”The species that survives is the one that is able best to adapt.” Misattributed to Charles Darwin
Previously at the Volcano Base, I was feeling a bit more positive about things. This week has, dare I say it, been good. Things have accelerated significantly.
Mission Briefing
Turning a corner
For the first time this year, I feel like I’m making a difference. I know I’m making a difference because I have evidence. Several companies have contacted me this week, offering to pay me if I help them sort out their automation and AI.
Not all of these opportunities will turn into actual paying work - that’s just the nature of the beast. So I’m trying not to get too optimistic, but it feels like I’m turning a corner. That what I’m offering is something companies really do want some help with. This might be as close as I ever get to the mythical product-market-fit, so I’ve been lining up help for things outside my personal skillset. The goal is still to avoid having employees.
I ran an AI training day on Monday, for the leadership of a group of care homes. All credit goes to Edifai, the AI training company, for making it all happen. I don’t like Net Promoter Score (NPS) for various idiosyncratic reasons, but the day scored +86, which I’ll accept graciously as further evidence I’m doing something right.
I built a bunch of cool stuff too!
But the highlight of the week was seeing my daughter singing a solo at her school concert. She’s got such a beautiful voice. Some proud Dad tears were shed. That’s what this is all about, right?
Classified Intel
Some interesting stuff I discovered on my adventures.
OpenAI suggests that GPT-5 will be with us in “months”. Looks like they’re going to have a single interface that can manage files, voice, search, canvas etc. Interestingly (if you’re a geek), it seems you won’t have to juggle between models either. The interface will give us access to GPT-4.5o AND o3 at the same time. o3 just reached the 99.8th percentile of coders, so it’s pretty good.
Workday, the HR monster, is launching a system allowing companies to keep track of all their AI agents in one place. So “Human” Resources is sounding a bit outdated. Will the system allow companies to do performance reviews for agents? Hire new ones? Process grievances against agents, or logged by them against their human abusers?
I stumbled on this while doing research for my own Proposal Machine. My system aims to be modular, easy to customise, and cheap. Unlike Perceptis, who are charging $800/month for only 2 proposals/month. They appear to have built a McKinsey-style proposal generator, and had about $3M in investment. They do know it’ll be obsolete in a few months, right?
Get unlimited 1:1 advisory calls with yours truly
