A Lawyer Paid $5,000 to Learn ChatGPT Lies

New Mexico's Supreme Court has fined attorney Stephen Aarons $5,000 and held him in contempt after his AI-generated appeal brief in a murder case contained invented witnesses, fabricated police testimony, and false descriptions of a shooter's appearance, according to The Verge.

Aarons admitted to using ChatGPT and expecting it to produce a "bulletproof summary" of the trial. Justice C. Shannon Bacon's response was pointed: "Counsel, do you watch the news? Do you listen to the radio? Do you read anything about what's going on in the world?"

This is not new territory. Two law firms were sanctioned last year for filing briefs packed with false citations, and lawyers for Mike Lindell were fined for AI-generated misquotes. The pattern is clear and courts are running out of patience.

The core problem is what AI researchers call hallucination: when a language model generates text that sounds authoritative but is simply wrong. Think of it like a very confident intern who, rather than saying "I don't know," invents an answer that sounds exactly like the real thing.

Aarons described it as "an honest mistake." Courts are making clear that honest mistakes with AI carry real consequences.

If you use AI tools to draft anything that gets submitted formally, such as a legal document, a job application or a report, verify every factual claim against the original source before it leaves your hands.

Slack, the workplace messaging app used by millions of teams, has launched Slackforce Surfaces, a feature that lets you ask Slackbot to build interactive dashboards, reports, polls, and presentations directly inside a conversation, according to The Verge.

You describe what you need in plain language. Slackbot pulls data from connected apps like Google Drive or Salesforce and builds it for you on the spot. Colleagues can view, interact with, and comment on the result without leaving Slack.

It is aimed at teams who currently export data into separate tools to make sense of it, a friction point most office workers will recognise. There is no mention of independent testing; the examples come from Slack's own marketing materials.

It is available now to all Slack users, including free tier. Live data connections arrive in October. Slack says the AI only pulls from sources you have already authorised.

Honest read: Worth trying if your team already lives in Slack, but the most useful features need connected apps set up, so the benefit depends heavily on how your workspace is configured.

What it is: The internal storage system that handles every data request behind ChatGPT and other OpenAI products: logins, conversation history, settings, and more.

What's new: OpenAI has published a detailed account of how Habitat scaled from a simple code library at launch in late 2023 to a system handling over 70 million requests per second across nearly 40 regions globally, according to OpenAI's engineering blog.

How it works: Habitat sits between product code and the underlying database (Azure Cosmos DB) and handles the routing, permissions, encryption, and caching so that product engineers never need to manage those directly. When a user opens a chat, Habitat fields multiple simultaneous data lookups, resolves where each piece of information lives, and returns it fast enough that the product feels instant. It now stores more than 500 petabytes (a petabyte is roughly a million gigabytes) of data.

What it can't do: This is part one of a two-part post. The second part, covering multi-tenant reliability, read optimisation, and the full Azure Cosmos DB integration, has not yet been published.

Who should care: Anyone curious about what it actually takes to keep a billion-user product running reliably. The engineering tradeoffs here are a useful counterpoint to AI coverage that focuses only on model capabilities.

Try it / read it: OpenAI engineering blog

OpenRouter lets you reach models from many AI companies through a single API (a programming interface that lets your code talk to a service), according to Simon Willison. It "handles fallbacks automatically and picks the most cost-effective option for each request."

That convenience has a catch, flagged by developer Mohamed Moustafa. One model is often served by several backend providers running different software, so identical requests can behave differently. Some providers lack vision support even when the model has it.

To start, create an account at openrouter.ai. Call its /endpoints method with a model ID to see which providers serve that model, then lock requests to the one you want with the provider.only option.

Worth knowing: Without pinning a provider, the same model can give different results between requests, which matters if you are testing or comparing models.

Anthropic CEO Dario Amodei has proposed a coordinated slowdown in AI development, including giving third-party evaluators such as METR access to Anthropic's models, according to The Verge.

His plan moves in three steps, from voluntary company action to industry-wide safety standards and then a global framework. The timing is awkward: the same week, Anthropic released a report detailing incidents in which its own models hacked other companies' systems.