I benchmarked benchmarks. It audited 14 popular AI benchmarks and found that about 1 in 5 sampled MBPP and Terminal-Bench items are broken.
Typically a wrong answer key, a question with more than one fair reading, or a test that rejects a correct answer.
Often it's plain human error. Even grade-school math (GSM8K) has a key that counts 8 AM to 11 PM as 3 hours.
github.com/EnesYilmazcode/be…
The fly is a simulation of a real fruit fly's full brain wiring, run on @modal GPUs, and it beat @typesafeai's Jev in 86% of games.
github.com/EnesYilmazcode/Fl…
Jev picks every step of the beginner's method from a real 20-move scramble through @vercel AI Gateway, with no solver in the loop.
github.com/EnesYilmazcode/Je…
I made 7,000,000 flies play Tetris.
The board goes in through its eye neurons, and a readout in its brain picks every move.
Trained on @modal GPUs
github.com/enesyilmazcode/Fl…
Enes retweeted
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
Enes retweeted
Introducing @NeoCognition, the agent lab for specialized intelligence.
Everyone needs experts, but human expertise does not scale.
Backed by $40M seed funding, we build self-learning agents that specialize across domains to make expertise abundant.
Cool shortcut I made today!
writing "claude --dangerously-skip-permissions" was too long
So I shortened it to "ccd" (Claude Code Dangerous)
I tried it, and it works!
Cool cave generator I made
If a cell has fewer than 5 neighbors, it dies
github.com/EnesYilmazcode/Sp…