NK

If I was organizing a tech team for the startup right now with a budget for three senior devs that's what I would do: * Hire two senior guys who know what they do for sure and arm them with all AI tools they need. * Find three juniors who are bright and promising and forbid them to use AI.

@ktaJ0202 @shub0414 If it makes you feel better that someone can be held accountable for the bad code -- I am happy for you. For me there is no difference about it. Especially if I have to fix something in it.

@NKA793 @shub0414 I hate this take because it is a bit dishonest. The bad human-written code before AI had a person who wrote it who could be held accountable, and chances are the person who wrote it ends up maintaining it. AI doesn't have memory and doesn't have a persistent instance.........

AI is pushing so much garbage code in production now that very soon they'll have to rehire more human than they laid off just to fix bugs created by AI and vibe coding.

@DenLoginoff My blind guess is that the ration of OSS code in Java to the whole OSS space is quite small. LLMs had no ability to weigh it enough. We have other complex and verbose languages (like Rust) where LLMs are quite good.

LLMs are so bad for Java! ..which is surprising, given how much Java code exists. Just goes to show they aren't really good at more complex and verbose languages, despite being statically-typed 🤔

@bsilone Please, look at the breakdown of added jobs. It is mainly leisure and hospitality due to the coming world cup. There are now opportunities created yet.

Like every other technological advance in history, AI is creating more opportunity than it is replacing humans

I see an economical reason for such behavior. But it does not make the situation better by any means. They do not want data center as an asset on their books and prefer to finance it through the SPV because they will have to build a new data center sooner than this one will become profitable (in 2-3 years). The old one will be "owned" by bond owners and will be at least barely profitable after being used by Google/Anthropic and others.

Please think about the stupidity 🤦 $GOOG and Anthropic could each build a fully equipped data center for 5-10 years of operation… Instead they each chose to rent a date center for ONE YEAR at the same price. They are essentially paying 5-10x the cost with absolutely zero ability to depreciate and sell the assets. THESE THINGS DON’T HAPPEN $GOOG shareholders should run, because this is incredibly poor management of capital.

@TTrimoreau I always was more tired after reviewing the code than writing it. Now I am only reviewing it.

Anyone else found that working with AI agents is more mentally tiring than just doing the work yourself?

@alex_lrz_nmv Then count it as an investment and set a limit after which you will consider it a sunken cost. Anyway this model cost is a capex, not opex.

@NKA793 Agree! This is the wise decision 😄 But what if your business is just starting 🤔

Founders, would you still use AI to code If it cost more than your rent?

The further from reality we are, the harsher will be the path back. Highly likely you are right. I also had such beliefs about crypto liberty and now we are building the confidential payments for big banks to come onchain. That's a hypocrisy. The most boring scenario will execute as always. This time it is not different.

Once the bubble pops, Anthropic and OpenAI will become the Coinbase and Block of the AI world. Mundane companies that ship narrative wrappers on mundane bytes. That the bubble will pop isn’t some apocalyptic doomsday prophecy. It’s not that complicated: AI is freakishly expensive to serve. If the returns on the other end are not justified, the bubble pops. And thus begins the decades long buildout to actually economically justifiable AI. It’s amusing how resistant reality is to our fictions and fantasies. In the peak of the crypto bubble we thought reality was going to be transformed into financial liberty and democratization for all, and network states and decentralized reserve currencies. Coinbase stood to be a multi-trillion dollar company and is now just a mundane tech startup. Today we spin similar narratives about the intellectual upheaval of AI, about the new democratization of intelligence and how everything will soon begin to orbit this new technology. At the end, Anthropic and OpenAI will be mundane IT providers with an insanely grim research outlook to make AI economically sensible and useful, no different from Google’s position in trying to make quantum commercially viable. Reality is, fortunately, pretty hardened against our delusions.

@NoahKingJr Transcribe an hour-long speech in a minute

Name something that AI can do that humans couldn’t

@JadeCole2112 I really hope they don't. It would be a GFC-style disaster. But this time government would be bailing out not the banks but the big techs which were "blessed" with low interest rate over the last years.

This is a bailout.

@paularambles You do not have wealth while it is on paper. We should see first how all this IPO book plays out. It may not go that smoothly.

kinda crazy that anthropic and miri used to recruit from basically the same talent pool and now one cohort has generational wealth and the other lives in berkeley

@maxkolysh Exactly. AI is not a customer but a tool. It won't buy your services. Customers will.

two YC teams i'm working with are pivoting. one is out talking to customers all day. the other is brainstorming with claude. you can probably guess which one is making more progress. the non-obvious ideas still come from being out in the world talking to real people.

@yarotheslav That's quite an average european salary in tech, isn't it?

Factorial pays Senior Ruby on Rails devs €60,000/year gross in Barcelona. Onsite.

You are right. But quantity of code does not correlate linearly (if correlates at all) with the speed up. It is an ephemeral effect. Quite often it comes not even because of the code but because of the processes being formalized and built in a better way after the software integration.

@NKA793 @FakePsyho If you want to catalyze a lot of processes, you need a lot of catalyzer, right?

Opening this blog post with a LOC graph makes me assume the only goal here was to boost your IPO valuation. I just cannot fathom how any engineer looked at this and thought "yeah this looks reasonable"

I prefer to look at the software dev product in a bit different. From my point of view we produce automatization and speed up for other processes. Like a catalyzer in chemistry. Code is just a tool for that. That's why I think more code is not always good. Only if it serves the purpose.

In the end, code is all that the software dev produces. That's the way they influence the world. A perfect software engineer is not the one who creates zero code. So pushing out more code may be a signal of low-quality slop, or actually pushing out more value through more good code. Anthropic seems to think it's the latter. I have the same experience with my coding productivity, so it may well be true (especially considering the quality of personnel at Anthropic).

@futurenomics Can we even define consciousness in a verifiable manner?

We were so busy debating whether AI is conscious that nobody stopped to ask if we were

@GiLgamesh111 @GergelyOrosz That's not a reason, that's a consequence and a trend at the same time. Last quarter they have shown weird paper profits that boosted their quarter and this quarter they are literally borrowing money. Seems suspicious.

@NKA793 @GergelyOrosz I see, and do you think that's a reason to increase short term profits over long term quality?

Everyone on Google’s Engineering Education team had been laid off very recently It suggests Google completely stops investing in this area… damn (Source is me: I confirmed with folks inside Google unfortunately this happened)

@GiLgamesh111 @GergelyOrosz I haven't said that it was unprofitable. I said that it wasn't that profitable. Out of 62B profits I have questions to 37B paper profits (highly likely from Anthropic) and 10B of revenue growth in Cloud services (matches the Anthropic investment). If we deduct it only 15B left.

@NKA793 @GergelyOrosz I see, so did they have an unprofitable quarter before that? When did that happen and how many did they have?