Mitchell Hashimoto
39.7K posts
Mitchell Hashimoto
@mitchellh
Creator of Ghostty. 👻 Prev founded @HashiCorp, created Vagrant, Terraform, Vault, and others.
Los Angeles, CA Katılım Ocak 2008
147 Takip Edilen197.6K Takipçiler
@mitchellh @jarredsumner are you one of his friends because clearly this is a message about bun
English
I strongly believe there are entire companies right now under heavy AI psychosis and its impossible to have rational conversations about it with them. I can't name any specific people because they include personal friends I deeply respect, but I worry about how this plays out.
I lived through the great MTBF vs MTTR (mean-time-between-failure vs. mean-time-to-recovery) reckoning of infrastructure during the transition to cloud and cloud automation. All those arguments are rearing their ugly heads again but now its... the whole software development industry (maybe the whole world, really).
It's frightening, because the psychosis folks operate under an almost absolute "MTTR is all you need" mentality: "its fine to ship bugs because the agents will fix them so quickly and at a scale humans can't do!" We learned in infrastructure that MTTR is great but you can't yeet resilient systems entirely.
The main issue is I don't even know how to bring this up to people I know personally, because bringing this topic up leads to immediately dismissals like "no no, it has full test coverage" or "bug reports are going down" or something, which just don't paint the whole picture.
We already learned this lesson once in infrastructure: you can automate yourself into a very resilient catastrophe machine. Systems can appear healthy by local metrics while globally becoming incomprehensible. Bug reports can go down while latent risk explodes. Test coverage can rise while semantic understanding falls. Changes happens so fast that nobody notices the underlying architecture decaying.
I worry.
English
@TommyGriffith @yacineMTB Think (use AI, but think)
English
Children of Magenta Line: youtu.be/5ESJH1NLMLs
YouTube
Mitchell Hashimoto@mitchellh
I strongly believe there are entire companies right now under heavy AI psychosis and its impossible to have rational conversations about it with them. I can't name any specific people because they include personal friends I deeply respect, but I worry about how this plays out. I lived through the great MTBF vs MTTR (mean-time-between-failure vs. mean-time-to-recovery) reckoning of infrastructure during the transition to cloud and cloud automation. All those arguments are rearing their ugly heads again but now its... the whole software development industry (maybe the whole world, really). It's frightening, because the psychosis folks operate under an almost absolute "MTTR is all you need" mentality: "its fine to ship bugs because the agents will fix them so quickly and at a scale humans can't do!" We learned in infrastructure that MTTR is great but you can't yeet resilient systems entirely. The main issue is I don't even know how to bring this up to people I know personally, because bringing this topic up leads to immediately dismissals like "no no, it has full test coverage" or "bug reports are going down" or something, which just don't paint the whole picture. We already learned this lesson once in infrastructure: you can automate yourself into a very resilient catastrophe machine. Systems can appear healthy by local metrics while globally becoming incomprehensible. Bug reports can go down while latent risk explodes. Test coverage can rise while semantic understanding falls. Changes happens so fast that nobody notices the underlying architecture decaying. I worry.
English
@Jae1ya @trashpandaemoji I think this is a demonstration of newton's third law
English
@Jae1ya @trashpandaemoji pushing out thoughts while pushing out turds
English
English
@mitchellh @BeauScheier @trashpandaemoji i envy your regularity. must be getting plenty of fiber and exercise.
English
@micneeley14 Yes, I'm a heavy AI user, but I think more on the rational side.
English
@mitchellh I’m on team “AI is properly hyped”. I think those dismissing the AI revolution are gonna have a difficult time. Maybe I misread your first post sorry 😅
English
@micneeley14 All secular trends are overhyped. It isn't a reason to blanket dismiss.
English
@mitchellh I think “AI is overhyped” is the more likely psychosis
English
@zeeg You are not one of the alluded to people, nor is your company. And also, just for the peanut gallery, I also almost exclusively write the code using agents nowadays (but I don't outsource the thinking).
English
There is a lack of grounding with many folks. HOWEVER, counter point.
