Tech-Reader AI Digest for Tue May 12 2026
.jpeg)
Tech-Reader AI Digest
Tuesday, May 12, 2026
#AI #TechNews #Digest
Story 1: Altman Takes the Stand — "I Believe I Am an Honest and Trustworthy Businessperson"
What happened: Sam Altman took the witness stand Tuesday in Oakland — the moment three weeks of trial testimony had been building toward. He testified in a blue suit and tie, initially speaking quietly and taking long pauses, visibly aware of what awaited him in cross-examination. By the time proceedings broke for recess, he had found his footing.
On direct examination, Altman told the jury what he believes the company he built actually is. He acknowledged that OpenAI had used "creative ways to keep it going" — taking outside investment and creating the for-profit arm — but said the result has been "one of the largest charities" in history. He said he did not understand Musk's allegation that he and other OpenAI executives were stealing from a charity. "It feels difficult to even wrap my head around that framing," he testified.
On Musk: Altman pushed back against the notion that Musk actually cares about OpenAI. "Mr. Musk did try to kill it," he said, adding that Musk launched a competitor, tried to poach its talent, and alleged that he engaged in "business interference." That comment was later struck from the record by Judge Gonzalez Rogers.
On the founding: Altman testified that the co-founders felt no single person should control AGI, and that Musk was not a good fit for the company. He described a "particularly hair-raising moment" when other OpenAI co-founders asked Musk what would happen if he had majority control of OpenAI and then died. Musk responded, according to Altman: "I haven't thought about it a ton, but maybe control would pass to my children." Altman said Musk's departure from the board in 2018 was a morale boost for employees who did not like his "hardcore" approach.
Cross-examination was pointed. Musk's attorney Steven Molo pressed Altman about concerns raised by former associates — employees of startup Loopt, Anthropic CEO Dario Amodei, and former OpenAI board members Tasha McCauley and Sue Yoon — and referenced a New Yorker article titled "Sam Altman May Control Our Future — Can He Be Trusted?" Molo asked directly: "Are you completely trustworthy?" Altman said he did not want to speak for other people who had made claims about him.
When asked "Have you misled people with whom you do business?" Altman responded: "I believe I am an honest and trustworthy businessperson." The judge later stepped in to give the jury context as the questioning grew increasingly hostile.
Fortune noted that more than two weeks into the trial, neither of the tech titans has emerged as an overly sympathetic character — and that a Cornell University expert on AI policy observed that "this is not looking good for any of them" at a time when public perception of AI is already under pressure.
Closing arguments are expected Thursday. An advisory jury is expected to deliver its view on wrongdoing during the week of May 18. (Source: NBC News / NPR / Fortune / CNBC / CoinCentral)
Why it matters: Altman was the last major witness. He walked into a courtroom where five witnesses had already described his leadership style under oath — none of them Musk allies. His answer to the trust question was not a denial of the specific allegations. It was a statement of self-belief. The jury has now heard the full record from both sides. They will decide what they make of it.
Aaron's take — What the trial record actually shows is a founding team trying to build something that had never been built before, under financial pressure that nobody fully anticipated, inside a governance structure that was always going to strain under that weight. Altman may have managed information differently with different people. He may also have been navigating a room full of brilliant, competing egos while trying to keep a fragile organization alive long enough to matter. The jury will weigh the legal question. The historical question is more complicated — and more interesting.
Story 2: The Verdict Preview — What the Jury Is Actually Deciding
What happened: With closing arguments Thursday and jury deliberations beginning as early as next week, it is worth being precise about what this jury is actually being asked to decide — because it is narrower than three weeks of dramatic testimony might suggest.
The jury's verdict is merely advisory. Judge Gonzalez Rogers will make the final call on any remedies. The jury must decide whether Altman and Brockman's communications with Musk around the formation of OpenAI established a formal "charitable trust" — and whether Altman and Brockman subsequently violated that trust when they restructured OpenAI so that its nonprofit board no longer had sole control over its for-profit arm. They must also decide on Musk's allegations that Altman and Brockman unjustly enriched themselves as OpenAI re-oriented from a research-oriented lab to a commercial entity. Most legal analysts say Musk's case is weak and that he's likely to lose.
Worth noting: a federal judge dismissed Musk's fraud claims before trial began on April 26, leaving only the breach of charitable trust and unjust enrichment arguments before the jury — a narrowing that OpenAI called significant and that Musk's legal team characterized as routine pre-trial procedure.
Fortune noted that Musk's apparent dual purpose was to sow investor doubt about OpenAI's for-profit structure ahead of its IPO — and to surface embarrassing internal documents through the discovery process. So far it's not clear the litigation has had much impact on OpenAI's ability to raise money. OpenAI has held several successful funding rounds since Musk filed his suit, including an additional raise at an $852 billion valuation that closed in March.
The $150 billion damages claim — directed at OpenAI's nonprofit arm — goes to the judge, not the jury, in Phase 2 if liability is found. (Source: Fortune / CoinCentral / Colombia One)
Why it matters: Three weeks of testimony have produced the most complete documentary record of OpenAI's founding history ever assembled. That record exists regardless of the verdict. The legal question — did a charitable trust exist and was it violated — is specific, technical, and ultimately the judge's call. The historical question the trial has answered is much larger than what any jury will decide Thursday.
