David Young

Lucky Husband to @youngnerual. Dad to Ruby and Bennie. Computer Person.

[Mitchell Hashimoto @mitchellh]

Mind boggling to me that I can make a thing faster and there's always people that ask "but why?" What kind of mentality is that? The pursuit of excellence does not need justification. Also, I find in so many cases, we can't know the impact of an improvement until we do it.

For example, one I've talked about before: Ghostty's high IO throughput has enabled terminal program (emulator and TUI) fuzzing at a speed thats incomparably fast to prior solutions. This has resulted in upstream patches to resolve issues in popular projects like btop, tmux, and more.

Speed enabled that anecdotally example that lifted the tides of adjacent communities that don't rely on Ghostty technology at all. I didn't predict this.

Make things better because they can be better and let the results naturally play out.

[Mitchell Hashimoto @mitchellh]

Amongst my friends, Spotify is the lowest quality consumer app we still pay for. It certainly hasnt gotten noticeably better in the last couple years (arguably worse). So, this is not the positive look Ant and Spotify are spinning here.

Bigger picture, this is the problem with a lot of AI reporting. It reports completely meaningless metrics like deploys per day or LoC. Why don’t we start reporting consumer satisfaction reports? Actually end state research results.

All the no nuance AI people always come out and think that this is anti AI. Again, I think AI is great and Claude is great. But this is bad marketing and makes both look like clowns.

[Ben Vinegar @bentlegen]

A short story about deferring tech choices to thought leaders:

Early days at Disqus (~2010-2012), we made several frontend choices based largely on what thought leaders were promoting at the time.

One example: there was a big movement toward "micro-frameworks." Instead of larger, well-tested libraries like jQuery, you'd stitch together tiny interoperable micro libraries (Ender.js was one). Disqus was an embeddable JavaScript app, so file size mattered. It fit our use case, so we went with it.

Then it went live, and we were serving millions of users. The reality of those choices became clear. Micro libraries meant that instead of one good semi-bloated library, you ran 6-7 smaller, less-tested, crappier ones. We burned a ton of cycles fixing bugs and covering corner cases when we could've been shipping product.

We made a few choices like this.

At conferences, I'd track down those same thought leaders and ask for advice. "I'm hitting problem X, Y, Z. How did you solve this?"

That's when I learned my lesson: they rarely had answers, because they'd never reached our scale. Their energy went into promoting new stuff, not running it.

You should know this has never stopped. It's happening right now with AI. It'll happen again with whatever comes next.

Do your own homework. Test a lot. Don't just go with what somebody tells you.

[Ben Dickson @bendee983]

Loopmaxxing has all the same problems as tokenmaxxing: Its proponents single out one or a few instances where the AI agents did something impressive without human oversight and while given unlimited compute.

And from there they quickly conclude that all you need is X (in this case loops). The reality is that AI loops are very good but only if they have the right structure and guardrails.

Without a verifiable goal and proper checks to make sure the AI is actually making progress on every iteration of the loop, your agent might get stuck in a situation where it will spend unlimited money on LLM tokens without achieving anything.

And interestingly, the loudest voices promoting loopmaxxing are the same ones that promoted tokenmaxxing: the companies that are selling the LLM tokens or the hardware the runs the AI systems.

Use AI loops but be skeptical of what you see on social media.

[David Young @davidallyoung]

@ThePrimeagen If anything, being smart is even more valuable now... You have to be able to discern more decisions with credible judgment... If you aren't trying to learn and build those skills we are in for a hell of a ride.

[ThePrimeagen @ThePrimeagen]

"being smart has no value anymore."

this is the mentality that paves the road to hell

[ThePrimeagen @ThePrimeagen]

the balls are back... and they are actually pretty good

[David Young @davidallyoung]

Me and my best bud at our favorite spot in Nacogdoches, Brendyn's BBQ! @youngnerual

[David Young @davidallyoung]

@ashleyschendel Who cares, he made a decision in the moment and it is not a big deal. The man that went insane and called the cops on him is the one who needs an interrogation. The father who patiently dealt with that freak deserves a medal.

[David Young @davidallyoung]

@EvangelosNikos @ryangrim @tparsi @bariweiss

[Mitchell Hashimoto @mitchellh]

The problem with the "if it works who cares what the code looks like" mindset for agentic work is that it assumes the agent has a perfect understanding of "works." Realistically, things are underspecified, agents make bad assumptions, etc.

To be fair, agents are pretty good at unit test coverage. They're pretty bad at designing human experiences (API, CLI flags, etc.), especially cohesive ones for future roadmap plans they may not have visibility into (unless your backlog is perfect and vision fully laid out, which I doubt). They're bad at knowing where performance matters and what type (CPU vs memory tradeoffs). They're bad at where compatibility matters and where it doesn't (and tend to err on the side of preserving it without further guidance). Etc.

Unless you have this ALL specified, you can't possibly claim "it works" without taking a look and thinking about it.

[David Young @davidallyoung]

@AnthropicAI Lmao

[Gergely Orosz @GergelyOrosz]

Meta had a SEV-0 outage today… less than two weeks after Meta’s most embarrassing undetected-for-too-long account takeover (also an outage)

It’s impossible to unsee Meta pushing AI for code + reviews and the end result being more massive outages vs before

They are connected

[dax @thdxr]

i see so many articles "analyzing" the ai industry and it reminds me why it's better to be stupid than only slightly smart

these people find very convincing, intelligent sounding ways to be completely wrong

[Bojan Tunguz @tunguz]

Our Anthropic overlords deciding which prompts the peasants are allowed to use.

[Gergely Orosz @GergelyOrosz]

Trusting any single AI vendor seems like an increasingly high risk for any team or company.

When using models: use it behind a router where it's trivial to switch providers as soon as one tries to force unacceptable T&Cs like Anthropic with Fable. When using harnesses: do the same. Use ones where models are trivial to switch out, like OpenCode, Factory, Cursor and many others.

Putting all your dependencies on one provider increasingly feels like a massive business risk that makes little to no sense to take.

Unless you have a hobby project, of course. Then convenience is all that matters. But if you're a professional, make it dead simple to offramp from one provider to the other!

[ThePrimeagen @ThePrimeagen]

I think one thing that is understated is that the hard parts can feel harder with more delegation.

Not only are you attempting to solve hard architectural ideas, you also are constantly solving it in a project in which you don't have the same deep familiarity that we did say a couple years ago.

It's like you're always on your first day, but you're given the task of refactoring the universe.

At least that's something I have noticed that makes me feel a bit angsty

[Gergely Orosz @GergelyOrosz]

Seeing how SOTA models are evolving: becoming more restrictive in usage (decided by the company), less transparent (you cannot tell if the AI lab nerfed your model) + less private (your prompts are stored, no opt out) makes me much more interested in open models + local inference

[nicole @nicolekcha]

announcing 4.8, our new model that 3x overthinks and doubles response fluff while reducing comprehension

announcing HTML instead of markdown, the format which is prettier for humans but uses 4x the tokens

announcing loops on top of loops, when you still haven’t optimized the last thing, so you will spend more tokens with the false hope of better results

[ThePrimeagen @ThePrimeagen]

something just doesn't add up... Can someone help me understand these two things?