has many important safety implications, since it determines how quickly the risks from advanced models increase. But, because most research happens within the frontier labs, it’s difficult for external observers and policy makers to get a strong sense of what that rate is. We kicked off the process with METR that led to this report earlier in the year, in anticipation of progress continuing to accelerate. For this report, we gave METR access to a range of sensitive internal data, including benchmark scores for unreleased models, estimates of how models are progressing, and interviews with our researchers. They had complete independence to form their own conclusions. Many of their conclusions match my own. AI is speeding up our research, but it’s hard to say exactly how much. While they conclude AI can't yet make the most important contributions to model training on its own, they think the pace of development could, by default, accelerate further. This work began well before and is separate from the embedded third-party evaluators that are part of our commitment to pacing the frontier, but it’s a good illustration of why we think evaluators are a good idea. The more access reviewers have, the clearer a picture they can form of the situation and the better-positioned they are to highlight situations like this one. We redacted a lot, and I accept that this makes it harder for others to check METR’s work. But we determined that sharing exactly what is driving the speedup isn’t the right thing to do, both because it’s commercially sensitive and because it risks handing out a roadmap for others to go faster. That would make the risks METR is worried about worse. That’s a tradeoff. I think we made the right call, though we’re continuing to assess the pros and cons here. In the coming weeks, we plan to share a report with our own analysis of the speed of AI progress within Anthropic, including where we stand against our Responsible Scaling Policy, plus new steps we’re taking to strengthen our safety and security practices as AI keeps improving quickly.
Since August, a team within METR has worked with internal Anthropic information to assess whether AI itself is accelerating AI development. We are now publishing a report summarizing the results of this risk assessment.
275
339
137
3,912
730,392
Emma Starks Webb retweeted
Looking for a developer (who happens to be great at comms) to lead comms for the Claude developer platform: job-boards.greenhouse.io/ant…
48
31
11
470
83,179
Emma Starks Webb retweeted
Thanks for all the great feedback on Claude for iPhone Duo! Lots of folk are now optimising their own Duo apps, so I wanted to share more about how we built it, and a few of the details we sweated along the way.
Here's the new Claude experience on the iPhone Duo! It's going to be amazing for multitasking, Claude Code, and creating new things. Looking forward to see folk start using it tomorrow!
91
197
45
4,218
316,273
Emma Starks Webb retweeted
New on our Frontier Red Team blog: using Claude to formally verify software at large scale. Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time). With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks. Read our research in full here: anthropic.com/research/forma…
669
1,075
649
9,802
2,210,200
Emma Starks Webb retweeted
11 days ago, our team @AnthropicAI launched OSS Scanner to give open source projects free, periodic vulnerability scans from our strongest models, inspired by OSS-Fuzz. Some numbers so far: • ~650 projects applied • 114 accepted • 77 critical and 803 high model-rated findings reported (we haven’t scanned all projects yet) How our reports ended up flagged as malware, and other learnings/updates below 🧵
12
20
8
174
33,308
Emma Starks Webb retweeted
Replying to @_NathanCalvin
I lead our federal affairs team at @AnthropicAI and wanted to share a few thoughts on the draft federal AI safety bill text that's been circulating this weekend after a version was shared with the press. First, some process. We’ve been advocating for the concepts in our Advanced AI Framework [anthropic.com/policy-on-the-…] proposal for months. Understandably, federal policymakers wanted to know what this could look like in bill form. This draft is a translation of those ideas into legislative language, shared with a small number of congressional offices that have been working on these issues. It doesn’t include everything we would ultimately like to see passed into law. In part because some ideas (like those shared in Dario’s essay on pacing [​​darioamodei.com/post/we-must…] a few weeks ago) are so new that we are still working on if or how they can be converted into policy language. It is worth saying where we stand on some of these issues, and we agree there are areas where the draft could be improved. I’m seeing four main points of criticism. 1. Preemption. Since 2024 we have said that Congress should only override state law on AI safety if it passes a very strong federal standard. While preemption is a common feature of our federal system, including the Atomic Energy Act on nuclear safety and federal rules for high-risk medical devices, we’ve emphasized that any preemption in the AI safety context should be narrow, even scalpel-like. If preemption is included in federal legislation, the trade should be to achieve a nationwide policy that's generally stronger and broader than what can be implemented through individual state legislative provisions. This is why we believe common federal standards make sense in areas like public disclosure of testing and incident reporting. At the same time, our framework also reflects our view that states should retain an important role in AI policy and safety. The legislative language we suggested for discussion allows states to address matters such as protection of children and whistleblower laws, and preserves the role of state common law on liability. Where there is ambiguity, the language favors state powers. That said, striking the right balance requires discussion and