Aaron's take — The verdict is not the story. The story is the record. Whatever Judge Gonzalez Rogers decides on remedies, the sworn testimony of Sutskever, Murati, McCauley, Brockman, Nadella, and Altman is now permanent public record. Every future board member, every IPO investor, every historian of the AI era will read it. That is what the trial has produced — independent of who wins.
Story 3: Thinking Machines Lands Google Cloud Deal — Murati Builds While the Trial Talks
What happened: While the Oakland courthouse has dominated AI headlines for three weeks, Mira Murati has been quietly building. Thinking Machines Lab has signed a multibillion-dollar deal with Google Cloud for AI infrastructure powered by Nvidia's latest GB300 chips. The deal is valued in the single-digit billions and includes access to Google's latest AI systems alongside infrastructure services to support model training and deployment.
The deal provided some insight into what Thinking Machines is developing. Google noted it can support the startup's reinforcement learning workloads, which Tinker's architecture relies on. Thinking Machines is among the first Google Cloud customers to access GB300-powered systems, which offer a 2x improvement in training and serving speed compared to prior-generation GPUs.
The Google Cloud deal follows a March partnership with Nvidia — a multiyear agreement under which Thinking Machines will deploy at least a gigawatt of Nvidia's Vera Rubin systems to train its models. Nvidia made what the two companies called a "significant investment," though they declined to disclose the size of the stake.
The company has not been without turbulence. In January 2026, co-founder and CTO Barret Zoph departed — the split described as not amicable — and returned to OpenAI alongside co-founder Luke Metz. Soumith Chintala was named the new CTO. The company has also lost co-founder Andrew Tulloch, who left for Meta.
For context on where Thinking Machines sits relative to its peers: after founding Thinking Machines Lab in February 2025, Murati secured the largest seed funding round in history — $2 billion at a $12 billion post-money valuation. In October, Thinking Machines released its debut product, Tinker, an API for fine-tuning language models. Nvidia and Google Cloud both now as strategic partners and investors.
Also worth adding to the standing Quick Hits roster going forward — Thinking Machines is now generating consistent digest-worthy news. (Source: TechCrunch / Axios / CNBC / Wikipedia)
Why it matters: Murati left OpenAI in late 2024, raised the largest seed round in history, shipped a product inside eight months, and has now secured infrastructure partnerships with both Nvidia and Google Cloud — all while managing co-founder departures that would have derailed most startups. The contrast with Safe Superintelligence — which has raised $3 billion and shipped nothing — is instructive. Murati is building. The compute is real. The product is live. The next question is what the frontier models she's training actually do.
Aaron's take — Murati testified in Week 2 about what she witnessed inside OpenAI. Then she went back to work. The Google Cloud deal and the Nvidia partnership together represent a credible path to frontier-scale training — not a research concept, not a mission statement. A gigawatt of Vera Rubin compute and GB300 infrastructure is the same tier of resource that Anthropic and OpenAI are working with. That is a meaningful development that got lost in three weeks of courtroom drama.
Quick Hits — The Rest of Today's AI World
Anthropic / Claude
- Agent View launched for Claude Code today — research preview. Single command-line dashboard for managing parallel coding sessions. Run via
claude agentsfrom the terminal. Available on Pro, Max, Team, Enterprise, and API plans. A quiet but significant step toward Claude Code as an agent operations layer rather than a chat-driven coding tool. (Source: Anthropic / TestingCatalog)
Gemini (Google)
- Google Cloud deepens Thinking Machines Lab partnership — see Story 3. No new model announcements today. (Source: TechCrunch)
VS Code / GitHub Copilot
- No new announcements. Token-based billing June 1 — 20 days remaining. (Source: GitHub)
Replit
- No new announcements.
Perplexity
- No new announcements today.
Microsoft Copilot
- No new announcements. Nadella testimony Monday — see yesterday's edition. (Source: CNBC)
Thinking Machines Lab
- Google Cloud multibillion-dollar deal confirmed — see Story 3. Now a standing Quick Hits entry. (Source: TechCrunch)
xAI / SpaceXAI
- No new announcements. Closing arguments Thursday. (Source: CNBC)
OpenAI
- Altman on the stand Tuesday — see Story 1. Closing arguments Thursday. Advisory jury deliberations week of May 18. (Source: NBC News / NPR)
Palantir
- No new announcements today.
Reflection AI
- No new announcements today.
Ollama
- No new announcements today.
DeepSeek
- V4-Pro and V4-Flash live since April 24. No new announcements today. (Source: DeepSeek)
Alibaba / Qwen / Z.ai
- No new announcements today.
Inflection Pi / Mistral
- No major news today.
That's your AI world for Tuesday, May 12. Closing arguments Thursday. Advisory jury deliberations week of May 18. Back tomorrow. — Aaron
Aaron Rose is a software engineer and technology writer at tech-reader.blog.
Catch up on the latest explainer videos, podcasts, and industry discussions below.
