First batch of alpha invites going out for lifebuild.me
Nervous, but excited! Feels so janky right now and there's a million things I want to fix, but really important to just get it out there.
The Software Factory solves an operational problem: how do we orchestrate agents to deliver high-quality software.
Once you know what to build, the factory gets to work.
It doesn't solve the alignment problem. How do you keep the factory fed?
Let me explain.
x.com/jessmartin/...
New from @jessmart.in:
"The job of building software is going to change so radically as to be almost unrecognizable to someone who wrote code professionally in the 2010s."
sociotechnica.org/notebook/sof...
We're cooking up something fun...
We will be writing more about this soon!
So, I spent two weeks trying to be my own personal executive assistantto build a stronger mental model of an AI admin. Ouch! Ouch! Ouch! sociotechnica.org/notebook/lif...
We are cookin'!
At @sociotechnica.org we've been working on an "agentic environment" called Work Squared for a few months and it's starting to feel truly useful!
youtu.be/1yfpGT6B50w
Core concepts:
- Agents and humans can collaborate on all of the work products
- Make it visual - see everything that's happening
AI advantages existing platforms.
The more data/tools you have, the more AI can do. Incumbents will sherlock every innovation.
Yet AI also requires rethinking software from scratch.
What will prevail: platform dominance vs startup disruption?
notes.jessmart.in/Lab+Notebook...
Good news: In my pursuit of work life balance, the scale is balanced at 50 / 50.
Bad news: The scale can't handle this much weight.
Can all the latest AI model breakthroughs now save me from myself?
WorkSquared Labnote #5: sociotechnica.org/notebook/ove...
Many people get very little out of AI chatbots.
What if we applied old school "edtech" (Kay / Piaget / Bruner) to humanize the AI experience?
WorkSquared Labnote #4: sociotechnica.org/notebook/sym...
Ran multiple Claude Code instances in parallel yesterday to speed up development. Wild experience with some surprising challenges.
The short answer: yes, Claude can definitely asynchronously get stuff done, but the setup is way more complex than expected.
Thread on what I learned 🧵
"Digital technology is infrastructure. Infrastructure is powerful, it shapes what runs on it and governs its users. Authoritarian infrastructure leads to an authoritarian society."
The crux
A 6 hour innovation sprint with Claude 4 Opus.
Maybe 4 days worth of work in 2023?
A play-by-play write-up with example prompts. sociotechnica.org/notebook/pla...
Moving my @sociotechnica.org projects from Google Docs to @obsidian.md + Relay from System 3[1].
Been using Obsidian heavily as a solo user for years now. Excited to try as a team. Also added an "agent" account.
Wish me luck!
[1] system3.md/relay
I was seeking UI inspiration for an AI native workspace. Plot twist: the best UI education was in my childhood games. Join me on a nostalgic 16-bit journey through retro games that solve the complexity paradox better than most modern day apps: sociotechnica.org/notebook/ws-...
We have started publishing Lab Notes, sharing our process and learnings from various explorations.
The first three notes are posted to read.
More coming soon!
sociotechnica.org/notebook
Reading about Civilization I (1991) for UX inspiration.
What's next?
We're building a AI intern in this new environment and planning a bakeoff against Pat Sharpe's approach.
Stay tuned for results and more technical deep-dives...
Read the full update here: sociotechnica.org/q1-2025-update
What might Pat's replacement be like?
An invention of necessity:
👁️ Environmental Awareness: A shared data environment where the AI sees context & remembers past actions
⚒️ Tool Autonomy: Agents discover, request & create their own tools
🤝 True Partnership: Genuine human-AI collaboration
Let's look at Pat Sharpe's 3-month performance review:
❎ Needs microscopic task breakdowns
❎ Forgets everything after each task
❎ Never learns from mistakes
❎ Can't do basic math
❎ Works in a windowless room with no awareness
❎ Needs everything translated to simplified formats
You're fired. 🙅
The results: $4.33/day average over two weeks ✅
We (almost) hit our $5/day target! But what we built was still fairly dumb—essentially just fancy scripts that don't truly leverage the reasoning capabilities that make agents interesting.
What worked well: Communication!
Our LLM kept us updated through Discord with helpful and fun pirate-themed messages. No need to specify exact text—the AI knew how to be engaging and clear when reporting back.
Problem 3: LLMs are bad at math 🧮
We saw this coming.
Calculating profit margins is crucial, and AI consistently makes elementary errors. Even giving it a calculator didn't help—calculators require knowing how to use them properly!
We ended up writing custom calculation tools instead.
Problem 2: Big context windows ≠ smarter
LLMs often drew incorrect conclusions or overlooked details when parsing the entire HTML for the page.
We had to transform everything into simplified, structured formats like CSV with filtered columns.
Problem #1: Claude has a vision problem 👀
Our AI confused browser UI with website content, misread digits, and struggled with navigation.
An LLM with vision issues + your credit card = danger! We had to shift to browser automation scripts instead.
Meet Pat Sharpe, our pirate-talking AI intern! 🏴☠️🏀
We built a simple trading system with clear rules:
- Buy low, sell high-ish
- Target uninjured stars on playoff teams
- Make offers below market
- Post purchases for resale
- Talk like a pirate, arrr!
Our Q1 goal: build a system where AI agents could make us $5/day in a marketplace—a stepping stone to economically-valuable AI agents.
We chose NBA Top Shot as our testing ground. Why a dying market? Because we know it well, and handing AI your credit card is weird enough without extra variables.
Sorry HAL, it's not me, it's you.
We gave an AI our credit card to trade NFTs, then had to fire our digital intern after 3 months. Here's what happened when we tried to make $5/day with AI agents...
🧵 A Q1 2025 Update from SocioTechnica
🦋
