FAANG SWE & AI realist. Honest takes on tech, finance, AI tools, and engineering life.
New York, NY
Joined March 2026
- Tweets967
- Following80
- Followers64
- Likes172
I think “software is solved” splits into two levels, and the second one is much harder.
1. Local level: AI writes individual classes, functions, tests, and small features accurately and fast. No one’s writing code en masse by hand anymore.
← We’re basically here.
2. Macro level: shipping and running big real-world systems or full SaaS products that paying customers depend on 24/7.
The jump from 1 to 2 is brutal because 90% accurate isn’t good enough. At scale you need near-perfect reliability, observability, security, and the ability to evolve the system without it collapsing.
You also need someone to take real liability when things go wrong. “AI did it” doesn’t fly with customers or regulators.
Production edge cases, integration messes, architecture trade-offs, and long-term ownership still demand real human judgment.
This is totally wrong.
Blaming the user is missing the point that (a) coding agents have been overhyped and (b) can’t reliably obey the rules given to them in system prompts and other guardrails.
Sorry, @lifeofjer, but this is YOUR failure:
1. Your failure to demonstrate extreme ownership for AI generated code; instead, you abdicated your responsibility and blamed the AI.
2. Your failure to have an adequate and predictive mental model for how LLMs work.
1/2
PR Review is a system that doesn't work because there are very few incentives to participate in it
- It's a boring, time-consuming task that feels like a chore
- Nobody commends you for reviewing code
- People perceive you negatively if you are nit-picky during PR review
for the most part, the reviewer is a glorified linter with some extensive knowledge of the codebase.
that's it.
we have updated our partnership with microsoft.
microsoft will remain our primary cloud partner, but we are now able to make our products and services available across all clouds.
will continue to provide them with models and products until 2032, and a revenue share through 2030.
The conclusions here feel wrong to me. The two lessons I see are:
1. Don't run agents anywhere they might be able to access production environment credentials - it's on you to know which credentials those are
2. Keep tested backups that are independent from your production host
The only people who believe any of this are non-coders.
I tried to build a game (an area I’m an n00b in.) The results are amusingly disastrous - I never before coded a decent game.
But I’ll crack out backend services w AI rapidly - because I coded dozens of them before…
feels like a good time to seriously rethink how operating systems and user interfaces are designed
(also the internet; there should be a protocol that is equally usable by people and agents)
WTF is Claude Code suddenly concerned about projects taking 1-2 days, or weeks, when it can generate the code in seconds/minutes.
You would expect the road to AGI is monotonically increasing, certainly not weird regressions like this.
If you are a software developer and dont use AI yet, then what are you even doing ?
AI definitely isnt perfect, but its the future
Start using it right now
I sometimes defend open source stuff that’s “vibe coded” because like why does it matter?
The common argument is something along the lines of “they don’t know what the code does. They don’t know the code quality.”
How many people who have written any code have been 100% satisfied with the code they wrote?
Stop demonizing technology.