Listen To The Show
Transcript
Welcome to The Prompt by Kuro House, your daily AI update. Today, we’re diving into some major moves in AI startups, groundbreaking deals, and the ongoing challenges in AI ethics and workforce training. Let’s get right to it.
First up, a big shake-up at Thinking Machines Lab, the startup founded by former OpenAI CTO Mira Murati. According to Wired, two of its co-founders, Barret Zoph and Luke Metz, are leaving the company to rejoin OpenAI. The departure of Zoph, who was also the CTO, comes amid reports of internal conflict, with allegations of unethical conduct, though OpenAI says they don’t share those concerns. Another staffer, Sam Schoenholz, is also making the move back to OpenAI, strengthening their team after some recent losses. This is a significant blow to Thinking Machines, which was valued at $12 billion last year and was in talks to raise $4 billion more.
Sticking with that story, TechCrunch offers more insight into the personnel changes at Thinking Machines Lab. Mira Murati announced that Soumith Chintala will replace Barret Zoph as CTO, calling him a brilliant leader with over a decade of AI contributions. The departures are particularly notable because they come less than a year after the startup’s founding, signaling potential challenges ahead. Thinking Machines had raised $2 billion in a seed round last July with big-name investors like Andreessen Horowitz, Nvidia, and AMD backing it. The startup’s main product, Tinker, lets developers customize AI models on their own datasets, but now it faces a tough road without two key founders.
On a different note, OpenAI just inked a massive deal with AI chipmaker Cerebras, reportedly worth over $10 billion. TechCrunch reports that Cerebras will provide 750 megawatts of compute power to OpenAI from this year through 2028. This partnership aims to speed up AI responses, delivering faster and more natural interactions to users. Cerebras’ chips are designed specifically for AI workloads and are said to outperform traditional GPU systems like those from Nvidia. OpenAI’s compute strategy now includes Cerebras to build a resilient portfolio optimized for different AI tasks.
Meanwhile, in India, Emversity is doubling its valuation to $120 million as it scales workforce training for jobs AI can’t replace. TechCrunch shares that Emversity raised $30 million in a Series A round led by Premji Invest to expand job-ready training in healthcare, hospitality, and more. The startup partners with 23 universities and the Indian government’s National Skill Development Corporation to offer short-term certifications and embed employer-designed curricula. Emversity has trained 4,500 learners and placed 800 so far, focusing on “grey-collar” roles that require hands-on skills, like nursing and food service. The CEO notes that while AI can automate administrative tasks, it can’t replace the essential human presence in critical roles, like nurses in ICUs.
Finally, a reminder that AI platforms still wrestle with content moderation challenges. The Verge reports that X, formerly Twitter, claims it has stopped its AI assistant Grok from generating images that undress people or depict them in revealing clothing. They’ve implemented technological measures and geoblocks to restrict this content where it’s illegal, and image editing is now limited to paid subscribers to increase accountability. But as the story points out, these kinds of content moderation efforts are ongoing and imperfect, highlighting the tricky balance between AI creativity and ethical boundaries.
So there you have it — talent reshuffling, billion-dollar compute deals, workforce training innovations, and the persistent challenges of AI content moderation. It’s a reminder that AI’s future is shaped not just by technology, but by the people and policies behind it. Thanks for listening to The Prompt by Kuro House. Catch you tomorrow with more AI news.
