Source: a leaked April 2023 internal document by Luke Sernau, a senior software engineer at Google, argues that open source AI will outcompete Google and OpenAI

A senior software engineer at Google wrote a critique asserting that the internet search leader is losing its edge … Source: SemiAnalysis .

Bloomberg 2023-05-05

Context & Ripple Effects

The leaked memo surfaced as Google was presenting AI as an extension of Big Tech’s existing product strategy, while related coverage identified the more consequential contest as centralized models versus open source. The document made that tension unusually explicit from inside Google.

It matters because the critique challenges the assumption that leading labs’ model capabilities alone determine durable advantage. Later coverage of Google’s work on reasoning models shows the company continuing to pursue frontier-model progress alongside that open-source pressure.

First-order effects

The leak publicly exposes an internal argument that Google and OpenAI could lose ground to open-source AI, adding strategic pressure to justify the advantages of their centralized model approaches.
Open-source developers gain a high-profile validation of the view that rapid community iteration can be a competitive force against large labs, even though the document itself does not establish an immediate market shift.

Second-order effects

Google and OpenAI face a sharper need to differentiate proprietary offerings through capabilities, product integration, and distribution rather than treating model access as the sole moat.
The framing raises the stakes for developers deciding between open models and hosted platforms: model quality is only one consideration alongside control, iteration speed, and ecosystem fit.

Third-order effects

If open-source model ecosystems continue to narrow the capability gap, AI competition could split between broadly available model layers and proprietary products, infrastructure, and distribution built around them.
The episode points to AI commons becoming strategic infrastructure: control may shift from ownership of a single model toward the ability to improve, deploy, and govern models at scale.

The trend: This is one data point in the widening contest between centralized frontier labs and open-source ecosystems for control of AI’s most valuable layers.

