Editorial illustration for Google Surges in Pre-Training Race, OpenAI Plots 'Shallotpeat' Comeback
Google Leads AI Race with Pre-Training Breakthrough
Google pulls ahead with pre-training; OpenAI's comeback plan named ‘Shallotpeat’
Everyone thought pre-training was a solved, boring part of building AI. Google just used it to take a lead. The company's latest models show that steady, unglamorous improvements to this foundational phase still translate into a real edge.
It’s the engineering equivalent of grinding out a two-yard gain on every play. It works.
Sam Altman confirmed it. He said Google’s recent work is excellent, especially in pre-training. For OpenAI, that’s an admission of a problem.
While Google methodically advanced, OpenAI reportedly hit walls trying to do the same. Their pivot was toward building “reasoning” models, a different and arguably more ambitious path. Now they need a response.
Its internal codename is ‘Shallotpeat.’
So the field splits. Google’s path is about incremental gains on a massive scale. OpenAI’s ‘Shallotpeat’ is a bet that reasoning can leapfrog all that grinding computation.
One strategy is disciplined. The other is speculative. The next phase of the race won’t be about who has the most data, but who has the better idea for what to do with it.
Google is ahead. For now.
Related Reading
Browse all OpenAI news →
Business & Startups Nordic pilot adds Gemini for Education, NotebookLM to boost AI literacy
Business & Startups Kling launches Video O1, all-in-one model with MVL bridge using transformer
Business & Startups DeepSeek Seeks More Capital Weeks After USD 7B Funding Round
Related Topic OpenAI researcher quits, citing distrust over ad‑driven engagement metrics
LLMs & Generative AI
Related Topic OpenAI launches GPT-Image 1.5 with precise editing for enterprise visuals
LLMs & Generative AI
From Archives AI startups use revenue as recruiting bait, citing demos and social media buzz
From Archives AI Agent Evaluation Supplants Data Labeling as Key Step to Deployment
From Archives GPT-5 helps mathematicians offload tedious tasks, says Timothy Gowers
Open Source
From Archives OpenAI adds group chat to ChatGPT for planning meals, trips and outlines
LLMs & Generative AI
Common Questions Answered
How has Google gained an advantage in AI pre-training techniques?
Google has recently demonstrated significant progress in pre-training methodologies, showing that this fundamental AI development phase still holds considerable potential. Their breakthrough suggests nuanced improvements are possible, even as dramatic performance leaps become less frequent.
What did Sam Altman acknowledge about Google's recent AI work?
Sam Altman publicly recognized that Google has been doing 'excellent work recently' in pre-training techniques. His acknowledgment signals a potential shift in the competitive dynamics of AI development, highlighting Google's meaningful advances in the field.
What is the significance of 'Shallotpeat' in OpenAI's strategy?
OpenAI is pursuing a mysterious strategy called 'Shallotpeat' as a potential response to Google's recent pre-training advancements. While the specific details remain cryptic, it appears to be part of OpenAI's efforts to maintain competitive edge in the AI development landscape.
Further Reading
- Google launches Gemini 3 with new coding app and record benchmark scores — TechCrunch
- The Scaling Wall Was A Mirage — Tomasz Tunguz
- AI Benchmarks 2025: Performance Metrics Show Record Gains — Sentisight
- A new era of intelligence with Gemini 3 — Google Blog
