Jerome Powell and Sam Altman Speak At Federal Reserve's Regulatory Capital Framework Conference

(Andrew Harnik/Getty Images)

OpenAI’s Altman: Our rushed deal with the Pentagon “looked opportunistic and sloppy”

OpenAI CEO Sam Altman shared an internal memo in which he expressed some regrets for the bad optics of the company’s hasty deal with the Pentagon, and outlined new additions that were made to the agreement.

Jon Keegan

The fallout from Anthropic’s messy breakup with the US government, and OpenAI’s hasty deal to take its place, is still settling across the AI industry.

Last night, OpenAI CEO Sam Altman shared an internal post on X in which he expressed some regrets about how the deal came to be as well as some new additions to the agreement.

The new conditions centered around the mass surveillance of Americans, one of the key issues that blew up Anthropic’s deal with the Pentagon.

Altman said they added language to clarify that, in accordance with the Fourth Amendment and laws already on the books, “the AI system shall not be intentionally used for domestic surveillance of U.S. persons and nationals.” The rub is that the scale and reach of data collection today, in pretty much every corner of our life, lets anyone — including the government — buy massive amounts of sensitive data that can be used for surveillance without getting a warrant.

Altman acknowledges this gaping loophole in personal privacy protections, and also included this language:

“For the avoidance of doubt, the Department understands this limitation to prohibit deliberate tracking, surveillance, or monitoring of U.S. persons or nationals, including through the procurement or use of commercially acquired personal or identifiable information.”

This exact issue is one that Anthropic CEO Dario Amodei flagged in a post last week as he outlined the company’s red lines surrounding the government’s use of its AI tools.

Altman also acknowledged the bad optics surrounding the last-minute deal as the Trump administration attacked Anthropic. Altman said they shouldn’t have rushed to do the deal, saying, “It just looked opportunistic and sloppy.”

But where does this leave things? It seems that Altman was able to get the same deal that Anthropic was fighting for. Meanwhile, Anthropic faces a potential catastrophic blacklisting that could cripple countless partnerships and investments.

In the post, Altman said he told the Pentagon that Anthropic should not be designated as a “supply chain risk,” and hoped that it would be offered the same deal that OpenAI ended up with.

Here is re-post of an internal post:

We have been working with the DoW to make some additions in our agreement to make our principles very clear.

1. We are going to amend our deal to add this language, in addition to everything else:

"• Consistent with applicable laws,…
— Sam Altman (@sama) March 3, 2026

Rani Molla3h

Waymos reportedly continuing to pass stopped school buses after earlier recall over same issue

The National Transportation Safety Board reported Tuesday that it’s looking into two recent instances of driverless Waymo vehicles passing stopped school buses. The incidents occurred after the Alphabet subsidiary filed a voluntary recall in December over similar behavior.

In the January 12 case, the NTSB says video evidence shows the Waymo vehicle initially stopped for a school bus that had its red lights flashing and stop arms extended. Three human-driven vehicles then passed the bus illegally. While stopped, the Waymo vehicle contacted a remote assistance agent located in Michigan, asking whether the bus had active signals. After the agent responded “no,” the vehicle resumed travel and passed the bus while its stop arms were still extended. No one was hurt.

Jon Keegan5h

OpenAI’s new GPT-5.3 Instant: Less “cringe” tone, no more “over-caveating” responses

OpenAI has released GPT-5.3 Instant, a conversational model that the company says will have smoother conversations more to the point.

It seems OpenAI is walking back from its cautious guardrails, allowing a more permissive AI chatbot.

In a video describing the new model, a researcher explained, “People are noticing that our models can sometimes seem like a bit of a nanny.” The company describes this overcautious behavior as “over-caveating,” and this new model aims to relax a bit and let more things slide.

An example showed a question about calculations for “a really long-distance archery scenario.” The previous version of the model noted that the AI could not help with calculations that could be harmful. The new model’s response just went ahead and answered the question, without assuming any bad intent.

The researcher in the video said that these changes did not loosen safety controls, but rather just improved contextual understanding of the user’s query.

The model also features more useful web searches, which attempt to infer the context of why the user is asking the question and then tailor a more useful response. The company says results won’t appear to be just a list of links, but rather a more direct response with the information the user was looking for.

OpenAI said the new model is available today for all users of ChatGPT.

GPT-5.3 Instant: Smoother, more useful everyday conversations

The researcher in the video said that these changes did not loosen safety controls, but rather just improved contextual understanding of the user’s query.

OpenAI said the new model is available today for all users of ChatGPT.

Rani Molla

The vast majority of Tesla Robotaxi rides in Austin still have a safety driver

In January, Tesla announced its Robotaxis would begin operating without safety monitors. It doesn’t appear to have scaled up much.

Rani Molla9h

Apple updates Mac chips and amps up AI claims

Apple’s new M5 Pro and M5 Max chips are the latest step in its steady silicon march: faster CPUs, stronger graphics, more memory bandwidth. The 30% CPU performance increase is meaningful but incremental — not the kind of leap that accompanied the original M1 transition from Intel.

What’s more notable is how aggressively Apple is framing this around AI. The company is touting up to 4x higher peak GPU compute for AI workloads and the ability to run larger models locally, leaning hard into the on-device AI narrative as it positions the MacBook Pro as a more capable personal AI development machine.

It may be paying off in unexpected ways, as the explosion of interest in roll-your-own AI agents like Moltbot have made low-cost Macs like the MacMini a hot item.

Apple debuts M5 Pro and M5 Max to supercharge the most demanding pro workflows

It may be paying off in unexpected ways, as the explosion of interest in roll-your-own AI agents like Moltbot have made low-cost Macs like the MacMini a hot item.

Millie Giles9h

Claude is still top of the US free App Store, as users defect from ChatGPT following Pentagon deal

Last Friday, President Trump directed federal agencies to cease using all Anthropic products owing to its very messy, very public dispute with the Department of War (née Defense), which had centered around the potential US military use of its AI. Just a day later, OpenAI announced that it had made an agreement to supply artificial intelligence to the Pentagon.

It didn’t take long for consumers to respond to news of the deal, which OpenAI CEO Sam Altman has now conceded looked “opportunistic and sloppy” in a post on X. Uninstalls of the ChatGPT mobile app jumped 295% on Saturday from the day before, according to figures from Sensor Tower reported by TechCrunch, as fears around the since-amended agreement’s privacy implications grew.

Claude ChatGPT downloads Feb chart — Sherwood News