I tweet on innovation at scale. 37 accelerators, 6 govs, 11 unis, 24 1B+ corps Newsletter: https://xitter.cf/t.co/eg2f1XGi1u Free tools: https://xitter.cf/t.co/9OCmjfqLgW
Brooklyn, NY, USA
Joined April 2009
- Tweets22.4K
- Following1.1K
- Followers10.8K
- Likes4.6K
Pinned Tweet
Testing an idea used to cost me a day of hand-coding. Now it's 5 min to prompt, 30 to make tea. That's why I rewrote the book: when building is free, building the wrong thing is the only real cost. kromatic.com/download/the-re…
Most founders get bad data because they ran the wrong experiment for the question they had. AI runs it 10x faster — it can't tell you which question to ask. New RSB 2nd-ed teaser: kromatic.com/blog/rsb-teaser…
Every method in The Real Startup Book now has a copy-paste AI prompt. Work through it with Claude, ChatGPT, Gemini, or Codex — paste it, give your context, let it help design the experiment. kromatic.com/mcp?utm_campaig…
The Real Startup Book is free, built by 51 contributors across 16 countries. Even Claude, the AI, earned a contributor credit — drew illustrations, ran research, and argued with me. kromatic.com/download/the-re…
The Fake Door Test: put a button or ad in front of people for something that doesn't exist yet, then count the clicks. A click is a behavioral signal, not a stated opinion. kromatic.com/download/the-re…
My favorite method in the book: Picnic in the Graveyard. Before you build, study startups that already died trying something close to your idea. Spot the landmines first. kromatic.com/download/the-re…
Genuinely humbled — @nireyal, @tendayiviki, and @danto_ma had kind words for the new Real Startup Book. Three people whose work shaped how I think about building. kromatic.com/download/the-re…
When a working prototype is one AI prompt away, the expensive mistake is building the wrong thing. The Real Startup Book 2nd edition is out today — 60+ methods for getting that part right. kromatic.com/download/the-re…
Tristan Kromer retweeted
Why AI agents should speak Latin, how to avoid the agentic convergence trap, and the power of the peanut butter and jelly sandwich...
Including insights from @TriKro, @erparker, @HarvardBiz, and @SanjanaPattan14
Listen to @theiopodcast at:
insideoutside.io/podcast/ai-…
94% of employees know generative AI. 1% of leaders say they're mature on deployment. The payoff isn't there.
kromatic.com/blog/why-ai-man…
"I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes." (Joanna Maciejewska)
Hook the chore, not the craft.
When your company rolled out an AI tool, did anyone actually use it? Or did people find creative ways to look like they did?
kromatic.com/blog/why-ai-man…
Who Should Own the AI Budget—and Why This Is a Terrible Question – WeAreOutcome weareoutcome.co/blog/who-sho…
The cheapest way to debug an AI agent is a Latin phrase. When my chief-of-ops agent stops saying _festina lente_, the persona has washed out.
kromatic.com/blog/why-my-age…
"A memory the agent has to remember to look up isn't memory." Four design choices separate production agentic memory from a markdown file.
kromatic.com/blog/agentic-me…
When your AI agent makes a decision in one session, does it remember WHY next week?
How are you handling memory across agent sessions right now?
kromatic.com/blog/agentic-me…
Claude hallucinated that we'd worked for the Austrian Federal Forestry Service. Completely false.\n\nThat's why every fact in our memory system now requires a citation. No source = assertion, not fact.\n\nkromatic.com/blog/agentic-m…
Harvard + Microsoft studied orgs with hundreds of active AI deployments. Productivity gains "remain trapped inside individual workflows." The bottleneck isn't the tech. It's org design. kromatic.com/blog/stop-putti…