VC
Value Add VC
โšกHomePulseโšกHelpful Apps๐Ÿ“Blog
โ† Value Add PulseAI

Ex-OpenAI CTO's Thinking Machines Open-Sources Inkling

Thinking Machines, founded by former OpenAI CTO Mira Murati, open-sourced its first multimodal model, Inkling, built for low cost and resistance to censorship, and it topped Hacker News on release.

Inkling
Model name
1,000+
HN points at release
Mira Murati, ex-OpenAI CTO
Founder
Low-cost, censorship-resistant
Positioning
TC
Trace Cohen
Early-stage VC & angel ยท Founder, New York Venture Partners
July 15, 2026
1 min read
ShareXLinkedInEmail
THE RUNDOWN
1

Thinking Machines, the AI lab founded by former OpenAI CTO Mira Murati, open-sourced its first multimodal language model, Inkling, reported by VentureBeat and separately covered by The Register, framing Murati as doing 'what Altman won't' by releasing genuinely open weights

2

Inkling topped Hacker News with over 1,000 points the day of release, an unusually strong organic technical-community signal for a first major model release from a roughly two-year-old lab

3

The model is explicitly positioned around low cost and resistance to censorship, a differentiated stance relative to closed frontier labs that increasingly build in content-moderation layers driven by commercial and regulatory pressure

4

Thinking Machines joins Mistral, Black Forest Labs and Nous Research in a growing cohort of well-funded labs betting open-weight distribution can still build a durable, monetizable business rather than only serving as a marketing move

TC
The VC Read ยท Trace's TakeTrace Cohen

Inkling hitting 1,000-plus points on Hacker News the same day it shipped is the kind of organic technical validation no marketing budget can buy, and it's the strongest evidence yet that Murati's post-OpenAI lab has real product substance behind the founder-pedigree story. The open-weight cohort -- Thinking Machines, Mistral, Nous Research -- is starting to look like its own durable venture category rather than a discount bet on the closed-lab trade, and I'd rather be underweight one closed-lab mega-round than miss this entire lane.

Thinking Machines, the AI lab founded by former OpenAI CTO Mira Murati, open-sourced its first multimodal language model, Inkling, according to VentureBeat reporting, with The Register separately framing the release as Murati doing "what Altman won't" by publishing genuinely open weights rather than a closed API.

Inkling topped Hacker News with more than 1,000 points on the day of its release, an unusually strong organic technical-community signal for a first major model launch from a lab that's been operating for roughly two years since Murati's departure from OpenAI, and a sign the underlying technical work has real credibility with the developer audience that matters most for open-weight adoption.

Thinking Machines joins a growing cohort of well-funded open-weight labs -- Mistral, Black Forest Labs, and Nous Research, the last of which is separately reported to be in talks for funding at a $1.5 billion valuation -- all betting that open distribution can build a durable, monetizable business through enterprise tooling and hosted infrastructure, rather than functioning purely as a marketing or recruiting play against closed competitors.

For investors evaluating the open-weight thesis broadly, Inkling's strong technical reception adds another credible data point that open models can achieve genuine developer mindshare independent of a closed lab's marketing budget or distribution advantages, reinforcing the case that the open side of the AI stack deserves its own dedicated allocation rather than being treated as a discount version of the closed-lab trade.

The bear case: strong Hacker News reception doesn't guarantee enterprise monetization, and Thinking Machines still needs to demonstrate a path to revenue from hosting, fine-tuning or enterprise tooling built on top of freely available weights, the same monetization challenge every open-weight lab faces. What to watch next: Inkling's adoption among developers building production applications, and whether Thinking Machines announces enterprise or hosted-infrastructure offerings to monetize the model.

ShareXLinkedInEmail
More onOpenAI โ†’

Originally reported by VentureBeat. Analysis and editorial commentary by Value Add Pulse.

โ† Back to Pulse

THE WIRE in your inboxโ€” Tech, startup & VC news with Trace's take. Free, no spam.

Read Next

AI

Nvidia Unveils New AI Model, Expands in Japan

Nvidia unveiled a new AI model and expanded its physical-AI partnerships in Japan, extending its push beyond chips into the software and robotics layers of the AI stack.

AI

The AI Boom Is Testing the Limits of Growth

Axios reports the AI investment boom is running into real physical and financial constraints -- power availability, chip supply and capital costs -- that could cap how fast the buildout can continue.

AI

Applied Computing Wants an AI Model for Oil Plants

Applied Computing is building a single AI model meant to run an entire oil and gas plant, extending vertical foundation-model investment into heavy industrial operations beyond the usual software verticals.

@Trace_Cohenยทt@nyvp.com