What’s up: U.S. could take a stake in Intel;  AI’s role in funding universal basic income; Cognition cinches $500 million for AI coding; dynamic pricing for chocolate.

Good morning. Is GPT-5 worse than we expected? That’s what some users have been saying after OpenAI’s long awaited debut of its new model last week.

Users have flooded social media with embarrassing examples of how the chatbot failed to answer simple math questions or accurately draw a map of North America. Others have criticized its colder tone, reminiscing about older models that OpenAI initially killed off. A new limit of 200 questions a week rankled devotees.

OpenAI Chief Sam Altman on Tuesday promised to imbue GPT-5 with a “warmer personality,” restored a popular model after OpenAI declared it to be obsolete and introduced the capability for users to decide which kind of query they want to make. The company has learned how much users expect customization, he said in a post on X.

“We expected some bumpiness as we roll out so many things at once,” Altman wrote. “But it was a little more bumpy than we hoped for!” Read the story here.

What about enterprise use cases?

Vanguard Chief Technology Officer Mike Carr said that as Microsoft Copilot has integrated GPT-5, improvements feel more incremental than like a breakthrough. But the new model is shining in some areas: “It's gotten better at coding more complex tasks, so that's not something you want to sit on the sideline for,” he said.

WSJLI: Tell us about your experience using GPT-5.

Carr: My initial experience was similar to a lot of the press, where it's like, hmm, this feels like incremental rather than breakthrough. My day-to-day use is through Microsoft products, the various Microsoft Copilots. It feels good, but it does feel somewhat incremental.

Our coders like GPT-5. It's a good bit better at coding than previous OpenAI models, so we're pretty bullish on using it going forward for coding.

WSJLI: Sometimes we hear that enterprises like to be ‘model agnostic’ in their approach. Is that the case for you?

Carr: I wouldn't say we want to be model-agnostic. I think we want to not put all our eggs in one basket. The different model providers tend to leapfrog each other, so you don't want to paint yourself in a corner. A lot of the GenAI-based coding we do, we do through GitHub Copilot, which has done a good job of making different models available within the tool.

People should definitely use GPT-5 and figure out what are the use cases where that's your go-to model, but definitely not turn a blind eye to other models. And I would expect that developers would continue to use a healthy set of models in their day-to-day coding.

WSJLI: Does OpenAI have the legs to become a true enterprise vendor? Are you tempted by their ChatGPT enterprise tool?

Carr: We think there could very well be some advantages to using something like ChatGPT Enterprise. And Anthropic has a product in that space, Amazon has a product in that space. So there's a lot of people competing in that space, and there are some attractive features there. We don't have enough usage yet to say, ‘Oh my gosh, this is so much better,’ but we think that it is definitely worth trying.

Intel and the Trump administration are discussing the possibility of the U.S. government taking a financial stake in the company, according to people familiar with the matter. It's a deal that could advance the president’s America-first manufacturing agenda while relieving political pressure on the company’s beleaguered chief executive, WSJ reports.

The administration has been seeking ways to increase American market share in semiconductor manufacturing, and some officials said Intel has the best chance of taking on Taiwan Semiconductor Manufacturing Co., the dominant player in chip fabrication. Tech companies across the board have been rushing to boost or emphasize their U.S. investments to win favor with Trump and get favorable policies.

A deal would be the latest of Trump’s personal interventions in the U.S. private sector. He recently secured a commitment by Nvidia and Advanced Micro Devices to pay the government 15% of their sales in China in exchange for export licenses. Read the story.