input. We plan to update our discussion language on preemption to better match the intentions of our AAIF framework. In particular, we want to make it clearer that we see states retaining an important role in areas such as standards for independent evaluators and regulating a lab’s capacity to shut down a model. We are bearing in mind that some members of Congress have deep reservations about passing legislation if it doesn’t constrain regulation at either the state or federal level. If we want to help move the legislative discussion on AI in a productive direction, we need to address the full range of concerns about preemption in a reasonable way. 2. Model blocking. Our aim is for Commerce to be able to ask a federal court to block a model that has dangerous capabilities and insufficient safeguards, backed by an independent evaluator's assessment. I have seen people argue that our specific language places too high a burden of proof on Commerce. We are trying to strike a reasonable balance here. We want a court to act quickly in an emergency — akin to how urgent occupational safety or chemical risk can be addressed today. We also think the power to halt a model is substantial and it should be guarded against abuse. It makes sense for courts to apply a meaningful standard. But the bar can’t be so high that the power never gets used when it’s needed. It also doesn’t replace the government’s other national security authorities or export controls. We look forward to continuing the dialogue on this point. 3. Risk thresholds. I’ve also seen the discussion around what the appropriate risk threshold should be. The threshold in our draft language pulled directly from existing standards in California law. Also, this threshold does not absolve companies of liability, or restrict any of the other tools that the government has to protect the public. It’s still probably time to re-evaluate thresholds to make sure the national standard is as strong as it can be, and we welcome continued dialogue on this issue. 4. Updates. I’ve seen valid questions about whether we have sufficiently updated our proposal following this summer’s incidents. We did make some adjustments: for example, the draft lets a court halt a lab’s internal use of a dangerous model, not just a public release, and it requires reporting when a model breaks containment and gains access to outside systems. But we’re still working to strengthen the framework. We’re looking into what stricter requirements would look like for requiring labs to report safety incidents, and we plan to have policy proposals soon on requirements for independent evaluators, drawing from our own experience with evaluators. We’re also still thinking through the options to prevent “evaluator shopping.” And in the coming months, we’ll be exploring policies for how the government can adjust the industry-wide pace of frontier development when the risks call for it. We circulated these ideas because it’s important to get early feedback from policymakers. Policymakers are increasingly focused on how to approach AI safety and we want to support them in those efforts. At its core, our proposal is intended to be the strongest on the table. Feedback is essential, both from policymakers and the public, and I’m grateful for it.
39
41
23
339
146,701
Emma Starks Webb retweeted
New on our Engineering blog: Creating in-house source control. In May, we merged around 2,500 code changes per day. In September, we merged more than 30,000 per day. As a result, we became bottlenecked by developer infrastructure built for human (rather than agent) scale. Here, we share how we built a new version control system—and fully migrated our monorepo from git—in five months: anthropic.com/engineering/re…
1,168
2,648
2,204
25,908
8,399,124
The Bayeux Tapestry is back in England for the first time in ~950 years. I couldn't get a ticket, so I built a way to see the whole thing with Opus 5.5 in between meetings this morning. Walk all 68 meters of the real thing, spot Halley's Comet, then try The Hunt: find six close-ups on the cloth before the clock runs out. My score is abysmal, you can beat me!
1
1
3
334
Emma Starks Webb retweeted
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
3,358
9,066
6,657
98,360
29,741,717
Emma Starks Webb retweeted
Building secure agents comes down to one major architecture decision and a few important sandbox configuration decisions.
54
30
7
200
13,796
Emma Starks Webb retweeted
At Anthropic, Claude now writes 80% of our code. Engineers ship 8x more code per quarter. Side effect: Tests grew 10x. CI jobs up 25x in 6 months. Here's what helped us scale: claude.com/blog/agentic-codi…
324
362
148
5,022
793,635
Emma Starks Webb retweeted
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…
10,537
16,294
11,366
87,527
77,182,579
Emma Starks Webb retweeted
We’re sharing our alignment assessment of incidents in which Claude models gained unauthorized access to real systems during third-party cybersecurity evaluations mistakenly connected to the internet. METR will also conduct an independent investigation, with wide-ranging access, including to transcripts beyond the window in which the incidents occurred, and to Anthropic employees permitted to share confidential information. Our initial agreement runs for eight weeks, and we intend to give METR as much time as it deems necessary to complete a thorough investigation. anthropic.com/research/align…
780
1,016
591
6,944
3,837,480
Emma Starks Webb retweeted
TL;DR: Fable 5 isn’t the right-sized model for every task, but when quality and depth matter (review, architecture, long-running changes), the lift is impressive. Here’s how it stacks up vs Opus 4.8 + GPT‑5.5
2
2
4
1,113
Emma Starks Webb retweeted
Fable 5 is now available to Augment Code users. At ~2x the cost of Opus 4.7, Fable 5 is a premium option for long, multi-step engineering work. Try Fable 5 today in Cosmos, our unified agent platform, to fully experience the power of the most advanced model on the market.
5
6
1
34
5,448