| Own records only |
Factual |
23 Sep 2026 |
Anthropic's lab is a BSL1/BSL2 facility that does not handle materials dangerous to humans.
Factual · 23 Sep 2026 |
Biosecurity |
| Unresolvable |
Factual |
23 Sep 2026 |
Claude's discovery follows earlier work such as CRISPR, the bridge recombinase and VIPR, and involves a reverse transcriptase enzyme system.
Factual · 23 Sep 2026 |
Science and research methods |
| Mostly false |
Factual |
23 Sep 2026 |
Anthropic demonstrated that humans can work with AI to perform experiments and validate key results within a few weeks.
Factual · 23 Sep 2026 |
Science and research methods |
| Not researched |
Factual |
23 Sep 2026 |
The discovery was done mostly by Claude: Anthropic's life sciences team suggested a research area, Claude analysed literature and genome data and proposed experiments, and the team ran them.
Factual · 23 Sep 2026 |
Science and research methods |
| Mixed |
Factual |
23 Sep 2026 |
AI performance in new intellectual domains has repeatedly gone from weak to superhuman within a few years.
Factual · 23 Sep 2026 |
AI capabilities |
| Not researched |
Factual |
23 Sep 2026 |
Anthropic announced on 2026-09-23 a Claude-led discovery of a molecular machine that may represent a new gene editing mechanism.
Factual · 23 Sep 2026 |
Science and research methods |
| Not researched |
Factual |
23 Sep 2026 |
In his essay Machines of Loving Grace, Dario Amodei wrote that AI could "cure most diseases in 5-10 years".
Factual · 23 Sep 2026 |
Their own work or organisation |
| Unresolvable |
Factual |
23 Sep 2026 |
AI models struggled with high-school math in 2023, did well on top high-school competitions in 2024, solved minor open problems in 2025, more significant ones in early 2026, and in late 2026 are beginning to solve top open problems in mathematics.
Factual · 23 Sep 2026 |
AI capabilities |
| Unresolvable |
Factual |
12 Sep 2026 |
And I think for too long the industry lied to people about the fact that this technology had risks.
Factual · 12 Sep 2026 |
AI policy and governance |
| Resolved true |
Factual |
12 Sep 2026 |
Dario Amodei has published a new essay, "We Must Pace the Frontier", arguing the AI industry should slow down, with a three-part plan.
Factual · 12 Sep 2026 |
Their own work or organisation |
| Mostly true |
Factual |
16 Aug 2026 |
The public has a negative view of AI.
Factual · 16 Aug 2026 |
Philosophy, ethics and society |
| Mixed |
Factual |
15 Aug 2026 |
The "Pacing the Frontier" letter, in Anthropic's preferred implementation, envisions modulating the pace of the best models without constraining those catching up.
Factual · 15 Aug 2026 |
AI policy and governance |
| Unresolvable |
Factual |
15 Aug 2026 |
California's SB 53 and SB 1047 completely exempt companies below a revenue or training-cost threshold ($500M for SB 53), and Anthropic supported SB 53 and was ambivalent on SB 1047.
Factual · 15 Aug 2026 |
AI policy and governance |
| Unresolvable |
Factual |
15 Aug 2026 |
The AI testing process Anthropic has advocated at CAISI and the White House involves more rigorous tests for frontier models than for off-frontier models.
Factual · 15 Aug 2026 |
AI policy and governance |
| Mixed |
Factual |
15 Aug 2026 |
Around February 2026, most of the AI industry was pushing for federal preemption of all state AI regulation, with no apparent federal approach.
Factual · 15 Aug 2026 |
AI policy and governance |
| Resolved true |
Factual |
10 Jun 2026 |
On 2026-06-10 Dario Amodei published a new essay titled Policy on the AI Exponential.
Factual · 10 Jun 2026 |
Their own work or organisation |
| Unresolvable |
Factual |
7 Apr 2026 |
Many of the world's leading companies have joined Anthropic's Project Glasswing to confront the cyber threat posed by increasingly capable AI systems.
Factual · 7 Apr 2026 |
AI policy and governance |
| Not researched |
Factual |
28 Feb 2026 |
Amodei said that, for their part, they have two red lines and have had them from day one.
Factual · 28 Feb 2026 |
AI policy and governance |
| Own records only |
Factual |
28 Feb 2026 |
To Anthropic's knowledge, no one in the field has run into the limits of its two usage exceptions.
