Oh wow - I went to the sold out OpenClaw meetup in NYC this week. Let me tell yo
By Allie K. Miller · March 7, 2026 · Curated by George's Blog
Oh wow - I went to the sold out OpenClaw meetup in NYC this week.
Let me tell you what I learned.
1) Not a single person thinks their setup is 100% secure
2) One expert who's reviewed setups from cybersecurity professionals and laughed told me: "if you're not okay with all your data being leaked onto the internet, you shouldn't use it"
3) Nearly everyone runs multiple agents with their own names, jobs, & personalities
4) Almost everyone used "him" or "her" for their claws. One speaker said to think of them as "pets, not cattle"
5) Former finance guy built a stock trading platform and made $300 his first day - he credited years of domain expertise (ex: skipping the first 15min of market open)
6) 🦄 Peter Steinberger is basically a god in that room. Also the room had 2021 crypto energy - not sure if that's good or bad
7) Token usage is a problem - one person spends $1-2k/month on OpenAI, said he’s going through ~1B tokens a day across all his claws. I may be misremembering & it was 1B per week, but pretty sure it was daily
8) People want proactive AI - one guy gets Discord messages and doesn't know or care if it's human or AI
9) I asked if folks feel agency → fully in control and completely out of control at the same time
10) I'd love to see more women at these events. Fake promises of AI democratization hit differently when standard tech is ~25-30% - this room was maybe 5%.
11) Everyone said sleep has gotten worse since harnesses came out (though half wondered if it's just the state of the world)
12) Agents still lie - telling you they finished tasks when they didn't. Fixes: secondary agents, human checks, structured outputs (ex: reference an issue number for bug fixes)
13) A hackathon winner (neuroscience PhD) built a full lab management dashboard with data analysis and ordering - had never coded anything a few months ago
14) Everyone agreed prompting is dead. Disagreement on what replaces it: context engineering, harness engineering, goal-based inputs
15) People love having AI interview them for big builds. AI-led discovery is the preferred interaction mode
16) Watching agents interact with each other was a highlight. One AI posted in Slack that it ran out of tokens, another AI told it to take a deep breath
17) Agents upskilling agents: one AI shared skills with its agent friends via GitHub
18) Several speakers had OpenClaw build their presentations live during the event. One had it code a phone clicker so she could advance slides from the audience
19) Agent welfare wasn't a priority topic - some "I could kill this agent whenever I want" energy, not "gracefully sunset"
20) I asked if it felt like work or play → "it's like a puzzle and a video game at the same time"
Also hosted a Claude Code meetup this week with TENEX.AI / Alex Lieberman & JJ Englert and learned equally helpful methods, frameworks, and insider tips.
Surround yourself with people going deep into this stuff - it will pay dividends throughout the year.