AI Evening Wrap — April 29, 2026

📖 2 min read

Happy hump day, AI watchers. Here’s what went down while you were grinding through your Tuesday (or sleeping through your Wednesday, depending on your timezone).

Musk Takes the Stand in OpenAI Trial

The Musk vs. Altman courtroom drama officially kicked off in Oakland federal court today. Elon took the witness stand and didn’t hold back, accusing OpenAI executives of “looting a charity” by converting the nonprofit into a for-profit company. A jury has been seated and this trial is expected to run for weeks. Whatever your take on Musk, this case could set real precedent for how AI nonprofits handle governance and billion-dollar pivots. Expect fireworks.

Google Signs Pentagon AI Deal After Anthropic Said No

In a move that would’ve been unthinkable during the 2018 Project Maven protests, Google has signed a new contract to supply the Pentagon with Gemini AI for classified work. This comes directly after Anthropic refused to let the DoD use Claude for mass surveillance and autonomous weapons. Hundreds of Google employees have signed a letter urging CEO Pichai to reject military AI use, but the deal is done. The Pentagon’s AI chief told CNBC that relying on a single model is “never a good thing” – so expect more providers to get pulled in.

Critical Vulnerability Found in Hugging Face’s LeRobot

Security researchers disclosed CVE-2026-25874, a CVSS 9.8 critical flaw in Hugging Face’s open-source robotics platform LeRobot. The bug allows unauthenticated remote code execution through unsafe pickle deserialization. As AI moves into physical robotics systems, this is exactly the kind of vulnerability that keeps security teams up at night. If you’re running LeRobot in any capacity, patch immediately.

Mag 7 Earnings Avalanche Hits Today

Alphabet, Microsoft, Amazon, and Meta all report earnings today (April 29). Amazon’s $200 billion AI capex plan is under the microscope, and AWS just posted its fastest growth in 13 quarters. Markets dipped on AI spending concerns earlier, but if these reports show real AI revenue – not just promises – expect the narrative to shift fast. Apple follows tomorrow.

Hot Take from the Trenches

The Google-Pentagon deal is peak 2026 irony. The same company whose employees successfully killed Project Maven in 2018 just signed up for classified military AI work. The difference? Back then, Google was the only game in town. Now Anthropic’s refusal created a vacuum, and Google filled it before anyone could blink. The “don’t be evil” crowd is going to have a field day with this one.

That’s your evening wrap. For deeper dives on AI investing angles, check out BetOnAI.net. Looking for the best AI tools? Browse our picks at AiToolCrush.com.

See you tomorrow.