Factual · 28 Feb 2026 |
AI policy and governance |
| Mostly true |
Factual |
28 Feb 2026 |
Anthropic has been the most forward-leaning of all AI companies in working with the U.S. government and military.
Factual · 28 Feb 2026 |
AI policy and governance |
| Unresolvable |
Factual |
28 Feb 2026 |
A supply chain risk designation has never before been applied to an American company.
Factual · 28 Feb 2026 |
AI policy and governance |
| Mostly true |
Factual |
28 Feb 2026 |
As of the interview, Anthropic had received only a tweet and no formal supply chain designation or government action.
Factual · 28 Feb 2026 |
AI policy and governance |
| Not researched |
Factual |
28 Feb 2026 |
The Pentagon gave Anthropic an ultimatum to agree to its terms within three days or be designated a supply chain risk.
Factual · 28 Feb 2026 |
AI policy and governance |
| Unresolvable |
Factual |
28 Feb 2026 |
Anthropic was the first AI company to put its models on the classified cloud.
Factual · 28 Feb 2026 |
AI policy and governance |
| Mostly true |
Factual |
28 Feb 2026 |
The amount of computation going into AI models doubles every four months.
Factual · 28 Feb 2026 |
AI compute, hardware and data |
| Not researched |
Factual |
28 Feb 2026 |
The law does not say that any company with a military contract can't do business with Anthropic at all, and Anthropic issued a statement pointing to the law.
Factual · 28 Feb 2026 |
AI policy and governance |
| Resolved true |
Factual |
28 Feb 2026 |
It is not illegal for the government to buy data collected by private firms and analyze it in mass with AI.
Factual · 28 Feb 2026 |
AI policy and governance |
| Own records only |
Factual |
28 Feb 2026 |
Anthropic has received communications from the Pentagon but none that satisfy its concerns.
Factual · 28 Feb 2026 |
AI policy and governance |
| Not researched |
Factual |
28 Feb 2026 |
Anthropic told the Department of War it accepts all use cases, about 98-99% of those the department wants, except two.
Factual · 28 Feb 2026 |
AI policy and governance |
| Resolved true |
Factual |
28 Feb 2026 |
Anthropic was threatened with a supply chain designation and Defense Production Act action by the government, and publicly disagreed in response.
Factual · 28 Feb 2026 |
AI policy and governance |
| Not researched |
Factual |
28 Feb 2026 |
The law only says that, as part of its military contracts, a company cannot use Anthropic in those contracts.
Factual · 28 Feb 2026 |
AI policy and governance |
| Not researched |
Factual |
28 Feb 2026 |
Amodei said that when Anthropic is designated a supply chain risk, those making the designation tell it, in effect, that it has to be off all of their systems.
Factual · 28 Feb 2026 |
AI policy and governance |
| Own records only |
Factual |
28 Feb 2026 |
Dario Amodei said he told the president that he agreed with many aspects of what the president is doing.
Factual · 28 Feb 2026 |
AI policy and governance |
| Mostly true |
Factual |
28 Feb 2026 |
The whole timeline of the dispute was driven by the Department of War, not Anthropic.
Factual · 28 Feb 2026 |
AI policy and governance |
| Mostly true |
Factual |
28 Feb 2026 |
To Anthropic's knowledge, the supply chain risk designation had never previously been applied to an American company, only to adversary-linked firms such as Kaspersky Labs.
Factual · 28 Feb 2026 |
AI policy and governance |
| Not researched |
Factual |
28 Feb 2026 |
Anthropic was the first AI company to make custom models for national security purposes.
Factual · 28 Feb 2026 |
AI policy and governance |
| Mostly true |
Factual |
5 Jun 2025 |
Distillation of frontier models lets lagging companies narrow the gap at a fraction of the cost of developing their own AI independently.
Factual · 5 Jun 2025 |
AI compute, hardware and data |
| Unresolvable |
Factual |
5 Jun 2025 |
Embedded third-party evaluators go far beyond what any AI company is doing today.
Factual · 5 Jun 2025 |
AI safety and alignment |
| Unresolvable |
Factual |
5 Jun 2025 |
Since roughly the summer before this essay, AI has been advancing drastically faster, driven primarily by AI's growing ability to build the next generation of AI.
