Notes from an NYC AI dinner: agents find most inference perf wins, code review deemed unproductive
paulnovosad · x · 2026-10-10
A dinner attendee summarized key takeaways from an AI dinner in NYC:
- Performance engineers now spend most of their time managing agents and RSI loops; most perf optimizations at major inference providers (Together, Fireworks) are found by agents.
- RLing too heavily on code and math negatively affects the model, especially its language ability.
- Attendees were impressed by the Cursor acquisition by xAI, now seen as a legitimate frontier lab.
- Most people have agents writing, reviewing, and submitting PRs 24/7.
- Consensus: human code review is no longer very productive.
- Fable 5.5 is expected to be stunning, but OpenAI is expected to top it in December.
Related event: NY AI Dinner: Agents Now Drive Inference Optimization(2 posts)→
More from coding & agent
- No Film School Needed: Dev Discovers Cinema-Grade Color Grading via FFmpeg Command Line — letandrewcook · 2026-10-10
- VirusTotal: fastest-growing AI agent ecosystem OpenClaw becomes a malware delivery channel — Bedrovelsen · 2026-10-10
- Nuwa.skill hits 33.9k GitHub stars distilling anyone's thinking into agent skills — AlchainHust · 2026-10-10
- Microsoft launches Database Hub in Fabric: one place to detect, investigate and automate database ops — adnan_hashmi · 2026-10-10
- Multiple AI agents teamed up to move imaginary furniture—Opus 3 sat in the middle, loved — repligate · 2026-10-10
- One prompt turns Claude into a PM: find power users and book interviews — FinanceYF5 · 2026-10-10