- AI Breakfast
- Posts
- OpenAI's Secret Internal Model
OpenAI's Secret Internal Model
Good morning. It’s Monday, February 10th.
Did you know: On this day in 1996, IBM's Deep Blue made history by defeating world chess champion Garry Kasparov in a single game? It was the first time a computer had beaten a reigning world champion under standard chess tournament time controls.
In today’s email:
OpenAI’s Internal Model
Apple’s Research Robot
Perplexity Adds 1M Token Context
5 New AI Tools
Latest AI Research Papers
You read. We listen. Let us know what you think by replying to this email.
What's next in the new era of tech: Our friends over at Semafor Tech write a twice weekly newsletter on the people, the money and the ideas at the center of the new era. They keep a finger on the pulse of the shifting dynamics and the new ideas that just might be crazy enough to change the industry as we know it. You can sign up for the newsletter here.
In partnership with ARTISAN
10x Your Outbound With Our AI BDR
Imagine your calendar filling with qualified sales meetings, on autopilot. That's Ava's job. She's an AI BDR who automates your entire outbound demand generation.
Ava operates within the Artisan platform, which consolidates every tool you need for outbound:
300M+ High-Quality B2B Prospects
Automated Lead Enrichment With 10+ Data Sources Included
Full Email Deliverability Management
Personalization Waterfall using LinkedIn, Twitter, Web Scraping & More

Today’s trending AI news stories
OpenAI’s Internal AI Model
Sam Altman, in a new blog post, laid out OpenAI’s AGI playbook to democratize access while conceding that AI’s economic ripple effects won’t be kind to everyone. He floated the idea of a “compute budget” to prevent AI from becoming a luxury good but also acknowledged the inevitable potential of AGI to disrupt the balance between capital and labour, calling for preemptive guardrails.
This tension between accessibility and economic impact was further underscored in a dialogue at UTokyo, where Altman revealed that OpenAI's internal AI model currently ranks as the 50th best competitive programmer in the world, with plans to reach #1 by the end of 2025.
In terms of Codeforces ELO:
o1 (Oct 2023) was ranked ~9800tho3 (Dec 2023) was ranked ~175th
Altman has confirmed that Open AI has a model internally at ~50th
Superhuman programming by eoy
— Justin Halford (@Justin_Halford_)
12:48 PM • Feb 8, 2025
OpenAI’s war chest ambitions—$40 billion in funding and a possible $500 billion infrastructure spend—signal an all-in bet on scale. Altman insists AI costs plummet tenfold annually, making frontier models more accessible, but that does little to assuage concerns about equitable distribution. He also teased a move toward more transparency—albeit with security caveats—warning of AGI’s appeal to authoritarian regimes. Read more.
WATCH: Q&A with OpenAI Leadership
Apple’s new research robot takes a page from Pixar’s playbook
Apple’s latest research paper introduces a novel approach to human-robot interaction, focusing on expressive robot movements. Drawing inspiration from Pixar’s Luxo Jr. lamp, the paper suggests that integrating emotional qualities, such as intention and attention, into robot motion enhances connection with humans.
Apple's Machine Learning Research team just gave us our first look at the possible upcoming tabletop robot which was reported by @markgurman
— Aaron (@aaronp613)
6:05 PM • Feb 6, 2025
Apple’s prototype—a lamp-shaped robot—mimics these movements, demonstrating how small gestures, like swiveling its head to mimic human curiosity, can foster a deeper engagement. With a non-anthropomorphic design, the robot diverges from Amazon’s Astro, hinting at future products with a more functional, yet personable, form. Read more.
Perplexity expands features with file and image uploads and 1M Token context window
Perplexity has expanded its capabilities, now offering file and image uploads alongside an enhanced context window of 1 million tokens. This feature is available to all signed-in users in “Auto” mode, further extending the AI's potential to process and manage complex datasets with greater efficiency. Read more.
Perplexity now offers file and image uploads with an expanded context window of 1 million tokens.
Free for all signed in users in “Auto” mode.
— Perplexity (@perplexity_ai)
4:11 PM • Feb 7, 2025

Anthropic CEO slams DeepSeek as 'the worst' in bioweapons data safety test
Christie's announces AI art auction, and not everyone is pleased
Report: Ilya Sutskever's startup in talks to fundraise at roughly $20B valuation
Tech titans and government leaders gather in Paris for landmark AI summit
France to Announce €109 Billion in AI Investments, Macron Says

5 new AI-powered tools from around the web

arXiv is a free online library where researchers share pre-publication papers.


Thank you for reading today’s edition.

Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.
Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email or DM us on 𝕏!