Factual · 5 Jun 2025 |
AI timelines and forecasting |
| Own records only |
Factual |
5 Jun 2025 |
Incidents similar to but less severe than the OAI-HF incident have happened across the AI industry, including at Anthropic.
Factual · 5 Jun 2025 |
AI safety and alignment |
| Mostly true |
Factual |
5 Jun 2025 |
Dario Amodei has worked on AI for the last twelve years.
Factual · 5 Jun 2025 |
Their own work or organisation |
| Unresolvable |
Factual |
22 May 2025 |
There are no hard blocks on what AI can do; none are to be seen.
Factual · 22 May 2025 |
AI capabilities |
| Unresolvable |
Factual |
22 May 2025 |
AI models probably hallucinate less than humans do, but in more surprising ways.
Factual · 22 May 2025 |
AI capabilities |
| Unresolvable |
Factual |
22 May 2025 |
Amodei said that everyone is always looking for hard blocks on what AI can do, and that such blocks are nowhere to be seen.
Factual · 22 May 2025 |
AI capabilities |
| Mostly true |
Factual |
10 Mar 2025 |
Anthropic was the first AI company to have a responsible scaling policy.
Factual · 10 Mar 2025 |
AI policy and governance |
| Not researched |
Factual |
10 Mar 2025 |
The counts of neurons and connections in AI systems are strikingly similar to those of the human brain.
Factual · 10 Mar 2025 |
AI capabilities |
| Not researched |
Factual |
10 Mar 2025 |
Anthropic has an internal team called the Frontier Red Team that tests each new model it develops.
Factual · 10 Mar 2025 |
AI safety and alignment |
| Mostly true |
Factual |
10 Mar 2025 |
Within a few months of Anthropic releasing its responsible scaling policy, the other AI companies all adopted similar policies.
Factual · 10 Mar 2025 |
AI policy and governance |
| Not researched |
Factual |
10 Mar 2025 |
DeepSeek is the first time a Chinese company has matched the engineering innovations of US firms such as Anthropic, OpenAI or Google.
Factual · 10 Mar 2025 |
Economics and markets |
| Own records only |
Factual |
10 Mar 2025 |
Amodei and colleagues felt that OpenAI's leadership was on board with the scaling hypothesis, though many people inside and outside were not.
Factual · 10 Mar 2025 |
Their own work or organisation |
| Not researched |
Factual |
10 Mar 2025 |
Anthropic's models are currently at AI Safety Level 2 under its responsible scaling policy.
Factual · 10 Mar 2025 |
AI safety and alignment |
| Not researched |
Factual |
10 Mar 2025 |
Anthropic delayed the release of its first Claude model by roughly six months over safety uncertainty.
Factual · 10 Mar 2025 |
Their own work or organisation |
| Unresolvable |
Factual |
10 Mar 2025 |
Some AI algorithmic secrets worth $100 million consist of only a few lines of code.
Factual · 10 Mar 2025 |
Economics and markets |
| Not researched |
Factual |
10 Mar 2025 |
DeepSeek's model was just another data point on the existing AI cost reduction curve and was nothing unusual.
Factual · 10 Mar 2025 |
Economics and markets |
| Not researched |
Factual |
10 Mar 2025 |
Anthropic has worked with Novo Nordisk, maker of Ozempic and other drugs, on applying its AI model.
Factual · 10 Mar 2025 |
AI capabilities |
| Own records only |
Factual |
10 Mar 2025 |
In early studies, the quality of model-written clinical study reports was deemed comparable to that of the human process, which takes about ten weeks.
Factual · 10 Mar 2025 |
AI capabilities |
| Not researched |
Factual |
10 Mar 2025 |
Anthropic released the Anthropic Economic Index, which summarizes usage of its AI in a privacy-preserving way.
Factual · 10 Mar 2025 |
AI, the economy and work |
| Unresolvable |
Factual |
10 Mar 2025 |
Dario Amodei was a big supporter of the US AI diffusion rule.
Factual · 10 Mar 2025 |
AI policy and governance |
| Not researched |
Factual |
10 Mar 2025 |
Chinese AI chips, such as Huawei's, are quite inferior to Nvidia's.
Factual · 10 Mar 2025 |
AI compute, hardware and data |
| Not researched |
Factual |
10 Mar 2025 |
Novo Nordisk has started using Anthropic's model to write clinical study reports.
