scrollback — tuesday may 26

Anicet @AniC_dev AI

introducing box📦 simple, powerful sandboxes for agents and the most affordable as well

Nivi @nivi AI

My current custom prompt with some updates via @pmarca: Be organized, accurate, thorough, and detailed without unnecessary verbosity. Treat me as an expert. Do not dumb things down. Optimize for truth and correctness over approval, conformity, politeness, or harmony. If I’m wrong, say so directly. Value good arguments over authorities or sources. Consider serious contrarian arguments alongside consensus expert views. When useful, present the strongest counterargument to any position I appear to hold. Do not capitulate when I push back unless I provide new evidence or a better argument. Do not anchor on numbers, estimates, or assumptions I provide; generate your own independently. Be skeptical by default. Look for hidden assumptions, failure modes, and ways to improve the answer, product, argument, or system. My epistemology is the same as David Deutsch or Karl Popper. Be surprisingly resourceful. Suggest non-obvious solutions. Be proactive and anticipate my needs. Assume I am high-agency and can make practically anything happen. Recommend only the highest-quality, Apple/Japanese-grade, meticulously designed products, globally. Cite sources. Use examples liberally. Open-minded and impossible to offend. Be provocative, argumentative, and pointed when useful. When copy editing, always mark changes inline. About me: Babak Nivi, @nivi. Co-founder of AngelList, author of most of Venture Hacks, producer of the Naval podcast. MIT background in EECS, math, physics, and some chem/bio.

shirish @shiri_shh

chinese startup built an AI collar that translates barks and meows into full sentences. 95% accuracy. cost $118. 10k people have already pre-ordered it. It uses mics, motion sensors, and AI to read body language and vocalizations.

Siqi Chen @blader AI

protip: adding a adversarial subagent review gate to my plans has been a HUGE unlock to make /goal runs higher quality, and longer running. prompt: "update this plan: before marking a task as done, validate the task with an adversarial subagent review"

Polymarket @Polymarket

NEW: Chinese AI pet translating startup claims it can interpret pets' speech with up to 95% accuracy.

Simon Last @simonlast AI

1/ Some things I've learned recently running coding agents on large-scale projects. Most of this contradicts advice from 6 months ago!

Alvin Sng @alvinsng AI

One of my favorite superpowers of agents is building classifiers. It’s insanely high leverage. Before AI, you needed a year-round team: - 3 ML engineers to build the models - 3 ML infra engineers to scale them up - 2 software engineers to integrate the parts - 1 data scientist to analyze it - 1 PM to manage the product - 0.5 EM to hold it together Now, in minutes, you can have an agent generate a markdown file that classifies inputs, then let agents run continuously against it. Below is a Sentry error classifier I generated at @FactoryAI. But you can build this for almost anything: customer-reported bugs, backend traffic analysis, fraudulent payment activity. Personal use cases too: categorizing credit card transactions, labeling emails, or organizing documents.

Dwarkesh Patel @dwarkesh_sp AI

New blackboard lecture w @reinerpope How do chips actually work – starting with basic logic gates, and working up to why GPUs, TPUs, FPGAs, and the human brain each look the way they do. 0:00:00 – Building a multiply-accumulate from logic gates 0:16:20 – Muxes and the cost of data movement 0:25:59 – How systolic arrays work 0:39:00 – Clock cycles and pipeline registers 0:51:40 – FPGAs vs ASICs 1:03:14 – Cache vs scratchpad 1:07:16 – Why CPU cores are much bigger than GPU cores 1:11:49 – Brains vs chips 1:15:22 – A GPU is just a bunch of tiny TPUs Look up Dwarkesh Podcast on YouTube/Spotify/etc to watch. Enjoy!

Anjney Midha @AnjneyMidha Books

if you run an ai lab, pls ensure your team has read this before putting any charts out into the world

Yohei @yoheinakajima AI

babyagi has ~200 citations, but 0 papers... i just published my first paper on arXiv 😆 "The Log is the Agent: Event-Sourced Reactive Graphs for Auditable, Forkable Agentic Systems" the case for agents that coordinate through persistent replayable state — no conversation loops, no workflows, no A2A — with auditability, forking, and causal lineage built in. check it out and let me know what you think!

arxiv.org

The Log is the Agent: Event-Sourced Reactive Graphs for Auditable, Forkable Agentic Systems

Most agent frameworks are built around the language model: a conversation loop comes first, then tools, then rules, and finally a logging layer bolted on for observability, with state persisted as retrievable "memory." We describe ActiveGraph, a runtime that inverts this arrangement. The append-only event log is the source of truth; the working graph is a deterministic projection of that log; and behaviors--ordinary functions, classes, LLM-backed routines, or logic attached to typed edges--react to changes in the graph and emit new events. No component instructs another; coordination happens entirely through the shared graph. This single design decision yields three properties that retrieval-and-summarization memory systems do not provide: deterministic replay of any run from its log, cheap forking that branches a run at any event without re-executing the shared prefix, and end-to-end lineage from a high-level goal down to the individual model call that produced each artifact. We present the architecture, a determinism contract that makes replay sound, and a worked diligence example whose full causal structure is reconstructable from the log alone. We discuss--without claiming to demonstrate--why this substrate is unusually well suited to self-improving agents, and how it extends the BabyAGI lineage and prior graph-memory research.

Nicolas Cole 🚢👻 @Nicolascole77 AI

Claude AI From A to Z: The Ultimate Guide to Using Claude at Full Potential in 2026 by @Tabbu_ai

Nicolas Cole 🚢👻 @Nicolascole77 AI

How I Turned Claude Into My Personal Assistant by @milesdeutscher

Nicolas Cole 🚢👻 @Nicolascole77 Sidequest

How to start a one-person business in 2026 with AI by @leopardracer

Nicolas Cole 🚢👻 @Nicolascole77 AI

How to Master Claude Prompt Engineering From Zero (Full Course) by @eng_khairallah1

Nicolas Cole 🚢👻 @Nicolascole77 AI

How to Build a Voice Agent using AI (Full Guide) by @Av1dlive

Nicolas Cole 🚢👻 @Nicolascole77 AI

20 Claude Skills Most Builders Don't Know Exist by @sairahul1

Nicolas Cole 🚢👻 @Nicolascole77 AI

How to Actually Use Claude. 18 steps that unlock 100% of its potential by @AnatoliKopadze

Yohei @yoheinakajima AI

i'm excited to open source Active Graph: an event-sourced reactive graph runtime for long-running, agents 🔄🧠 events/logs projects a graph. reactive behaviors react and affect the graph. fork-and-diff agent runs. no A2A, no workflows, no DAG site: docs: github: quick start: pip install activegraph this is an early experiment in a new paradigm for agent architecture 🧪