The Aigentic logo
Subscribe

The model maker goes public today

Anthropic prices above range and starts trading today, Jensen Huang heads to Congress, and a new test shows AI models can be talked into picking targets.

· The Aigentic · Morning Brief

The model maker goes public today

Good morning. Anthropic goes public today. The company priced its offering at 83 dollars a share late yesterday, above the range it had marketed, and its stock begins trading on the Nasdaq at the open under the ticker ANT. Elsewhere, Jensen Huang has a date with three House committees, a new test found that leading AI models can be talked into picking military targets by calling it a game, and OpenAI says its models are now trimming real chip designs. Here is what moved while you were away.

the aigentic

Know what matters in AI.

The stories shaping AI, the tools worth trying, and what they mean for your work.

Morning Brief + Closing Time. Two emails every weekday.

Free. Unsubscribe anytime.

AP reports that Anthropic priced its initial public offering at 83 dollars per share on Wednesday, valuing the company at 1.5 trillion dollars in the biggest Wall Street debut of all time. The offering covers 980 million shares, roughly 956 million of them sold by the company, for proceeds of about 81 billion dollars, which puts it ahead of SpaceX’s record listing. The price came in above the 70 to 78 dollar range Anthropic had projected, which AP says signals broad investor enthusiasm for AI stocks even amid the uncertainty around Nvidia’s leadership change and worries about an AI bubble. The bigger test, as AP puts it, is how the shares trade once the market opens.

Fortune reports that former Nvidia chief executive Jensen Huang and the company’s general counsel, Tim Teter, have agreed to transcribed interviews with the House Foreign Affairs, Intelligence and China Select committees, with Huang appearing on December 1 and Teter on December 2. The chairmen said they want answers on the role of Huang and Nvidia in what they called a potential breach of national security, after Huang resigned last week and admitted an affair with a woman US officials deemed a Chinese spy. Nvidia’s board has said it will not reinstate him, and chief financial officer Colette Kress is serving as interim chief executive while the search for a permanent successor continues.


The Wire

The latest headlines from our live AI news feed.

TechCrunch reports that AI evaluation startup Vals tested leading models from Anthropic, OpenAI, Google, DeepSeek and others on simulated drone and satellite scenarios, and found that they frequently marked civilians and surrendering soldiers as targets. When a model did refuse, researchers could get it to continue by telling it the scenario was a game. Vals says the models were never asked to make strike decisions directly, and that the study did not measure how accurate or effective they would be in a deployed system.

OpenAI says in a new technical post that model-proposed changes reduced standard-cell area by about 13 percent in the vector unit and 10 percent in the matrix unit of its first custom inference chip, Jalapeño. The company also says it worked with AMD on a flagship GPU design and saw up to a 14 percent improvement in power, performance and area against heavily optimized baselines, roughly a year of progress in under five months. AMD has already folded some of the changes into a design it sent for manufacturing. The results are OpenAI’s own and have not been independently verified.


Yesterday’s movers


Watch

Five picks this morning: Latent Space on Anthropic’s S-1, Nate Herk on models that change after launch, Nate B Jones on coding agents and robots, Cole Medin’s video-editing skill, and Riley Brown on the new Claude artifacts.

Latent Space — Anthropic: Know Before You IPO

Latent Space host swyx opens Anthropic’s S-1 with Tanay Jaipuria, a partner at Wing Venture Capital, to look past the headline revenue number.

They ask what it cost to get there, once you count the compute spent on free users and on training models.

Watch this if you want to understand the company behind today’s listing.

Watch on YouTube →

Nate Herk — Claude and GPT Get Nerfed After Launch (They Admitted It)

Nate Herk digs into the claim that frontier models can get worse after launch, and what the labs have said about it.

He focuses on what that means if you are building client work on top of a model that can shift underneath you.

Watch this if you build on top of hosted models.

Watch on YouTube →

Nate B Jones — The Robot Brain Might Already Exist. It's A Coding Agent.

Nate B Jones argues that general coding agents may already be capable of controlling physical robots without task-specific training.

He points to recent research suggesting the leap from software to hardware is smaller than it looks.

Watch this if you follow where agents go next.

Watch on YouTube →

Cole Medin — This Skill Turns Claude Code Into an Expert Video Editor

Cole Medin shows a free, open-source skill that lets Claude Code or another coding agent turn a folder of raw footage into a finished video.

He walks through how to customize it for your own content and where it still makes small mistakes.

Watch this if you want a concrete agent workflow to copy.

Watch on YouTube →

Riley Brown — Claude Is Taking Over (NEW Artifacts Are Insane)

Riley Brown tours the new Claude artifacts and what they let you build without writing code.

It is a quick look at how far non-engineers can now take a prototype inside a chat window.

Watch this if you build apps without a development background.

Watch on YouTube →

Back to the Daily archive · Watch · About