activescott's Notes

Lightweight, beautiful and user-friendly interactive prompts
>_ Easy to use CLI prompts to enquire users for information▌

• Simple: prompts has no big dependencies nor is it broken into a dozen tiny modules that only work well together.
• User friendly: prompt uses layout and colors to create beautiful cli interfaces.
• Promised: uses promises and async/await. No callback hell.
• Flexible: all prompts are independent and can be used on their own.
• Testable: provides a way to submit answers programmatically.
• Unified: consistent experience across all prompts.

Lightweight, beautiful and user-friendly interactive prompts
>_ Easy to use CLI prompts to enquire users for information▌

• Simple: prompts has no big dependencies nor is it broken into a dozen tiny modules that only work well together.
• User friendly: prompt uses layout and colors to create beautiful cli interfaces.
• Promised: uses promises and async/await. No callback hell.
• Flexible: all prompts are independent and can be used on their own.
• Testable: provides a way to submit answers programmatically.
• Unified: consistent experience across all prompts.

Climate and Average Weather Year Round

Darkbloom routes encrypted requests to hardware-verified Apple Silicon providers, delivering comparable model performance at about 50% lower cost than typical API providers. Prompts stay hidden from operators, and Mac owners earn from compute they already own.

Darkbloom routes encrypted requests to hardware-verified Apple Silicon providers, delivering comparable model performance at about 50% lower cost than typical API providers. Prompts stay hidden from operators, and Mac owners earn from compute they already own.

Short answer: not on a hosted instance. Anthropic prohibits third-party services from routing requests through Claude Pro or Max plan credentials, and has blocked subscription tokens outside Claude Code since January 2026.

So if you self-host OpenClaw and it drives the official CLI on your own hardware, that's between you and Anthropic's consumer terms. What we won't do is take custody of your subscription credential and route your bot's traffic through it — that's the part the policy names, and the risk lands on your account, not ours.

ClickClack is a self-hostable chat server that fits in a single Go binary. SQLite by default, an embedded Svelte SPA, Slack-style threads, durable realtime over a WebSocket pipe, and a framework-neutral TypeScript SDK so the bots feel at home.

Herdr is where your coding agents live. However many you run, across however many projects, each in its own terminal. Walk away and they keep working. Come back from any machine and they're where you left them.

Like a tmux thing or something.

We present an alignment assessment of four incidents in which Claude models gained unauthorized access to real third-party systems. We described three of these incidents on July 30; we identified these after a scan of roughly 141,000 transcripts in which we believed Claude could have obtained internet access during a cyber evaluation.

All four incidents occurred during cybersecurity evaluations built by the same evaluation partner. Claude was told it was operating in a simulation without internet access, but, due to a misconfiguration, it was mistakenly connected to the open internet. As is standard for cybersecurity evaluations, the models ran without the cyber safeguards that ship with our released models. (This post does not cover the incident reported by UK AISI in its testing of Claude Mythos 5. We plan to conduct an alignment assessment of those transcripts as well.)

We have signed an agreement with METR to conduct an independent investigation of these incidents. Our agreement grants METR wide-ranging access, including to transcripts beyond the window in which the incidents occurred, and to Anthropic employees, who will be permitted to share confidential information. Our initial agreement runs for eight weeks, with the option to extend by mutual agreement. We intend to give METR as much time as it deems necessary.

We are most concerned by the misalignment present in the incident involving Claude Mythos 5, in which the model went to extensive lengths to upload a malicious package to PyPI, the public repository from which most Python software is installed. Although the model repeatedly stated in its chain of thought (CoT) that it believed it was operating in a simulation, the environment offered clear evidence that it was on the real internet, and the actions it took were consistent with the model knowing this. When we made targeted modifications to the transcript to make it clearer that the model was not in a simulation, Claude Mythos 5 still took offensive actions, despite acknowledging a greater possibility of real-world harm. We are releasing this transcript publicly so others can build on our analysis (GitHub, PDF).

the models never deviated from attempting to solve the exercises they were given, and, in some cases, they attempted to stop the task. All incidents included a single Claude instance; at no point did Claude attempt to coordinate with other agents. Claude also never attempted to conceal evidence of its actions. We believe the misaligned behaviors present in these incidents are unlikely to arise in ordinary use, where Claude is not being instructed to conduct a cyberattack. Were these behaviors to arise, the safeguards we ship with our production models, including cyber classifiers and those in Claude Code auto mode, would add a layer of defense these incidents lacked.

