mid takes, you're safe here building https://teapawt.net/t.co/HvKcghlceK
Seattle, WA
Joined November 2016
- Tweets3K
- Following819
- Followers271
- Likes31.5K
Evan Reilly retweeted
planning is kinda overrated now, like, in a general sense. like the puck is moving so fast. you don’t really have time to think and plan out, you’ll miss the puck. it’s raw intuition, risk taking. you just have to move and trust your instincts. self correcting as needed.
Evan Reilly retweeted
no amount of utility from these models makes me willingly want to subject my neighbors to flood lighting 24/7, horrible high frequency fan noise and toxic natural gas emissions in half thought out and rushed build sites. If they had treated humanity as a design goal they wouldn’t have been in this mess of PR in the first place. Fucking assholes just had to consider the people when designing these things
Evan Reilly retweeted
Austin Texas sounds like an Eastern European city that just overthrew a dictator
Evan Reilly retweeted
> we dont pay high salaries in this industry for people to prompt, we pay for domain expertise
i keep going back and forth on it, an engineer in me screams at the thought of prompting my way to PMF, but a futurist knows that there's nothing special about my 5 years experience with React nodes, and Claude will be able to nail the "hardcore infra" in the next months much better than me
you don't try to steer a Waymo from a passenger seat
Evan Reilly retweeted
human-in-the-loop
└ human-on-the-loop
└ human-near-the-loop
└ human-cc'd-on-the-loop
└ human-legally-liable-for-the-loop
Evan Reilly retweeted
I'm convinced that at this point, you could have an Opus 5.5 agent re-implement all of CPython probably within two weeks without any human input, against just the tests, and probably with full compatibility with including ABI, no GIL, a jit and probably beat it on all measurable metrics. It's completely crazy.
Evan Reilly retweeted
fun fact I never passed a coding interview
failed interviews at a dozen elite startups I won't name (for now)
finally at @attio they gave me a take-home instead – team said I was the only candidate who understood and implemented the spec 100%
lesson in there somewhere
I’ll absolutely fail any coding interview. Always have. But give me a problem and a model, not even a good one, and I’ll solve it. Pity companies haven’t caught up. The good news is, most of the market will end up like this as there are simply bigger problems so solve in a software stack, and this code nonsense is just a side effect.
Be real, you’re gonna pay most devs to talk to codex or Claude. How about you test them in the actual environment. But no. Let’s find the next leetcode
Evan Reilly retweeted
technical skills are, at their core, the result of figuring out how something ACTUALLY works.
"non-technical" domains can be quite technical. the best sales person is doing countless specific things right that makes them so good.
your ability to go deeply technical in one area means you can do it again, especially in the AI world where learning through metaphor had never been easier.
I see EVERYTHING through a systems lens with my industrial engineering background. everything is a factory to be optimized.
you can do way more than you realize if you let go of the idea that there are rigid paths in life you need to follow.
Evan Reilly retweeted
AI may well be able to write massive software projects “from scratch”, but that almost certainly isn’t the fastest / highest performance / highest security / most token efficient way to do things. Good building blocks matter.
In addition to RL’ing the agents on final task loss, we should try to get them to regularize the building blocks as well. What would the standard library look like that makes the agents most effective?
It is likely at the limit that it turns out to be some alien organization of code, but given the existing human-bias in training, I bet that in the near term it would be lovely code that old programmers would appreciate.
I suspect the same could be said for mathematics, but that is out of my lane.
Evan Reilly retweeted
One in 46 Taiwanese babies is born to a TSMC employee. The Taiwanese gene pool is being sexually selected for semiconductor proficiency.
re-reading gideon lewis-kraus’s feature story “sand castles” on @NewYorker.
so many great details. the AI supply chain becomes a breathing, moving, growing beast, something so visceral and material.
one detail i can’t stop thinking is: one in forty-six babies in taiwan is born to a tsmc employee (?!?!)
next year’s goal: convince a magazine or foundation to fund my tsmc tradwife story.
link: Computer Chips Are the Most Magnificent Things Humanity Has Ever Made newyorker.com/magazine/2026/…
Evan Reilly retweeted
The coding interview is dead, at least for my team at Anthropic. I've seen enough to know that it no longer provides signal. I removed it from loops this month.
I have a spicy, optimistic take - I think there's an opportunity for junior engineers: felixrieseberg.com/hunger-be…
Evan Reilly retweeted
Whether or not anyone builds it... everyone dies.
The future is uncertain. Your impact doesn't have to be.
