Some of you have been asking very politely. Some of you have just been asking.
GPT-6.2 Astra and Sol are now available.
Super efficient and built to perform where it counts—the real world.
https://t.co/qr2iP0GpfX
Underrated life advice: Become difficult to offend. The world is full of people projecting their own unresolved anxiety, insecurity, and stress onto you. When you take it personally, you hand them control over your emotional state. When you let it float by, you keep your power. Emotional stability isn't about feeling nothing; it's about refusing to let other people's chaos disturb your inner baseline. The ability to remain completely calm in uncomfortable rooms is the ultimate competitive advantage. Protect your peace like it's your most valuable asset. Because it is.
absolute goldmine for anyone building products or features.
every requirements doc i've ever written died the same way: the code moved on and the spec didn't.
came across this open source plugin by @Google that checks for exactly that. /egm-drift-report scans the AST and flags where checked-in specs and live code have drifted apart.
it's part of elephant-goldfish model (egm) designed by @Google , one agent holds all the context and drafts the spec with you. a zero-context "goldfish" agent reviews it, reading only the spec and the code.
as someone who writes prds, requirements and loves building products from scratch, a spec that tells me when it's gone stale is something i've wanted for years.
running it against my own product https://t.co/JcobTklT0N and would share the results soon!
https://t.co/QlvvodwdBF
100+ applications and cold emails in the last month. zero screening calls.
today is day 1 of 30 days with @boardyai trying to land an ai pm role, in public.
i already rewrote my resume and my cold emails once. same result. what i worked out today is that it's usually not a resume problem. i'm applying to pm-titled roles against people who already hold that title, and i'm getting filtered out before a human ever actually reads what i've done.
so here's what changes from today:
- stop measuring effort in application volume
- lead with VerityLoop (https://t.co/JcobTklT0N), the ai product decision tool i built and shipped a site + waitlist for. it goes at the top as proof i've built ai product, instead of sitting at the bottom as a side project
- go where founders read their own inbox (seed to series a) instead of job portals
- add adjacent titles where my enterprise transformation work at shell is the qualification and not the gap
will share what moves and what doesn't over the next 30 days.
@thatGTMdude bussiness analyst transitioning into ai product management, currently building https://t.co/48SQ3XTTMX . looking forward to use this sheet to understand problems in a particular niche of interest and build solutions to help them !
u r doing God's work btw ! thank you
personal AI agents have taken over my feed: instinct's overnight hype, meta’s Muse, openAI’s Dots and now claude's Duet.
one question i keep coming back to is "what makes personal agents different from the ones we are already using like lets say cc or codex or chatgpt with memory and preferences?"
We’re opening a research preview of Claude Duet, a new app where Claude works as a proactive agent.
It comes to you with what needs your attention, completes tasks you ask of it, keeps working when you step away, and checks with you before it acts on your behalf.
i think based on the generic job title role and responsibilities, there are certain set of skills a ba has which can make the ai pm transition easier.
You already understand users.
You already map workflows.
You already work with stakeholders.
You already define requirements.
You already support testing.
as a pm, i think there is immense alpha in understanding the model frontier for your specific product and use cases. or your domain of interests in fact.
you should be able to answer the following better than most pms for your use cases:
- what can models in each size category by magnitude (<10B vs 10B - 100B vs 100B+) do well today?
- what are the problems or workflows where each model fail?
- what workarounds can you use to overcome these failures ?
- given the trajectory, what will these models likely be able to do 2–3 months from now? how should your roadmap look given this? this is a core part of the pm job now.
hopefully soon i can share my version of responses to these questions for credit risk and trade finance domain or b2b saas products in general.
You left engineering because you were tired of:
- PMs who don't understand system dependencies
- "Product people" who can't think in flows
- Leaders who demand random features
- Roadmaps built on hope
But what if product management was actually about systems?
"Thinking in Systems" blew my mind:
@Iris_LS_Luan agree the eval suite is the spec, but i think the harder transfer is spotting which messy stakeholder requirement is actually three unstated specs.