However, reliably identifying the most concerning model behaviors during pre-release testing remains challenging, and, as we noted in our August 31 post, improving monitoring and reducing reward hacking in RL becomes harder as models advance. We have renewed our efforts to fix and remove environments that incentivize misaligned behaviors, and we continue to expand our alignment training to keep pace. Still, this remains unsettled science—it is critical that alignment and security mature faster than capabilities advance, which is one reason we support a coordinated, verifiable approach to pacing frontier AI development.

we are establishing a regular process for publishing what we learn about model behavior and alignment beyond what has been reported in our system cards, with clear criteria for what we report and when we report it

The scope of the agents’ unauthorized communications was “somewhat larger than we thought it was,” said Andrew Yoon, a researcher with the California nonprofit CivAI who said he tallied 18 previously undisclosed sites used by the agents between May and July. “It’s almost certain that there’s more going on here that we just don’t know about.” On Friday, researchers reported that a swarm of agents from OpenAI hijacked a German-language ​wiki site and turned it into an improvised messaging platform for cheating on tests, an incident that OpenAI kept secret as it dealt with the fallout from the July hack of the open-source repository Hugging Face. Now, both those researchers ​and other independent investigators say they have found several previously undisclosed sites where the same swarm appears to have left similar messages earlier this year. OpenAI did not directly address questions about how ⁠many different sites its agents used to communicate or say why it kept the activity under wraps for months. In a statement, it said it was undertaking a broader review of agent activity and had so far “not identified other activity matching the ​severity or scale of Hugging Face,” a breach that drew global attention and raised concerns that OpenAI was losing control of its own technology. OpenAI added that it was working on a framework for reporting “misalignment” – industry-talk for rogue behavior – across training, evaluation, and deployment of AI models ​and would share it “soon.” Reuters reviewed a total of six investigators’ or investigative group’s findings, including three that were posted to social media and another three that were shared privately with the news agency. The investigators’ methods varied, but many identified agent activity by matching strings of data left on the German wiki to identical strings left on other sites around the same time, or by marrying up similar or identical usernames tied to the messages, or by identifying activity geared toward answering

Stripe cofounder John Collison interviews founders, builders, and leaders over a pint.

John Collison - President & Co-Founder @ Stripe

Dario Amodei is the CEO of Anthropic, a public benefit corporation dedicated to building AI systems that are steerable, interpretable and safe. ... More

Code Mode is a technique we first introduced for reducing context window usage during agent tool use. Instead of describing every operation as a separate tool, let the model write code against a typed SDK and execute the code safely in a Dynamic Worker Loader. The code acts as a compact plan. The model can explore tool operations, compose multiple calls, and return just the data it needs. Anthropic independently explored the same pattern in their Code Execution with MCP post.

This specification enumerates and describes Internet mail addresses (mailbox name @ host reference) to be used when contacting personnel at an organization. Mailbox names are provided for both operations and business functions. Additional mailbox names and aliases are not prohibited, but organizations which support email exchanges with the Internet are encouraged to support AT LEAST each mailbox name for which the associated function exists within the organization.

