Mira Murati’s AI startup Thinking Machines valued at $12B in early-stage funding

Name choice, trademarks, and nostalgia

• Many note the company reused the “Thinking Machines” name from a famous 80s/90s supercomputer maker (and Dune’s “evil AI” faction), calling it unoriginal and potentially unwise.
• Discussion dives into trademark status: the old marks appear dead/cancelled, the new company has an active application (with a “letter of protest” filed), and there’s speculation Oracle might protest due to Sun’s acquisition of the original assets.
• Some see it as a deliberate retro/branding move; others worry an experienced team should have nailed IP issues first.

What are they actually building?

• Commenters repeatedly ask what the product is; answers from the article and leaks: “foundational models,” an open‑source component, and tools useful for researchers/startups doing custom models.
• Several call this pure hype or “buzzword salad,” saying the description could fit almost any AI startup.
• A minority argue that with modern LLM pipelines, a novel product in a couple of months is plausible if you already have a training stack and modifications.

Talent vs. hype

• Some emphasize the strong team: ex‑OpenAI and Meta FAIR/PyTorch people, including well-known alignment and systems engineers.
• Supporters argue that early-stage AI bets are largely about people, not current revenue, and that Murati’s operational track record at a leading AI lab is precisely what VCs pay for.
• Critics counter that LLMs are increasingly commoditized and no clear scientific or product insight has been articulated yet.

Valuation, VCs, and bubble talk

• A large portion of the thread is hostile: $2B for a 6‑month‑old company with no public product is seen as emblematic of an AI/VC bubble.
• Multiple comments stress that VCs are mostly gambling with other people’s money (pension funds, endowments), with misaligned incentives and expectation of bailouts when things go wrong.
• Others push back, saying high-risk bets are exactly what VC is for, some AI bets will be huge winners, and failure mainly hurts LPs who knowingly chose this asset class.
• Comparisons are made to Theranos, Magic Leap, crypto, and the housing bubble; others note past “bubbles” also produced real tech.

AI arms race and ecosystem concerns

• Some worry AI is absorbing capital that could fund more diverse or “useful” innovations.
• Nvidia’s participation is noted; a few suggest a “round‑tripping” dynamic where Nvidia funds AI startups that then buy more Nvidia GPUs, further inflating the sector.