📖 2 min read
Five things that actually happened in AI in the last 24 hours — fast, no filler.
1. Google ships Gemini Omni, going head-to-head with GPT-5.5
Google rolled out Gemini Omni Flash globally today, pushing it to every AI Plus, Pro, and Ultra subscriber through the Gemini app and Google Flow. YouTube Shorts and the YouTube Create app are getting it free this week — a clear bid to seed the model into creator workflows before OpenAI’s next push.
📧 Want more like this? Get our free The Ultimate AI Tool Database: 200+ Tools Rated & Ranked — Downloaded 5,000+ times
It’s the multimodal sibling to Gemini 3.5, which DeepMind has been quietly using to tackle long-horizon coding tasks across full codebases. If you’re picking a daily driver, this one’s worth a serious look — our breakdown of the top models lives over at AiToolCrush.com.
2. OpenAI’s reasoning model cracks an 80-year-old Erdős conjecture
On May 20, OpenAI announced one of its general-purpose reasoning models autonomously disproved a discrete-geometry conjecture Paul Erdős posed in 1946. Not “assisted a human” — disproved it, end to end.
This is the kind of result that quietly resets the ceiling. Pure-math problems were supposed to be the last bastion. They’re not anymore.
📧 Want more like this? Get our free The Ultimate AI Tool Database: 200+ Tools Rated & Ranked — Downloaded 5,000+ times
3. GPT-5.5 Instant becomes ChatGPT’s new default
OpenAI flipped the switch on GPT-5.5 Instant as the default ChatGPT model — smarter responses, fewer hallucinations, and finer-grained personalization controls. Free users get the upgrade automatically.
If you’re using ChatGPT to research bets or markets, the new model is noticeably better at sticking to facts. Worth pairing with the AI sportsbook picks we track at BetOnAI.net.
4. Andrej Karpathy joins Anthropic’s pre-training team
One of the biggest talent moves of the year: Andrej Karpathy is now at Anthropic, working on the pre-training team responsible for Claude’s core training runs.
📧 Want more like this? Get our free The Ultimate AI Tool Database: 200+ Tools Rated & Ranked — Downloaded 5,000+ times
Karpathy was a founding member of OpenAI and later led Tesla’s AI division. His landing at Anthropic — right as Claude Code dominates developer mindshare — is a strong signal about where serious researchers think the frontier is heading.
5. Anthropic nears its first profitable quarter as Claude revenue surges
Anthropic is closing in on its first quarterly operating profit, with Claude sales projected to hit roughly $10.9B in the June quarter. That would make it one of the only frontier AI labs in the black — a rarity in a field defined by burning cash on compute.
Meanwhile, Claude Code’s London launch is making the AI-coding anxiety conversation officially mainstream. Developers love it, junior engineers are nervous, and the org charts are starting to shift.
That’s the wrap. More tomorrow — same time, same stack.