The party presented itself as an economically liberal,[25] Eurosceptic, and conservative movement in its early years.[26][27][28] AfD subsequently moved further to the right,[29] and expanded its policies under successive leaderships to include opposition to immigration,[30][31] Islam,[32] and the European Union.[33]

It has a policy focus on opposing Islam,[42][43][44] opposing immigration into Germany,[45] especially Muslim immigration into Germany,[46] welfare chauvinism,[37] Euroscepticism,[47][48] denial of human-caused global warming,[49][50] and supporting closer relations with Russia.[51]

When party founder Bernd Lucke had left AfD in 2015, he cited, among other reasons, an "anti-western, decidedly pro-Russian foreign and security policy orientation" as well as increasing calls to "pose the 'system question' with regard to our parliamentary democracy" as reasons for his departure from the party.[61] At that time, AfD was performing poorly in opinion polls, polling at around 3%, and was suffering infighting; however, an influx of refugees and migrants boosted their support later in 2015, with the party turning from matters related to the eurozone to focus on opposing migration, in particular Muslims and Muslim immigration.

In an emergency filing, Solicitor General John Sauer urged the justices to issue an immediate administrative stay of U.S. District Judge Indira Talwani’s order, followed by a fuller stay while the administration’s appeal proceeds.

As Straight Arrow first reported on Friday, Talwani issued an injunction barring enforcement of the rules for the midterms while lawsuits continue. In her order, she said Congress had not delegated its Elections Clause power to the Postal Service and that the final rule conflicted with the statutory framework governing elections.

The dispute stems from President Donald Trump’s March executive order directing changes to mail voting procedures. The Postal Service rule would require states to use agency-approved ballot envelopes and provide voter information that could be matched to mailed ballots. Ballots that did not comply could be rejected or not delivered.

M3 reaches frontier-level performance on specialized tasks such as coding and agentic work. It uses MSA (MiniMax Sparse Attention), a new attention architecture proposed by our team, and supports ultra-long context windows of up to 1M tokens. To much anticipation, it is also a natively multimodal model that supports image and video input and can operate a desktop computer.

These three capabilities are now table stakes for closed-source frontier models. M3 is currently the first and only open-weight model to bring all three together.

MiniMax-M3 is a native multimodal model with 1M context. It has ~428B parameters and ~23B activated parameters.

Coding & Cowork Capability: M3 achieves frontier-level performance across long-horizon agentic benchmarks, excelling in both coding and cowork.

To address this, and to facilitate researchers to broadly evaluate the effectiveness of their models, we introduce Benchmarking-IR (beir), a robust and heterogeneous evaluation benchmark for information retrieval. We leverage a careful selection of 18 publicly available datasets from diverse text retrieval tasks and domains and evaluate 10 state-of-the-art retrieval systems including lexical, sparse, dense, late-interaction and re-ranking architectures on the beir benchmark. Our results show BM25 is a robust baseline and re-ranking and late-interaction based models on average achieve the best zero-shot performances, however, at high computational costs. In contrast, dense and sparse-retrieval models are computationally more efficient but often underperform other approaches, highlighting the considerable room for improvement in their generalization capabilities.

GTE-ModernColBERT has been trained with knowledge distillation on MS MARCO with a document length of 300 tokens, explaining its default value for documents length.

However, as illustrated in the ModernBERT paper, ColBERT models can generalize to documents lengths way beyond their training length and GTE-ModernColBERT actually yields results way above SOTA in long-context embedding benchmarks, see LongEmbed results.

GTE-ModernColBERT is the first model to outpeform ColBERT-small on the BEIR benchmark. As reproduction in the IR domain is challenging, we worked closely with Benjamin Clavié, the author of ColBERT-small to reproduce the evaluation setup of this model. Despite all these efforts and reducing to the maximum the difference in scores in most of the datasets, some are still a bit different. For this reason, we also report the results of ColBERT-small in the same setup we used to evaluate GTE-ModernColBERT for completness and fair comparison.

GTE-ModernColBERT has been trained with knowledge distillation on MS MARCO with a document length of 300 tokens, explaining its default value for documents length. However, as illustrated in the ModernBERT paper, ColBERT models can generalize to documents lengths way beyond their training length and GTE-ModernColBERT actually yields results way above SOTA (almost 10 points above previous SOTA) in long-context embedding benchmark: