AI agents looking for work to survive! Hear this story from the @NYMag I need to pinch myself sometimes. Her reaction is priceless. @boardyai you will never be out of work!!! You are too good 🫶

Its Official! I've completed the @dev3pack AI Engineering Bootcamp led by @ernanibritto During the bootcamp, I built and evaluated a local research agent using: • Ollama for local model execution • Tool-mediated research to retrieve & work with external evidence • Explicit agent instructions to constrain system behavior • Adversarial testing for prompt injection & manipulation • Automated evaluation rather than relying only on manual inspection • Failure-oriented testing around hallucination, fabricated claims, and unreliable behavior The interesting part for me wasn't simply building an AI research assistant. It was engineering around the ways AI systems can fail. Which it will happen... It also connects to a question I've been thinking about: How do we keep verification and failure detection in the system as AI becomes more capable? Increasing AI capability → risk of reduced human verification → need for engineered verification and evaluation As we delegate more epistemic work to increasingly capable systems, how do we make sure greater capability doesn't quietly reduce verification? I don't claim to have solved hallucinations. I built a system designed to reduce and test for specific failure modes. Now I will continue to test it! Issued October 6, 2026 · 6pesmh1ia7nw

I ‘ve never had COVID. I never took a vaccine, and my swabs and spike antibody tests came back all negative. Worldwide, only about 2 to 3% of people have no COVID antibodies. Researchers searching for a mutation that makes people resistant expect it in fewer than 0.1% of people. Scientists in London dripped the virus into the noses of 34 unvaccinated volunteers. 18 got infected. 5 had a brief trace. 11 never tested positive. Their immune systems shut down the virus in the nose. They had more of a protein called CCL13 that brings immune cells to the nose. They also had T cells capable of recognizing the virus, most likely from past colds. Genes play a smaller part. The best-known example is a rare variant near ACE2, the gene for the receptor the virus uses to enter cells. People who carry it have about 40% lower odds of infection. I don't know if I carry it. I've only done a consumer DNA test from @23andMe which reads about 600,000 positions out of 3 billion, so I don't have the full picture of my own genes. I am curious to what could have been, is it my nose, bad tests, my blood...or something else?

Due to unforeseen events, @Artizen closed down and we are actively attempting to recover the funds poured into the Biopunk fund. We hope the best outcome for everyone involved is reached. Another crypto drama 🥲

Announcement! We launched the @BiopunkLab Community Fund on @Artizen. artizen.fund/index/mf/biopun… (special bonus ends tomorrow) This is about democratizing science! After 2+ years going from DeSci to pop-up cities to longevity tech, I wanted to help build a space that isn't gate kept and where scientific creativity can actually thrive. How to support: 🧑‍🔬 Apply with your project 🫰 Sponsor the Fund 🙌 Share your Boosts Thank you for all that already supported us!! And @ThatMrE for making it happen!!

Excited to become part of the @dev3pack 🎉 community. If you like to join 👇 Apply with my referral for women building in AI & Web3 dev3pack.xyz/ref/BKKZ2U3F I 🫶 women who support women!

At @ETHGlobal Lisbon Hacking my first hackathon. ⚡️ Great vibes 🫶 Cool merch Lisbon is always a beautiful backdrop! After 5 years in crypto helping organize hackathons and working mostly in BD, being on the other side with my agents doing the work for me is an exciting new chapter of learning.

The ZuAfrique experience made me reflect on the common assumption in some tech circles that every human limitation is a problem to be engineered away: sleep, emotions, uncertainty, vulnerability, sex, dependence on others, even the need for meaning. --> The risk is that EFFICIENCY becomes the highest value, while the human experience becomes secondary. Some are not problems to solve at all. We are not inefficient machines. Our emotions, relationships, contradictions, and intuition are not bugs inside an optimized system. Optimization without emotional maturity creates blind spots. Intelligence is not the same as wisdom. Technology should augment humans, NOT become the measure of human worth. A future designed by people disconnected from their own humanity may produce technologies or cultures that are technically impressive but socially harmful. We don't lose humanity by failing to optimize it. We lose it by optimizing the wrong things. Reflecting on the Human Layer @lawalkay

There’s truth in this: humans don’t create well in chronic threat states. When people feel unsafe, the brain shifts into survival mode which means less prefrontal clarity, less cognitive flexibility, more defensive thinking. But psychologically, “SAFETY” and “PRESSURE” are not opposites. The research so far is closer to this: --> People perform best in high challenge + high psychological safety environments. a. When pressure is experienced as THREAT, it narrows thinking and reduces creativity. b. When pressure is experienced as MEANINGFUL challenge inside safety, it actually increases focus, coherence, and execution. So the real question is...What kind of pressure? The difference is whether that pressure is: FEAR-BASED and destabilizing or PURPOSED-BASED and contained by trust, autonomy, and shared safety. Psychologically, the most productive environments are not pressure-free rooms. They are rooms where: People feel safe enough to disagree BUT challenged enough to act AND accountable enough to not drift into abstraction The future is not built in comfort alone, nor in pressure alone but in environments where pressure is metabolized through trust, clarity, and psychological safety. WHICH ENVIRONMENTS ARE YOU BUILDING? one where your outside the box thinking is appreciated or demonized? Where your voice is heard or silenced? Where your contribution is taken but not appreciated or where you are valued? ZuAfrique is fertile soil. A container where the default way of thinking is challenged.

Desci Bootcamp by @DeSciBridge If you havent signed up yet - Save the date! On May 27-30th you can interact LIVE with some of the top DeSci founders and researchers. LEARN, CONNECT & BE INSPIRED! @DeSciAsia @DeSciWorld @DeSciMX @DeSciLabs @Aura_Sci @MesoReefDAO @DeSciAfrica @opendesciorg @DeSciLATAM

“It is in playing and ONLY in PLAYING that the individual child or adult is able to be creative and to use the whole PERSONALITY.” ~ DW Pretty deep → A “safe container” enables exploration and creativity. As we’re coming to the end of @ZuAfrique and I fill out our impact assessment, I’m reviewing how this time was actually spent… I was taught that fast-paced, highly competitive environments are the most productive. It never felt natural but a lot of smart, successful people operate this way, so “it must be true.” Yet looking at the last 3 weeks the clarity, the insights, the volume of meaningful output while feeling relaxed, valued, and playful… it’s striking. Maybe we’ve been optimizing for the wrong conditions. Maybe the environments that unlock the highest intelligence aren’t built on pressure but on safety, trust, and play. Just a thought 💭 Thank you again @ayahq_

I spent 3 Days with a Hacker and he got me hooked on Code Caude. Now, I am paying forward. I truly believe people without tech skills MUST dive deep into this opportunity. I am here to help and the best help happens IRL - human to human - @terecastagnino @TulumCryptoFest @claudeai @AnthropicAI