- The AI Optimist
- Posts
- The Week the AI Lobsters Came for Your Computer
The Week the AI Lobsters Came for Your Computer
From Assistants to Agents. AI is getting arms and legs.
Hugo Pickford-Wardle
January 30, 2026 • Estimated Reading Time: 5 minutes
Friends,
Welcome to The AI Optimist - your weekly dose of AI optimism. Here's what matters…
This week marks a genuine inflection point. While we've been tracking the steady march of agentic AI, this is the week it went viral. Moltbot (formerly Clawdbot) crossed 60,000 GitHub stars and moved Cloudflare's stock 14% in premarket trading. Meanwhile, both Alibaba's Qwen3-Max-Thinking and Moonshot's Kimi K2.5 dropped as open-source releases rivalling the frontier labs. Your AI strategy just got more options - and the question is shifting from "which provider?" to "what problems are we solving?"
Let's get into it.
🔥 Urgent Priorities
✅ No fires this week - use the space to experiment
✅ Open-source models are now genuinely viable alternatives
✅ AI tools are moving from "generate text" to "do things"
Use the calm to explore what's changed.
🎯 Strategic Insight
Moltbot Represents the Agentic Tipping Point
Peter Steinberger's personal AI assistant has become the fastest-growing open-source project on GitHub. Users are managing calendars, booking flights, debugging code, and running entire companies through WhatsApp messages to their AI assistant.
Why this matters: The shift from "AI that generates" to "AI that acts" has been anticipated for months. Moltbot makes it tangible. Users describe it as an "iPhone moment" - a fundamental shift in how people interact with AI. One user (@therno) simply stated: "It's running my company."
The security trade-off is real. Moltbot requires deep access to your digital life - email, calendar, files. Security researchers have already found exposed instances and demonstrated supply-chain attacks. This is early-adopter territory with genuine risks.
Your move: For technical teams, this is worth understanding even if you're not ready to deploy. The architecture shows where personal AI is heading. For everyone else, watch and learn - this pattern will become mainstream within 18 months, likely through more polished, enterprise-ready implementations.
🤔 Geek-Out Stories
JudgeAI: An AI Arbitrator for Economic Disputes
JudgeAI is an automated judicial system designed to resolve economic disputes using legal algorithms, Nash equilibrium modelling, and multimodal evidence analysis. It operates through a seven-stage process from preliminary analysis to judgment.
Why it matters: This is AI moving from "assistant" to "arbiter" - from helping humans make decisions to making decisions itself. Whether or not JudgeAI succeeds, the pattern is significant: we're now asking which decisions require human judgment and which can be delegated to machines. That question will define the next phase of AI adoption.
Anthropic Launches Claude in Excel
Claude Opus 4.5 is now available directly inside Excel via a sidebar chat. Ask questions about your workbook with cell-level citations, update assumptions while preserving formula dependencies, debug errors, and build financial models through conversation.
Why it matters: This is the "AI in your workflow" vision made real. No copy-pasting, no context switching - the AI sees your entire workbook including structure, formulas, and relationships. Currently beta for Max, Team, and Enterprise plans.
AI Mirrors for Blind People
AI applications are providing visual feedback about appearance to blind users - essentially creating digital mirrors that never existed before. Users can now ask AI to describe their outfit, check if something's on their face, or see themselves in photos for the first time.
Why it matters: This is AI creating genuine human value with no hype required. When considering your own use cases, ask: where could AI provide access to something that was previously impossible for your customers or staff?
🎨 Weekend Playground
🎨 WEEKEND PLAYGROUND
📖 Read: Daemon by Daniel Suarez
This 2006 techno-thriller imagined distributed AI agents taking autonomous action in the physical world. Reading it in 2026, after a week where Moltbot went viral, is genuinely unsettling. Suarez predicted the pattern we're now seeing emerge.
✅ Consider what "AI that does things" means for your organisation
✅ Think about where human oversight matters most
✅ Notice your assumptions about what stays human
If The AI Optimist helps you think more clearly, forward it to someone else navigating the shift. If it's not quite landing, hit reply and let me know - I read every message.
Stay strategic, stay generous.
Hugo & Ben
