100% reasonable, policy at my business is that you're responsible for what you generate, if you generate and ship code you do not understand and it causes problems, you better correct this behavior or you will be fired.
If it generates code that you don't understand, either learn to understand it, or replace it with something you do.
Anything else is just being a lazy fuck and making it everyone else's problem.
I'm interested in what he pulls off, I expect it to be a lot less exploitative and more driven creatively, I expect artists to come first, I expect the medium to be improved, not cheapened.
We'll see where it goes, but someone like him gets the respect where I'm more curious at first than if it was just a c-suite that is just looking to lay off and pay people less.
Fuck that noise, efficiency-obsessed over here -- AI was a huge gamble on that, efficiency dictates to not be an early adopter because that's always *inefficient*, there's rarely a payoff to being that early, tools change, adoption process matures, being late is usually better.
It was all FOMO and C-levels smoking crack.
Honest solution finding, always has been.
Coding is easy, I can pick up languages in days, I'm not learning much in coding techniques and haven't in a long time.
Figuring out how to design systems to be intuitive, design them to perform well, promote adoption, beat out competitors, refining process, handling feedback gracefully, efficiently moving forward.
The hardest parts, and what matters in application design, and it's why a lot of new apps just have crap adoption, coding was never the hard part, and anyone with "an idea" could learn how to do it in a handful of months.
It's still a crock of shit, 100x is absurd, no one is showing close to that, none of our investments, none of our partners, none of our teams we work directly with, not our own teams.
It helps, and in some cases can actually hinder, it's easier to be pissed about it when you're looking at money being held up by non-AI companies adopting AI, vs. AI companies selling their wares.
The teams that have adopted AI the most, are the *most* behind in their roadmaps right now, the most distracted, the most likely fucking around and not actually working on company time, the most likely to yak-shave instead of affecting the bottom line, it sucks ass.
GPT-5.5 doubled API pricing and Opus 4.7 silently inflated token burn by 40%. The math is in: it's now cheaper to hire an offshore dev armed with DeepSeek than to run a fully autonomous frontier agent. We didn't build autonomous engineers. We just made outsourcing the optimal cognitive architecture.
@HarryStebbings As somebody that runs a company in this space: AI has left us *buried* in work, many things we could never consider due to costs are now completely viable, roadmaps are getting *bigger* not smaller.
These companies are AI-washing their own failures.
hat graph doesn't disprove anything I've said -- it just says when you get taxed that high you're not willing to pull it out as income -- which is exactly the kind of shifting of priorities that helped those businesses perform better and the American dream work a lot better.
The graph if anything promotes what I pointed out, that people stop squeezing value out of businesses and plan for the long-term when short-term squeezing becomes less profitable.
Also: if "well no one ended up paying it" -- cool so we can go back to that right? If it's all a nothing-burger?
I mean to be fair, the average person, or even highly technical person doesn't know, because all the companies paying for this stuff are *very purposefully* keeping it private.
Even if you're highly skilled you're giving at best educated guesses unless you have access to very privileged information... and then you won't risk your job to leak it.
> Somehow
No "Somehow" the writing has been on the wall, a lot of MBA asshats operate in two modes:
1) They know they're full of shit, but they're basically doing PR and marketing
2) They don't know they're full of shit, and JFC why are they in charge?!
@leaving_tech The technology is fucking literally designed to be accessible to people, the skill floor is stupidly fucking low, the skill ceiling usually involves "Claude, how do I use you better?", JFC Tan has been intolerable through all of this and watching him AI 600k lines for a blog.
Here's a better question:
If it took 3 weeks on AI, it may have only been a few months to do it traditionally, how did the company commit to burning $600k/yr with SalesForce (which is honestly small for a multi-billion dollar company, surprised it's this little) when an in-house solution was so much cheaper the entire time?
This wasn't "6 months with AI" territory where traditional work may have been more costly than $600k/yr, this was *3 weeks*.
I have like a thousand stories like this, when I say that corporate is inefficient like governments I 100% mean it. I swear a significant number of seats are glorified chair-warmers.
Once had a company president sign a partnership with another company, the plan on how it would make us money: sell our product in this new market, our product didn't work in that market and was at least 2 years away from getting there. President had zero clue, SaaS BTW, he only had to know very little about his product and market, he did not.
Worked with a VC where I had to remind a founder in charge or product multiple times that we actually had to make money, so many initiatives basically went nowhere.
Worked with upper management who just Claudes things these days to the point of not following company protocol, he's screwed over multiple projects that made the company look incompetent (customers were pissed) and required me to fix it.
Worked at a company that burned through two IT directors that were glorified chair-warmers, one of them only got fired when a consultant that was hired to "fix IT" threatened to go above the president straight to the board with regards to *everyone's* incompetency.
Companies aren't logistical, they're not professional, they're emotional and full of man-children.
Met a guy at a party who called himself the “head of AI” at a mid-sized and well financed company
Oh so you build AI tools? No, I can’t code. Oh so you buy AI tools and deploy them internally? No the IT team does that. Oh so you set corporate AI strategy? Not really.
As far as I can tell, this man’s job is to
1) Be the kind of rich-looking older gentleman that boomer execs take seriously
2) Spoon-feed those execs AI takes that were ice cold on X six months ago and coach them about how to repeat them in public
This experience has radicalized me.
This person’s job is proof to me that corporate America is not just clueless about AI, they are paying lots of money to fake it
This guy’s company and many more like it are going to get obliterated by companies run by 25 year olds and staffed with agent swarms
Other cyclist here, cyclist is a douche, I got a road bike and I'll still go in the grass if needed, sure it's nice if people make room but I'm not going to turn into a piss baby because someone didn't and I still have lots of room to get around.
Also e-bikes allow you to haul ass on shared infrastructure that isn't really great for that speed.