Discussion

@sashamtl Dr. Sasha Luccioni on x
Hope y'all took the time this week to read this! https://www.semianalysis.com/ ... It really reflects the power of open-source technology and its threat to AI hegemony. No amount of company secrets and poaching employees can beat sharing and building upon each other's work.
@simonw Simon Willison on x
Leaked Google document: “We Have No Moat, And Neither Does OpenAI” The most interesting thing I've read recently about LLMs - a purportedly leaked document from a researcher at Google talking about the huge strategic impact open source models are having https://simonwillison.net/…
@eladgil Elad Gil on x
I always remember when I was at Google in the mid 2000s and social was happening I was at a meeting where many people (now known as social “experts") were pounding the table saying social products had no moat and would never be sticky https://www.semianalysis.com/ ...
@jelleprins Jelle Prins on x
🔥 If you read one thing about AI this week, let it be Google's leaked doc “We have no moat, and neither does OpenAI” https://twitter.com/...
@emostaque Emad on x
While this article fits with much of our thesis I think it has a misunderstanding of what moats actually are It is v difficult to build a business with innovation as a moat, base requirement is too high Data, distribution, great product are moats https://www.semianalysis.com/ ...
@loomdart @loomdart on x
This is a very dramatic leak that will probs go under the radar but wow https://www.semianalysis.com/ ...
@nathanbenaich Nathan Benaich on x
Fascinating read on AI strategy. Google “We Have No Moat, And Neither Does OpenAI” “The value of owning the ecosystem cannot be overstated.” I'll predict some huge M&A for OSS AI cos. Community is everything. https://www.semianalysis.com/ ...
@daveyalba Davey Alba on x
@byJuliaLove @rachelmetz UPDATE: We can confirm the author of the document is Luke Sernau, a senior software engineer at Google. He published the document internally in early April, and it got thousands of shares among Googlers before it was leaked and published by SemiAnalysis. …
@adamkovac Adam Kovacevich on x
Stuff like this is a big reason to be skeptical about AI “Monopolization” Panics https://twitter.com/...
@wolfejosh Josh Wolfe on x
Opensource ftw @huggingface https://twitter.com/...
@adcock_brett Brett Adcock on x
Interesting leaked internal google memo: → “[OpenAI's] ability to maintain an edge is necessarily in question” → Open-source LLM's will win vs. Bard/ChatGPT → Author advocating for “LoRA” = quick model fine-tuning https://www.semianalysis.com/ ...
@davetroy Dave Troy on x
Important thread here; leaked doc from Google argues that big companies are not advantaged in the AI arena because the proliferation of lightweight open source AI models allows for faster, more creative and wider innovation. May be a serious inflection point for big tech. https:/…
@mobileraj Raj Singh on x
I agree with @Scobleizer Owning the AI developer platform relationship which OpenAI is doing is the moat. It's the same moat MS had with Windows developers. It's the same moat AWS has with cloud developers. https://twitter.com/...
@steipete Peter Steinberger on x
This leaked Google doc is everything. 🍿🍿🍿 https://www.semianalysis.com/ ...
@scobleizer Robert Scoble on x
This is wrong. Developers and community are moats. Like I said yesterday. You think @khanacademy is going to rip out GPT underneath its new AI-centric education system? And it is one of many building on top of GPT. You are insane if you think that. Google will never get Khan... h…
@hosseeb Haseeb on x
A powerful realization, leaked from a Google researcher. LLM companies do not have moats. Open source is winning as always. The collective intelligence of the world is racing ahead of secretive research orgs when it comes to optimizing LLMs. https://www.semianalysis.com/ ...
@counternotions Kontra on x
“Leaked” Google document: “...the uncomfortable truth is, we aren't positioned to win this arms race and neither is OpenAI. While we've been squabbling, a third faction has been quietly eating our lunch. I'm talking, of course, about open source.” https://www.semianalysis.com/ ..…
@t0nyyates Tony Yates on x
AI is not going to be regulated by focusing on a few large companies. This is already clear from descriptions of how the large language models have ‘leaked out’ of Meta and elsewhere. https://twitter.com/...
@sergeykarayev Sergey Karayev on x
Google has no moat. They don't have over 90% search traffic. They don't have everyone's emails and the most used email client. Their OS is not powering 70% of smartphones. They will never be able to deploy LLM features into these products — instead, people will run OSS LLMs.
@alexgdimakis Alex Dimakis on x
My thoughts on the now famous Google leak doc: https://www.semianalysis.com/ ... 1. Open source AI is winning. I agree, and that is great for the world and for a competitive ecosystem. In LLMs we're not there, but we just got OpenClip to beat openAI Clip and Stable diffusion is b…
@goldman Jason Goldman on x
I think I was on the receiving end of this table pounding at Google. And the forces of market concentration will work as they do and sticky, defensible AI products will be built. BUT! If its true as the doc alleges that open source LLMs are better tech, that's a bfd https://twitt…
@mont_jiang @mont_jiang on x
It would be unexpected and kind of glorious of Google were the first FAANG to get totally decimated by technological competition. They always seemed to be the most entrenched technologically, like a coding superpower. https://www.semianalysis.com/ ...
@simonw Simon Willison on x
I've now heard from an ex-Googler that they heard from a current-Googler that this is indeed a legit internal document, for whatever that's worth
@nitashatiku @nitashatiku on x
new from me and @GerritD For years the tech giant published scientific research that helped jump-start its competitors. But now it's lurched into defensive mode. https://www.washingtonpost.com/ ...
@hughlangley Hugh Langley on x
New: Google leaders have told staff they plan to be more strategic about the AI research the company publishes. Google wants to keep more knowledge in-house as rivalries with Microsoft etc heat up. With @tomaxwell https://www.businessinsider.com/ ...
@oablanchard Olivier Blanchard on x
Interesting approach given the need for interoperability and standards in the A.I. space. This also raises questions for me about IP: If you don't publish or patent your innovations until after they become products, it's going to be difficult to ramp up licensing and adoption. ht…
@kantrowitz Alex Kantrowitz on x
OpenAI is driving AI closed https://twitter.com/...
@jfischoff Jonathan Fischoff on x
Considering the ratio of papers to products for google is like a million to one, they should just say they aren't going to publish any more. https://twitter.com/...
@mathemagic1an Jay Hack on x
This makes sense from a game-theoretic perspective, esp. considering OpenAI's existing policies Will drive up the value of alumni of leading AI labs and generally have a chilling effect on startup competition https://twitter.com/...
@taliaringer Talia Ringer on x
This article by @nitashatiku and @GerritD is really good, and also makes me feel really sad. It was beautiful that there was ever an era in which an industry giant was willing to selflessly share research with the outside world before monetizing. https://www.washingtonpost.com/ .…

Chronicles