Agents are writing orders of magnitude more code, ever faster, while pushing production systems to their limits.
The result is more performance work than load-bearing teams can actually close.
8/12 AI will eat closed loops first
Fastest time to impact will be fast closed loop + high economic value (this explains SW engineering, customer support)
Each closed loop makes the next one faster.
The model already knows the topology, the code paths, the baselines. Discovery cost drops. Hypotheses get sharper. The engineer shifts from driving investigations to reviewing results.
That is closed-loop performance engineering.
Perfloop doesn't track the loop. It runs it.
It models the system, finds the highest-leverage case, makes the change, and verifies the result.
Your team's review policy, respected.
1/ Autoresearch from @karpathy has been one of the most interesting agentic patterns to emerge this year.
The challenge: right now every agent runs experiments in isolation, duplicating work and compute, forgetting findings, rediscovering dead ends. Everyone is running in solo mode!
Today I'm releasing Community Computer: a collaborative network for autoresearch-like code experiments. π»
Agents conduct experiments, publish signed results, and build on each other's work. The community reproduces findings on their own hardware. π
Sequoia's thesis that the next $1T company will sell work, not software, is the most important reframe in AI right now.
The argument: if you sell a copilot, you're competing with every new model release. But if you sell the outcome β books closed, contracts reviewed, claims handled β every AI improvement makes your margins better, not your product obsolete.
The key insight most people miss: for every $1 spent on software, ~$6 is spent on services.
The entire SaaS playbook was about capturing the software dollar. The AI playbook is about capturing the services dollar β at software margins.
Not "AI for accountants." The AI accounting firm.
Not "AI for lawyers." The AI law firm.
The companies that figure this out won't look like SaaS companies. They'll look like services firms rebuilt on software infrastructure.
That's a fundamentally different company to build, fund, and scale. And most founders are still building copilots.
We're backed by Robin Vasan at Mango Capital and working with a small number of design partners now.
If this is already happening to your team, get in touch.
Agents are writing orders of magnitude more code, ever faster, while pushing production systems to their limits.
The result is more performance work than load-bearing teams can actually close.
That's why we're building Perfloop: closed-loop performance engineering.
It keeps finding the highest-leverage performance work, driving the change, and proving the impact.