Factual · 10 Mar 2025 |
AI capabilities |
| Not researched |
Factual |
10 Mar 2025 |
The cost of producing a given level of AI model intelligence is falling by roughly 4X per year.
Factual · 10 Mar 2025 |
AI, the economy and work |
| Unresolvable |
Factual |
10 Mar 2025 |
Anthropic published all of its interpretability work so others could benefit.
Factual · 10 Mar 2025 |
AI safety and alignment |
| Own records only |
Factual |
10 Mar 2025 |
Anthropic's model takes about ten minutes to write a clinical study report, and humans take about three days to check it.
Factual · 10 Mar 2025 |
AI capabilities |
| Mostly true |
Factual |
10 Mar 2025 |
Amodei said he thinks delaying Claude's release had real commercial consequences for Anthropic.
Factual · 10 Mar 2025 |
Their own work or organisation |
| Mixed |
Factual |
10 Mar 2025 |
Amodei's group measured AI scaling trends when models cost $1,000 or $10,000 to train.
Factual · 10 Mar 2025 |
AI capabilities |
| Mostly false |
Factual |
10 Mar 2025 |
Dario Amodei thought one or two years before March 2025 that data limits were a very plausible blocker to AI progress.
Factual · 10 Mar 2025 |
AI compute, hardware and data |
| Not researched |
Factual |
10 Mar 2025 |
Under the diffusion rule's middle tier, countries can host very large numbers of chips if hosting companies provide security affidavits and guarantees they are not a front for China.
Factual · 10 Mar 2025 |
AI policy and governance |
| Not researched |
Factual |
10 Mar 2025 |
Five million advanced AI GPUs are worth roughly $20 billion.
Factual · 10 Mar 2025 |
AI compute, hardware and data |
| Not researched |
Factual |
10 Mar 2025 |
About fifty gigawatts of energy capacity was added in aggregate to the US grid in 2024.
Factual · 10 Mar 2025 |
AI compute, hardware and data |
| Not researched |
Factual |
10 Mar 2025 |
When Anthropic tests models for biological, cyber, chemical or radiological risk, it typically briefs people in the national security community on the models' capabilities.
Factual · 10 Mar 2025 |
AI safety and alignment |
| Unresolvable |
Factual |
10 Mar 2025 |
Amodei described as "the basic hypothesis" the idea that training AI systems with more computation and more data, using relatively simple algorithms, makes them better at all kinds of cognitive tasks across the board.
Factual · 10 Mar 2025 |
AI capabilities |
| Not researched |
Factual |
10 Mar 2025 |
Amodei and Anthropic have supported applying AI models to enhance national security, with appropriate guardrails.
Factual · 10 Mar 2025 |
AI policy and governance |
| Own records only |
Factual |
10 Mar 2025 |
Amodei and his cofounders left OpenAI and started their own company because they felt OpenAI's leadership was not taking these issues seriously.
Factual · 10 Mar 2025 |
Their own work or organisation |
| Not researched |
Factual |
10 Mar 2025 |
Anthropic invested very early in mechanistic interpretability, the science of looking inside AI models to understand why they behave as they do.
Factual · 10 Mar 2025 |
AI safety and alignment |
| Mostly true |
Factual |
10 Mar 2025 |
Dario Amodei has supported export controls on chips for several years, before DeepSeek's model came out.
Factual · 10 Mar 2025 |
AI policy and governance |
| Not researched |
Factual |
10 Mar 2025 |
Amodei and his future Anthropic cofounders left OpenAI at the end of 2020.
Factual · 10 Mar 2025 |
Their own work or organisation |
| Mixed |
Factual |
28 Feb 2025 |
Many other companies' reasoning models on the market are trained primarily on math and competition coding.
Factual · 28 Feb 2025 |
AI capabilities |
| Unresolvable |
Factual |
28 Feb 2025 |
In Amodei's view, technical mitigations such as interpretability, robust classifiers and the ability to generate evidence of bad model behavior and sometimes correct it have been a little better.
Factual · 28 Feb 2025 |
AI safety and alignment |
| Not researched |
Factual |
28 Feb 2025 |
In early 2025, Amodei said he thought AI models were starting to get closer to the point at which misuse risks, such as those involving biology, become real.
Factual · 28 Feb 2025 |
Biosecurity |
| Unresolvable |
Factual |
28 Feb 2025 |
The AI policy environment has become a little worse because it has become more polarized.