Most strong leaders arent very public with their opinions - especially on this platform - and a surprising amount of them are very grounded in the way they're adopting things. They look at the information the same as we do, have made similar well-reasoned decisions, and are cautiously moving forward with increasingly forward looking approaches.
I think its important folks continue to push back on too-futurist opinions amongst corporate leaders (and others), but moderation is always the key component. Facts change over time, and you should be willing to change your way of thinking as they do.
Two years ago there's not a chance in hell I would be generating code with an LLM (in any meaningful degree), but now I do it exclusively.
My adoption and stance have changed over time as the technology has changed, but there are also many problems that I still see with the tech today that were very visible six months ago, let alone two years back. I dont see a lot of evidence for some of the things that have become popular opinion (the personal fav being engineering being a less valuable skill).
I think the biggest issue that is pretty widely visible is that the public discourse is often wildly disconnected from the reality inside of these organizations. "AI is causing layoffs", "we're not hiring new grads", "engineering is dead", etc. Those are toxic, and do not reflect real truths.
Mitchell Hashimoto@mitchellh
I strongly believe there are entire companies right now under heavy AI psychosis and its impossible to have rational conversations about it with them. I can't name any specific people because they include personal friends I deeply respect, but I worry about how this plays out. I lived through the great MTBF vs MTTR (mean-time-between-failure vs. mean-time-to-recovery) reckoning of infrastructure during the transition to cloud and cloud automation. All those arguments are rearing their ugly heads again but now its... the whole software development industry (maybe the whole world, really). It's frightening, because the psychosis folks operate under an almost absolute "MTTR is all you need" mentality: "its fine to ship bugs because the agents will fix them so quickly and at a scale humans can't do!" We learned in infrastructure that MTTR is great but you can't yeet resilient systems entirely. The main issue is I don't even know how to bring this up to people I know personally, because bringing this topic up leads to immediately dismissals like "no no, it has full test coverage" or "bug reports are going down" or something, which just don't paint the whole picture. We already learned this lesson once in infrastructure: you can automate yourself into a very resilient catastrophe machine. Systems can appear healthy by local metrics while globally becoming incomprehensible. Bug reports can go down while latent risk explodes. Test coverage can rise while semantic understanding falls. Changes happens so fast that nobody notices the underlying architecture decaying. I worry.
English
> If I review 100% of the code I've produced by an agent to 100% understanding, I too will find much to dislike.
Its not that I dislike it stylistically. I stopped correcting that long ago. Its that its actually wrong masquerading as something that works. This happens every day!
And yes, I do react by building additional tooling, augmenting system prompts, using a mix of models for review, etc.
Besides my own personal work, I see this in our OSS issues/PRs all the time (again, daily or every other day). We're outsourcing the thinking way too much.
The illusion of it works and the reality of it works to an experienced eye still have a wide gap.
English
Yeah, I understand the difference in shipping to different kinds of consumers. The ability to move quickly doesn't change the need for taste, or what happens when you overstuff a product.
If I review 100% of the code I've produced by an agent to 100% understanding, I too will find much to dislike. Because as a human with 30+ years of experience, I have developed taste and patterns that really matter to me (and that no doubt impact the quality of my work, especially when I'm the one doing it.) The domain you work in matters for this, but ultimately - questions of taste start to fall away, and instead you're looking for high quality repeatable patterns. External quality remains the immovable bar. But internal quality takes on a very different character.
English
@mitchellh If the choice is between preserving what we have now and heading into the danger I choose the second option
English
@BeauScheier @trashpandaemoji I type really fucking fast, and I write my posts with minimal edits in one pass. People who have worked with me have seen me do it. I also poop fast.
English
@mitchellh @trashpandaemoji You either type fast or had an enormous lunch
English
This works better in infrastructure because you can update online systems such that the MTTR hits every user within a reasonable time-bounded window. When you're shipping software others integrate with or run on their own (libraries, desktop software, mobile apps, etc.) it doesn't matter how fast you can move because its going to be limited by host fast the user moves.
I also disagree about the quality of work of machines building machines. And I say that as someone who probably hasn't written a line of code in... over a month? But I've reviewed every line of code I've produced by agent to 100% understanding (not locally, but also within the whole-system architecture).
