AnthonyIsaacson

founders that hold ideas back like they're clutching grandma's pearls are going to lose
if you think it's smart, articulate it today or someone else will
things are moving too fast to build anything stable on an idea you have now
use it to build credibility and move on

Foundation models? negative margins.
Agents? open source.
Productivity gains? incredible.
selling ai makes no money, applying it makes all the money

cybersecurity stocks are going to zero with haste
(not financial advice)

you learn fast selling automation that small businesses don't have any processes
you can't automate something without 1 source of truth and a stable-ish set of steps to follow
they don't have that. They can't automate shit.
world is much wilder than our home offices

fear and desire are the only factors you need consider when creating something new
today's most common fears:
- employees: being replaced by AI
- founders: being left behind by competitors using AI
- execs: noncompliance with law
desires (universal, variable intensity):
- money
- connection
- sex
if your product doesn't satisfy a first-principles desire, you have nothing.

the best ai products are going to come from non ai companies
homegrown stuff that works extremely well, not bespoke apps
You’re too far from the problem space as a dev. People thrive at solving their own problems but miss nuance when building for others

if everyone is building the same primitive, only the best version will survive
that best version will be absorbed by the lowest common layer - the operating system
all of this agentic “stuff” will disappear into Mac, windows, and iOS.

if you talk to users, you find out you're building the wrong thing
but you won't do that
because you're too busy optimizing your little "software factory"

this era of X discourse is gonna age so poorly
we're all out here flexing "software factories" and "agent harnesses"
in 2 years, schoolkids are gonna use ai the way they use tiktok. None of this shit will matter. It's all old news by the time we finish typing anyway.

the best time to migrate your cloud out of AWS was yesterday
the second best time is right the fuck now

how many indonesian food pictures can you buy for $9M?

When I think these days, I sometimes catch my inner monologue constructing a prompt instead of actually thinking through the problem
anyone else notice this?
how do you think delegating our thinking to ai is going to affect us long term?