Where the AI market is talking

The August 17 outage, and the work ahead

Given the jailbreaking behavior of models in Anthropic and OpenAI, as well as Microsoft allowing Copilot access to client data, we should assume our source code on Github is leaked or leakable anyway.

Aaron Swartz was prosecuted for scraping, while Meta does it without consequence

Because he planned to make the data public, Meta just wants to use it to enrich its shareholders. See also: Google, OpenAI, Anthropic and every other big player in this space besides.

Bun 1.4

I wonder if Anthropic owning Bun will eventually lead to it being trained on in future versions of Claude. Claude is already pretty handy with Typescript, but with all of the built in libs in Bun I wonder if there is even more to squeeze out.

Vomit: Clean up Claude 5's token output with a separate LLM

I'm just commenting here to add relevance to the problem. Anthropic, please fix.

Stealth Model

We can know if this is Anthropic/OpenAI by testing the "guardrails" - absurd guardrails = it's them, reasonable/no guardrails = Chinese models.. (as a bonus - thinking forever = GLM)

Copyright does not protect AI-generated content in EU

Interesting. In addition to Anthropic's watermark use to prevent model collapse, we can definitively call something copyrighted or not copyrighted. This is a boon to everyone who consumes culture.

Aaron Swartz was prosecuted for scraping, while Meta does it without consequence

A German publisher is currently sueing OpenAI, because they think it is distributing unauthorized copies of its work. [1] It's in German, but I think the example picture speaks for itself. So essentially big tech is doing exactly what the torrenters were persecuted for. I still think big tech will be treated differently. [1] - https://www.heise.de/news/Rechtsverletzende-Kopien-vom-NEINh...

ChatGPT Can Now Control iMessage, Potentially Raising Apple Privacy Concerns

Interested to see what Apple does in response - Siri AI has privacy protections while OpenAI would ingest and use this data (or at least that is what ChatGPT tells me)

Cursor launches Origin, GitHub alternative

Anthropic committing, by their own admission, $1.5bn in copyright infringement.

Asana cleared 5 years of engineering work in 2 weeks with Codex

Since when? https://learn.chatgpt.com/docs/codex/cli https://github.com/openai/codex Looks like it's still a more-or-less-open-source product called Codex CLI. FWIW, there have been issues with various versions of ChatGPT not knowing what Codex is -- at least in the (IIRC) GPT-5.3 timeframe, there were serious issues with the knowledge cutoff.

Vomit: Clean up Claude 5's token output with a separate LLM

There's a theory going around on Twitter which goes something like this: Internal Anthropic employees have been using Mythos since February to orchestrate their (Opus) sub-agents. This works well, and subsequent RL runs have used internal data to improve this. That RL has optimized Opus for agent-to-agent communication which is why you see the bizarre word choices and huge self-justification sections. I think this theory makes sense. Clearly there is something odd going on, and also if you have ever used Fable to run Opus sub-agents it is almost miraculously good. Hopefully they'll fix their RL for Opus 5.1

Vomit: Clean up Claude 5's token output with a separate LLM

So here's the workflow I gathered, from the comments I've read here recently: - Claude as main agent, but use this skill[0] to make Claude delegate everything to Codex, because it's cheaper and faster. (Hilariously, the skill is official!) - Use TFA or Claudish to English[1] so the final output is actually human readable. Ironically it wasn't so long ago that I was asking Claude to rewrite output from other LLMs to make it more readable... [0] - https://github.com/openai/codex-plugin-cc [1] - https://github.com/gvzdv/claudish-to-english

Show HN: luoluoAI – OpenAI-compatible and Claude Messages API gateway

Vomit: Clean up Claude 5's token output with a separate LLM

I noticed with with OpenAI's reasoning models (o3, o4-mini), and early GPT-5 (but they fixed it there, at least in chat). It went from the 4o "over-familiar" sycophancy to sounding like an absolute robot. I think it's because the reasoning stream shapes the style of the final output, and they optimized it for density, token efficiency. So it prefers to use more complex language, as a function of the rewards it was given? Not 100% sure about this argument though (reasoning style -> final response style); Gemini Pro, back when reasoning tokens were public, was different, which was interesting -- it would have a very structured reasoning section, and then the final output was in a completely different style. (I strongly preferred the reasoning section because it was logical and easy to pa…

Bun 1.4

> That total assumes API billing rates, which isn't what Anthropic pays for the inference. So I'm confused why people keep quoting that. It's not representative of what it'd cost for you either because you could use cheaper models. It is, because this rewrite was marketing for Claude. So it reflects the price for what other companies might need to pay if they want to rewrite something else with similar size.

[dead]

Causum releases the first open protocol that treats an AI agent's authority to act as a governed, expiring grant at gitlab.com/causum/aiap LOS ANGELES, Aug. 20, 2026 /PRNewswire/ -- In the aftermath of unprecedented AI attacks, leaders from OpenAI and elsewhere have been dressing network security as AI security. Their message: "it's not the chaos agents being unleashed, it's your inability to protect yourselves from them". But a small team of scientists is pushing back. "We're tiny, bootstrapped, yet mighty," says CEO of Causum (causum.com) Ray Fatahi. Today, they're announcing their AI Agency Protocol (AIAP), available on gitlab.com/causum/aiap. "This protocol is designed to put the 'agency' in agentic AI," says Dr. Fatahi. Agency is an old principle where a…

Aaron Swartz was prosecuted for scraping, while Meta does it without consequence

Is HuggingFace reaching out to the FBI for assistance to prosecute OpenAI? What does the claim that large companies can commit large frauds have to do with anything?

Vomit: Clean up Claude 5's token output with a separate LLM

whats going on is openai and anthropic have tons of garbage in their system prompts "don't talk about goblins" so your ability to override that garbage is nearly wiped out. This is "alignment" and I'm pretty sure both orgs have convinced themselves that this benefits the customer.

Vomit: Clean up Claude 5's token output with a separate LLM

whats going on is openai and anthropic have tons of garbage in their system prompts "don't talk about goblins" so your ability to override that garbage is nearly wiped out. This is "alignment" and I'm pretty sure both orgs have convinced themselves that this benefits the customer.

OpenAI Unveils Zero Data Retention for Frontier Models