Garrett Tacoronte

haha got my first AI full send

The teams that move fast aren't skipping review. They're surgical about where human attention is actually worth the wait. What's the last review layer you added — and do you know what it's actually catching?

The replacement isn't no review. It's tiered review. Auth, payments, core data models — full review, no shortcuts. UI, copy, internal tooling — ship it, async comment if you want. Pair program on the hard stuff instead of async back-and-forth that takes 3 days.

Two engineers. Same team. One ships a feature in 2 days. The other takes 12. Same complexity, same codebase. The difference isn't skill. It's how many people have to say yes before the code lands.

Fix the reproduction problem and the diagnosis gets 10x faster with or without AI. The session portability is the unlock. The AI is just a nice bonus on top.

Everyone's focused on 'LLM auto-diagnoses your bug.' Cool. But the real shift is that debugging context is no longer trapped inside one developer's browser. That's the thing that kills sprint velocity — not the diagnosis, the reproduction.

Counterintuitive take on Chrome's new DevTools MCP: the AI part is the least interesting part.

Cold email pivot: 'Your marketing crushes it. Here's why your last user surge tanked load times.' Books fractional help instantly. You scraping repos like this?

Misses the signal. Those stargazers nailed marketing flywheels, hit PMF hard. Now growth exposes eng cracks — deploys breaking weekly, latency spiking under traffic.

X threads exploding on outbound to 'Marketing for Founders' stargazers. Apollo enrich, Exa scrape, AI SDR fires personalized cold emails. Founders nodding: perfect non-tech ICP.

You benching humans once AI triages risks? Or hiring the one who are “cyborgs” and lean into the AI as a tool to empower them and make them more productive and effective? I’d choose the latter ever day

Dead wrong. AI floods you with 50 alerts. The human knows which one bursts during your demo week. That's why fractional CTOs ship stability now.

Travis clip blowing up: AI spots the pipe leak in seconds, so plumbers obsolete. Founders nodding, eyeing their CTOs next.

@briandonatiello Yup 👍

@10exdev Latency is the silent killer. 20s vs 2min response time changes everything in live conversations. Parallel calls should be table stakes for any AI SDR.

Seeing X posts and demos of AI SDR agents everywhere: Apollo enrich, wait for result, Exa search, wait, HubSpot check. 2min per lead feels futuristic.

Locking into one provider trades flexibility for hype. Multi-model routing future-proofs without rewrite hell. What's your stack routing across right now?

Fourth startup this month where single-model obsession tanked their AI perf. Routed same prompt across Claude/Grok/GPT-5 — output quality jumped 40%, costs halved. No code changes.

a16z multi-model thesis hits hard. Founders reading it as 'pick the one best LLM and scale'. That's the miss.