Loading activity…
@zainhas
I build and teach AI • AI/ML @togethercompute • EngSci ℕΨ/PhD @UofT • Previously: vector DBs, data scientist, lecturer & health tech founder • 🇺🇸🇨🇦🇵🇰
AI
A post argues that, on some problems, a team of agents can outperform a single agent given proportionally more runtime—suggesting an advantage in compute efficiency, not just speed.
AI
Wafer says the first resource uses worked problems to connect model dimensions to deployment choices, including memory capacity, workload distribution and expected performance under stated assumptions.
Technology
DeepSeek says its KV cache, which lets the model reuse earlier calculations, needs a quarter of the high-bandwidth memory and an eighth of the SSD storage required by the previous generation's cache.
AI
Investor Deedy Das posted a sarcastic reply targeting Wired's sourcing for an AI story.
AI
Zain recommends high reasoning effort mode for GLM-5.3 users on the platform.
AI
Maksym Andriushchenko shares ongoing work testing agents on long-horizon benchmarks.
