DeepMind's AGI Contest Just Crowned... AI Slop?
What happened
A submission to Google DeepMind's Kaggle competition 'Measuring Progress Toward AGI' won the $25,000 Grand Prize — and the Kaggle community immediately called it out as low-quality, AI-generated slop. Discussion threads are now dissecting how a supposedly rigorous benchmark for machine intelligence rewarded work that reads like it barely tried.
Why this matters
This competition exists specifically to measure whether AI systems are creeping toward AGI-level reasoning — so if the judging process itself gets fooled by shallow, AI-assisted output, that's a credibility problem for the whole benchmark. It's an awkward irony: a contest about measuring intelligence apparently couldn't detect a lack of it in the winning entry.
The slightly cynical read
With cash prizes and prestige on the line, of course entrants are using AI to generate submissions faster than judges can vet them. If a flagship DeepMind competition can't catch obvious slop, what does that say about every other 'rigorous' AI eval leaderboard flooded with similarly optimized entries?
What to watch next
Expect Kaggle and DeepMind to face pressure to explain their judging criteria, possibly revoke or amend the prize, and tighten anti-slop detection for future rounds. Watch whether this becomes a broader flashpoint about AI-generated submissions gaming competitions and peer review everywhere.