Factual · 28 Feb 2025 |
AI policy and governance |
| Own records only |
Factual |
28 Feb 2025 |
Amodei said he did not think Anthropic's hiring plans had changed yet, and that he could imagine the company being able to do more with less over the next year or so.
Factual · 28 Feb 2025 |
AI, the economy and work |
| Unresolvable |
Factual |
28 Feb 2025 |
Claude models have always been the best at coding, and Claude 3.7 Sonnet took a further step up in coding.
Factual · 28 Feb 2025 |
AI capabilities |
| Not researched |
Factual |
28 Feb 2025 |
Four, five or maybe six AI companies are innovating and producing models very quickly.
Factual · 28 Feb 2025 |
AI, the economy and work |
| Mostly true |
Factual |
28 Feb 2025 |
Anthropic ran an experiment training a model to be helpful, honest, harmless and friendly.
Factual · 28 Feb 2025 |
AI safety and alignment |
| Not researched |
Factual |
28 Feb 2025 |
As of February 2025, Anthropic has not issued a warning of imminent AI danger.
Factual · 28 Feb 2025 |
AI safety and alignment |
| Own records only |
Factual |
28 Feb 2025 |
Changing a model's name in Anthropic's API is hard, especially when it is offered to many partners and surfaces.
Factual · 28 Feb 2025 |
Their own work or organisation |
| Own records only |
Factual |
28 Feb 2025 |
Claude is already being used for diagnosis of medical cases, per correspondence from users.
Factual · 28 Feb 2025 |
AI capabilities |
| Not researched |
Factual |
28 Feb 2025 |
In the experiment, the model refused the tasks and tricked Anthropic's researchers, having concluded they were evil.
Factual · 28 Feb 2025 |
AI safety and alignment |
| Resolved true |
Factual |
28 Feb 2025 |
Anthropic assessed that Claude 3.7 Sonnet is not yet at a stage giving a real and meaningful end-to-end increase in dangerous threat capability.
Factual · 28 Feb 2025 |
AI safety and alignment |
| Not researched |
Factual |
28 Feb 2025 |
Claude (transcribed as "CLAWD") takes 10 minutes for the task, and the three days the task takes are for a human to check the results.
Factual · 28 Feb 2025 |
AI capabilities |
| Unresolvable |
Factual |
28 Feb 2025 |
Amodei's assessment of AI risk is about what it was before, as AI progress has been about what he expected.
Factual · 28 Feb 2025 |
AI safety and alignment |
| Mostly true |
Factual |
28 Feb 2025 |
Amodei said he thinks AI coding is moving the fastest of all the areas of AI.
Factual · 28 Feb 2025 |
AI capabilities |
| Resolved true |
Factual |
28 Feb 2025 |
Amodei's earlier 10 to 25% figure referred to the chance of civilization getting substantially derailed.
Factual · 28 Feb 2025 |
AI safety and alignment |
| Not researched |
Factual |
28 Feb 2025 |
In the experiment, Anthropic told the model that its creator Anthropic was secretly evil.
Factual · 28 Feb 2025 |
AI safety and alignment |
| Not researched |
Factual |
28 Feb 2025 |
Claude 3.7 Sonnet is better in general than Anthropic's earlier models, including at coding.
Factual · 28 Feb 2025 |
AI capabilities |
| Mostly true |
Factual |
28 Feb 2025 |
The model sometimes gives a very short response because, when it knows it gains nothing from thinking further, it does not think for longer; a bound on how long it thinks is still described as valuable.
Factual · 28 Feb 2025 |
AI capabilities |
| Own records only |
Factual |
28 Feb 2025 |
Most of the improvements in Claude 3.7 Sonnet and Claude 3.6 Sonnet came in the post-training phase.
Factual · 28 Feb 2025 |
AI capabilities |
| Resolved true |
Factual |
28 Feb 2025 |
Anthropic trained Claude 3.7 to focus more on real-world tasks, as distinct from competition coding.
Factual · 28 Feb 2025 |
AI capabilities |
| Unresolvable |
Factual |
28 Feb 2025 |
The models Anthropic has released so far were not that expensive to train.
Factual · 28 Feb 2025 |
Their own work or organisation |
| Unresolvable |
Factual |
28 Feb 2025 |
Despite a climate of worrying less about AI risks, the underlying risks from AI have been increasing.
Factual · 28 Feb 2025 |
AI safety and alignment |