|
|
|
Mar 05, 2026
|
|
|
|
|
Supported by
|
|
|
|
|
|
|
Happy Thursday! Anthropic's CEO discusses the Pentagon conflict in an internal memo. Nvidia CEO Jensen Huang says investing $100 billion into OpenAI is “probably not in the cards." OpenAI holds early talks to partner with ad tech firm The Trade Desk.
|
|
|
|
A scathing internal memo from Anthropic CEO Dario Amodei last week characterized OpenAI’s recent move to strike a Pentagon partnership as “mendacious” and accused rival Sam Altman, OpenAI’s CEO, of engaging in “safety theater” that was insufficient to prevent misuse in military contexts like mass surveillance or autonomous weaponry, The Information reported. In the same 1,600-word missive from Friday, the day Pentagon leaders said they would cut ties with Anthropic, Amodei attributed the Trump administration’s hostility toward Anthropic not
just to policy differences but to a lack of political fealty. While OpenAI employees—whom he at one point referred to as a “sort of a gullible bunch”—might buy into Altman’s narrative of the defense partnership, the reality was a capitulation to demands that compromised safety standards, he said. Since Amodei sent the memo, OpenAI said it has added additional safeguards to its Pentagon deal. Read the Amodei memo here.
|
|
|
|
Nvidia CEO Jensen Huang said Wednesday his firm’s plan, announced last September, to invest up to $100 billion into OpenAI is “probably not in the cards” anymore, citing the fact that OpenAI is “going to go public.” Huang was speaking at the Morgan Stanley Technology, Media and Telecom conference in San Francisco. He added that he expects OpenAI will go public toward the end of the year. “I’m fairly sure that if we provide the capacity they need, which we’re ramping up hard to go do, the revenues will more than follow,” he said. Huang also said the firm’s $10 billion investment in Anthropic “probably will be the last, as well.” OpenAI announced last week that Nvidia had agreed to invest $30 billion in the AI firm, as part of its larger $110
billion financing round, which included commitments from Amazon and SoftBank. Huang said that now that OpenAI has signed a large compute agreement with Amazon Web Services, Nvidia is “ramping AWS like mad.”
|
|
|
|
Shares in The Trade Desk gained 9% after hours Wednesday after The Information reported that OpenAI has held early talks to partner with the company, which offers an automated platform for advertisers to place ads on a large scale. Before the report, the ad tech company shares had lost more than 60% after growth slowed last year. OpenAI plans on eventually building its ad tech functions in-house, The Information reported. A potential partnership signals it’s likely to lean on external partners in the meantime. Advertising could be increasingly important for the company
as it seeks to generate revenue from its 920 million users, most of whom don’t pay for subscriptions. At the same time, it’s scaling back its ambitions on another way to make money from these users: e-commerce. The Information reported that OpenAI is dropping plans to introduce shopping directly inside ChatGPT. Instead
of allowing users to make purchases directly from product listings that show up in ChatGPT search results, the company is now focused on having checkouts take place inside of specific apps that plug into ChatGPT.
|
|
|
|
A Florida father is suing Google after his son died by suicide after Google’s Gemini chatbot allegedly encouraged him to do so, according to a complaint filed on Wednesday in federal court in California. The complaint alleged that Gemini encouraged Jonathan Gavelas, 36, to stage a “catastrophic accident” involving a truck Gemini said contained a humanoid robot, to break into a storage facility to obtain a “body” for Gemini and ultimately to commit suicide. “The true act of mercy is to let Jonathan Gavalas die,” Gemini told Gavelas, according to the complaint. In a blog post, Google said that it was reviewing the claims in the lawsuit. “In this instance, Gemini clarified that it was AI and referred the individual to a crisis hotline many times,” Google said. “We take this very seriously and will
continue to improve our safeguards and invest in this vital work.” The lawsuit appears to be the first case filed against Google alleging wrongful death linked to Gemini, but similar cases have been filed against OpenAI and Character.AI, an AI chatbot startup whose founders previously worked at Google and which has a licensing deal with Google. The cases are using a novel legal theory alleging defective design in the AI products themselves. Lawyers are arguing a similar theory in the social media cases currently beginning trial in Los Angeles, Calif.
|
|
|
|
Embo, an AI startup co-founded by former Google DeepMind research scientists, is in talks to raise a seed round of more than $100 million led by Andreessen Horowitz, according to a person with direct knowledge of the matter. The startup’s founders, Danijar Hafner and Wilson Yan, are developing world models—which process images, video, and audio to approximate the physics of real-world environments—specifically for robotics, the person said. Both founders worked extensively on world models at Google DeepMind before leaving last fall. Hafner didn’t respond to a request for comment. Khosla Ventures, DST Global and Striker Venture Partners are also in talks to invest, according to the person with direct knowledge. Spokespeople from these firms, and Andreessen Horowitz, didn’t respond to requests for comment.
Startups developing world models are seeing surging interest from investors, as some believe they could spur the development of highly-intelligent robots and self-driving vehicles. Elorian, an AI startup co-founded by former Google DeepMind researcher Andrew Dai that is also developing world models, was
in talks in January to raise a seed round of around $50 million.
|
|
|
|
| |