English
It's fair to worry. The last sentence is the key to relaxing, though - "changes happen so fast that nobody notices the underlying architecture decaying". Assuming you have an architecture in first place (which is the very first thing working on pure vibes destroys) - you will 100% notice when the underlying architecture is decaying. Yes, the agent will let you live with it for a little while - but soon enough, software that is poorly factored, has the wrong boundaries, or wasn't designed to extend in the way you need it to bites the same way it always did. Just faster. :)
But what actually can happen is the opposite - changing architecture becomes possible in a way that it has never been before. When you realize the seam is wrong in the design, pulling it into a new shape is remarkably low risk - if you've been building a machine to build the machine. If it's always starting from a blank page + hopes and dreams, it'll be tough. But if you have a strong internal architecture, vocabulary that makes sense, patterns that repeat - you aren't just piling slop into a garbage palace.
It is not evenly distributed yet - but you can crank out highly coherent, well architected software.
English
@trashpandaemoji Trust me my tweet I post while pooping isn’t gonna stop them
English
@ascetic_tweeter I have, they won't listen. But I think its an important point to share more broadly.
English
@mitchellh If you greatly respect them (idk why), then wouldn't you go to them in private to share your concern? And if this is impossible because of psychotic reactions, then why isn't it skillful to lose respect for the psychotic?
English
Long term reputation means more than short term eyeballs. Bun isn't a flash sale, its already a respected and acquired startup. You don't want to be the internet clown that dances around for views, you want to be the respected elder teaching the next generation what the future looks like.
English
@jarredsumner @mitchellh nah this way the blog post will get many more eyeballs
never underestimate the power of a big controversy
English
It isn't unexpected that the focus of the Bun Rust rewrite is on the anti-Zig side more than anything, since the internet loves to hate. What is unexpected and unfortunate is that leadership within Bun hasn't tried to steer the conversation away from that at all.
There are so many positive and interesting takeaways from this and I'm not really seeing any of them pushed as the primary message.
A positive thing that hasn't been talked about at all is how far Bun came thanks to Zig. And even if you dump it now, its meaningful for how good Zig was to even build a product to this point and impact by any metric. I would've loved to see anyone in leadership say this.
On the interesting side is how fungible programming languages are nowadays. Programming languages used to be LOCK IN, and they're increasingly not so. You think the Bun rewrite in Rust is good for Rust? Bun has shown they can be in probably any language they want in roughly a week or two. Rust is expendable. Its useful until its not then it can be thrown out. That's interesting!
There's been a lot of talk about memory safety and no doubt Rust provides more guarantees than Zig. But I'd love to see a better analysis of why Bun in particular suffered so much rather than take the language-blame path. How could engineering as a practice been more rigorous to prevent this? What were the largest sources of crashes other programs should watch out for? How does Rust prevent them? How could Zig theoretically prevent them? That's interesting.
I know the official blog post hasn't come out yet from Bun. But they're smart enough to know that that PR would stir up controversy the moment it opened, or they should've been. And plenty in the company have been tweeting and writing about it. Its somewhat telling to me in various dimensions what they chose to talk about first.
I tend to think I'm pretty good at corporate PR/comms (especially when it comes to developer audiences) and I think appealing to the negative is never the right long term strategy; it does work to get short term eyes though.
English
@mitchellh @astrostl I also thought the Ghostty blog was weird. I love your work, I just thought it was weird that a post was made with no mention of a replacement, or even a remark about what alternatives are being considered.
English
I have no problem with the anti-Zig angle or negative opinions against Zig. There's a lot more interesting things here to talk about though and its a shame those aren't leading. I also think the comms burden of a nearly trillion dollar company are significant different than an individual like me but I'll take the compliment of being compared in kind.
English
@mitchellh Funny seeing complaints about “anti-Zig negativity” now when he spent years mocking Rust and calling the Rust community insecure. Bun itself heavily pushed the Zig angle. You can’t market that for years, rewrite into Rust over stability issues, then expect nobody to talk about it
English
@misterfitzie Nope, not at all what I said. Totally respect what Bun did.
English
@mitchellh bun does something opinionated, and the tech world shits on them and tells them that bun being attacked is okay, because it's bun's fault for having an opinion or whatever.
English