The AI platform for human intelligence that will match 1 billion people with their dream role.
External link for micro1
San Francisco Bay Area, CA, US
micro1 reposted this
great chat with Raymond Ravaglia at Forbes on why every enterprise in the world will eventually spend millions a year on evals (structured human judgement). its the only route to real automation & AI adoption. https://lnkd.in/gQUaykZs
Exactly this. The path to AGI depends on structured human judgment across every domain, and the opportunities are endless. (via Inc. Magazine / Joshua Browder)
Human judgment is the bottleneck in AI. The world keeps changing, which is why human data is becoming a trillion dollar market with micro1 at the frontier Proud investor in this company since the pre-pre-seed https://lnkd.in/gxVkQhbM
Great listen ⬇️
𝐁𝐑𝐄𝐀𝐊𝐈𝐍𝐆: 𝐓𝐡𝐞 𝐭𝐫𝐞𝐧𝐝 𝐞𝐯𝐞𝐫𝐲𝐨𝐧𝐞 𝐤𝐞𝐞𝐩𝐬 𝐫𝐞𝐩𝐞𝐚𝐭𝐢𝐧𝐠 𝐢𝐬 𝐰𝐫𝐨𝐧𝐠: 𝐀𝐈 𝐢𝐬𝐧’𝐭 𝐫𝐞𝐩𝐥𝐚𝐜𝐢𝐧𝐠 𝐢𝐧𝐯𝐞𝐬𝐭𝐢𝐧𝐠. 𝐈𝐭’𝐬 𝐫𝐞𝐩𝐥𝐚𝐜𝐢𝐧𝐠 𝐢𝐧𝐯𝐞𝐬𝐭𝐦𝐞𝐧𝐭 𝐥𝐚𝐛𝐨𝐫. That distinction matters. A lot. Let that sink in. This came up clearly in my conversation with Ali Ansari (CEO of micro1) and it reframed how I think about where real edge actually comes from. Here’s what’s actually happening 👇 AI is getting very good at the mechanical parts of investing: models, spreadsheets, comps, scenario analysis, data cleanup. That’s not investing. That’s table stakes. The central insight: Alpha was never created by building models. Alpha was created by deciding where to look, who to trust, and when to act. AI compresses time. It doesn’t replace judgment. The paradox? As AI gets better at analysis, human judgment becomes more scarce—not less. Everyone can run the same model. Almost no one can ask the right question. That’s the bottleneck. AI takes over the 90% of work that looks like intelligence. Humans focus on the 10% that actually compounds: • Relationship formation • Industry pattern recognition • Market timing • Conviction under uncertainty That’s where returns come from. Here’s the part most people miss 👇 Firms deploying AI fastest aren’t “automating investors.” They’re raising the floor of execution so humans can operate at the ceiling. Bad investors with AI stay bad. Great investors with AI become lethal. Practical takeaway: If you think AI is “replacing investing,” you’re asking the wrong question. The real question is: What human work is now freed up—and are you good at that work? Most aren’t. That’s the opportunity, We’d like to thank AlphaSense for sponsoring this episode! Thank you Alex Pattis for the kind introduction! #investing #AI #Leadership #Founders #DecisionMaking Link to Podcast in Comments Below 👇
micro1 reposted this
Join Imran Nasim, PhD and I on February 4th at 10am PT for a conversation on micro1's forum about how enterprises build trust in AI agents. As agents move from demos into real workflows, teams are running into a new set of challenges around reliability, evaluation, and scale. In this discussion, we’ll explore what breaks in production, why traditional benchmarks fall short, and how domain expertise enables more accurate evaluation and sustained improvement. Register below to learn why contextual evaluations are critical to deploying enterprise AI with confidence: https://lnkd.in/dwdFhE4p
High-quality human data starts with deep, domain-specific vetting. That’s why we just released v2 of our On-the-Fly Exercise model in Zara. Instead of relying solely on static or conversational assessments, teams can now generate real-world, role-accurate exercises using natural language, tailored to how experts actually think, decide, and work. In this walkthrough, Gustavo Farias builds an On-the-Fly Exercise and shows how it works end to end.
micro1 reposted this
On January 29th, I’ll be joining LIU ZHANG to discuss our recent research at micro1 (https://lnkd.in/dfCz8tDw), which looks at how AI safety behaviors show up in real-world use as models scale. The conversation will be moderated by Mark Esposito, PhD, our Chief Economist at micro1 and a mentor to me since I joined, which makes this discussion especially meaningful. The core idea is that effective safety is less about blocking and more about competence under constraint: helping models respond in ways that are informative, grounded, and aligned with user intent, even in constrained settings. 📅 January 29th, 9am PT Register here :)) [https://lnkd.in/eBtJjZsF]
The last mile doesn’t exist in AI. Ali Ansari and Aaron Levie share how automation creates a continuous loop of new functions and, in turn, new work across the AI economy.
micro1 reposted this
great live chat with Aaron Levie just now on a $1T/year human data market that's emerging (and what Enterprises must do to prepare). https://lnkd.in/durnKkeE