Five researchers at Thinking Machines Lab, the AI startup led by former OpenAI technology chief Mira Murati, posted on X within about three hours of its Inkling 2 Small launch on 5 November that they had recently joined the company or were in their first months there. Four said so directly; Murati replied to three of them, writing to one, “Great to have you on the team, Yiwen!”.
The posts land as Business Insider reports the company is raising $5 billion at a $40 billion pre-money valuation, and after at least ten founding-team members left, by our count of reported departures. By our arithmetic, that is five public new-hire posts against ten sourced founding exits.
Business Pill · Power Law Explained
A short explainer on why a few take most. It teaches the general idea only and does not discuss this story.
The key insight: As we read the launch-night posts, Thinking Machines is answering the departure story with arrivals: five researchers said publicly within about three hours of the Inkling 2 Small launch that they had joined in recent months, against at least ten founding-team exits we have counted from reports, five public arrivals for ten sourced departures by our arithmetic.
Yiwen Yuan wrote: “I joined Thinky in August to work on coding post-training.” Hao Zhu wrote: “Personal news: I joined Thinking Machines in September.” Idan Shenfeld wrote: “Maybe a good time to announce that I joined @thinkymachines three month ago.”
Vincent Cho called the model his “First release since joining Thinking Machines!” Anchit Gupta wrote of “a really fun few months at thinky”. Three of the five give a month, August or September 2026; none of the posts gives a job title beyond Yuan’s “coding post-training”.
Founding-team departures from Thinking Machines Lab as reported by The Next Web and Fortune, by destination, against the five researchers who posted on X on 5-6 November that they had recently joined or were in their first months at the company. Counts are ours.
What They Said About the Model
Shenfeld listed the model as “pre-trained in less than 3 days” and called it “very finetunable, way more than other OSS models”. Horace He, also at the company, wrote that it “took just under 3 days to pretrain” and gave its Tinker prices as “0.30/1.00/1.50 for in/out/train”.
Read with our launch piece, the first two numbers are the per-million-token input and output prices; the third, $1.50, is the training price, which our launch piece did not carry. These are statements by employees, not company filings.
Against the Exits
Our earlier count found at least ten founding members gone by July 2026: five to Meta, four to OpenAI and one to xAI, as The Next Web and Fortune reported. Lilian Weng, the last on that list, left in July. The three joiners who give a month all arrived after that, in August or September, by their own posts.
Five posts are not a headcount. The company has not published how many people it employs or has hired, and the posts show only the researchers who chose to announce. The round itself is in our first piece on the raise.
As we read the dates, the arrivals are recent: the three researchers who give a month name August or September 2026, after the last reported founding exit in July.
As we read the posts, the joiners talk about the model rather than the company: pretraining in under three days, a single joint reinforcement-learning run, and a model they call very finetunable.
As we read what is missing, five posts are a floor, not a total: the company publishes no headcount, so the net change in staff cannot be measured from these posts.
Yiwen Yuan on X, 6 November 2026
“I joined Thinky in August to work on coding post-training.”
Three Implications
FOR THE $5 BILLION ROUND Business Insider reports Thinking Machines is raising $5 billion at a $40 billion pre-money valuation, citing sources familiar with the matter.
FOR TINKER USERS Horace He gives Inkling 2 Small’s Tinker prices as 0.30/1.00/1.50 for in/out/train.
FOR RIVAL LABS Our earlier count found nine of ten reported founding exits went to Meta or OpenAI.
The framework says: “The transformer is a commodity now; what changes between paradigms is where intelligence accumulates and who can capture it.”
As we read it, the people moving between labs are one way that intelligence moves: Thinking Machines lost founders to Meta and OpenAI and is now showing new arrivals. The framework predates this story and does not discuss Thinking Machines.
What Is Not Established
No post or company statement gives Thinking Machines’ total staff, the number of hires since July, or whether the $5 billion round has closed. Business Insider’s report on the raise cites sources familiar with the matter, and the company declined to comment to it. We did not contact Thinking Machines or the researchers.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
Five Thinking Machines researchers posted around the Inkling 2 Small launch that they had joined in recent months, three of them in August or September. Set against at least ten reported founding exits, that is five public arrivals for ten sourced departures, by our arithmetic.
A note on sourcing. The joining dates, model claims and prices are from the researchers’ and Mira Murati’s own posts on X on 5 and 6 November 2026. The departure count and the funding report are from our earlier coverage, citing The Next Web, Fortune and Business Insider. The comparison is our arithmetic. We did not contact Thinking Machines or the researchers.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
