FinnLobsien

In a way it’s good that the equation is so lopsided. A massive part is that software can be replicated infinitely for free. This is why the scale of the value created is so giant.

I definitely do think it’s crazy that someone whose software gets dozens of millions of downloads A DAY can end up making less than someone building a mediocre SaaS app and getting acqui-hired 2 years later.

For the record, I think the VC - Startup ecosystem is incredibly valuable. But it IS crazy how essential software can go essentially unrewarded.

Sure, but I also feel like at the higher level, you’re not necessarily looking for implementation alone, but for “what should we be doing in the first place?”, which AI can’t help with

The value created vs value captured equation of OSS must be one of the most lopsided things ever.

If you’re at Google and invent Kubernetes you might still capture 0.000001% (probably less) of the economic value created by Kubernetes, but you probably enjoy very generous comp.

OSS doesn’t have any of that, besides being extremely in demand as a consultant or whatever.

How would you enforce that? I'm genuinely curious. It's nearly impossible to conclusively prove someone rewrote your codebase with AI.

If you have a hobby project like writing a blog, crocheting, or almost any other creative hobby, you can dip in and out however it suits you. If you deal with major life events, sicknesses, etc., you can leave the hobby and come back. Nobody is paying you for it, so nobody can complain (maybe the friends who miss you, but it's not actively impacting the real world).

Open source is one of those weird things where your hobby project can become an essential piece of infrastructure.

It's like if you loved crocheting, but somehow if you stopped crocheting everyone in your city would no longer have clothes and need to walk around naked.

There’s one thing that I find interesting about this. I can’t judge whether this is a good product, whether it’ll work, and what it’ll add to the medical establishment, or whether more data is always more useful.

In general, I think we should applaud this though.

Any genuine attempt to create novel medical technologies is probably a good thing (assuming they’re non-invasive and non-painful).

Unless it’s a Theranos situation, I think it’s a great thing to attempt, even if it fails. So many things we rely on today are the result of a successful attempt, but the failures were just as necessary for the eventual success.

That ambition is very positive to me.

Private Equity is engaged in buying artisanal semi-businesses, turning them into businesses, propping up the numbers while destroying them —then, hopefully, destroying itself. -Nassim Taleb

I agree and I think some of the comments here have shifted my perspective on this, especially given the age.

I think with teaching anything, there’s always going to be a difference between teaching the tactics (e.g. how to use AOL) vs the fundamentals (what’s the internet, how do computers work)

Yep good point. I’m not defending using AI to do homework.

Fair point!

I would disagree that they’re adjusting fine. So much of the stuff we see now is full AI slop clearly created with the first output of ChatGPT. It’s like saying we don’t need to teach kids about the internet because older generations grew up without it and they’re scrolling TikTok.

a) what’s the actual percentage of professionals who actively use AI? It’s much smaller than we think in tech.

b) what percentage of those people understand the very basics of how LLMs work (e.g. token prediction, context windows, etc)

c) what percentage of those people understand AI Agents (or any of their ingredients (APIs, credentials, etc.)

You quickly arrive at a tiny fraction that has a real clue about what they’re doing.

I agree to some degree. But by that logic, should kids in the 2000s not have learned about the internet because the internet fundamentally changed between then and now?

I think that if anything, it’s really good I learned how to operate a computer and the Internet BEFORE what the Internet became.

I pity the generation who don’t understand a computer’s folder structure because they grew up with smartphones and TikTok.

I think it’s quite tricky. On one side, writing is a form of thinking and cognitive training.

Just NOT doing that work by having AI simulate it is not good for anyone’s cognitive development.

At the same time, anyone growing up today will be using LLMs for massive parts of the jobs they grow up to do. So they should learn about it.

I really feel for teachers/educators right now. It must be hard to remain demanding and insist on educating kids well while also preparing them for the future they’ll actually live in.

I got 75k words, which I’m happy with as a non-native speaker. Others here have also mentioned that the math may be off and that you can juice the game by looking at how answers are phrased etc.

I do wonder how much of these were “what AI thinks are hard words to know” vs. actually hard to know.

There are discussions on the forums on this. Nothing is decided yet.

It's a beta from the team that built Bear (Bear.app), which has themes. Would assume they're coming later.

I'm curious about the reasoning for this. Must've been really good terms for both sides or wanting to do well by compatriots. ElevenLabs previously raised a $500m and I don't know why they'd need $11m when they're not even fundraising

I think this whole genre is an issue of theory of constraints.

-Before the internet, space to publish anything was scarce (and so were methods of finding specific information if it did exist), so newspapers/magazines were valuable and textbooks/guides were great media.

It was difficult to find extremely specific information because a lot of information was so niche it wasn't worth publishing, or because it wasn't worth the effort to find something SO specific.

-The internet made publishing space infinite and search easy, so specific information became super valuable, e.g. "how to do x in 5 steps" blogs. They architected information in ways that wouldn't have been worth publishing in a regional newspaper or small-circulation magazine, but was economically viable at worldwide scale and when publishing was free.

This is the era this article describes.

-Then everything went mainstream and the above production of content was often more frequently done by large companies, with bloggers and other individuals filling even smaller niches

-Now that LLMs have vacuumed all of that up and you can create even more specific content for free, instantly (e.g. you no longer need a generic tutorial on "how to start a business" but summon one that already knows your prior skills etc.)

This has made the previous era of "how-to" content less useful because you can decide whether you want expert/beginner/intermediate level content and what personality etc.)

Now the frontier is first-person narratives and opinions, because even specific information is no longer the constraint.

That's true. Although how the product is sold doesn't necessarily affect how we enjoy working on it imo.

You can love writing code and working on tricky projects the same way you can love hand-making boots.

You can love managing an agent swarm the same way you can love managing a shoe factory.

I was talking about enjoyment of the work, not the economics of it.

I think your take is very balanced and true. I've also never felt as technologically capable as I do now.

I guess the most annoying part to me is the expectation of producing a multiple of the output in the same time, when what I really enjoy is the process of slowly making things.

> I still think it is though, that technical knowledge is still supreme.

Ultimately, the ability to get results is what matters most. I've seen quite a few AI maximalists who have Claude mass-produce a bunch of things that nobody needs and then have something running on autopilot that doesn't need to be running in the first place.

I work in growth now, and I definitely resonate with this. Producing content about Kubernetes operator architecture is always going to be hard.

But you have to decide whether you alienate the beginners by assuming people know what CRDs and memory errors are — or the advanced practitioners by explaining pods, clusters, and other jargon from scratch.

You can't make everyone happy.

I think emotions are an underrated aspect to the "AI changes how we work" thing. Everyone I know extensively uses AI in most parts of their work. Basically everyone has multiplied their output by a lot.

But I don't know many people who say they enjoy their work more because of AI. I definitely resonate with the pain of watching something that required so many little bits of knowledge you earned over years melt into pushing a button.

I might just be the artisan cobbler watching factories rise during the industrial revolution. Making good boots was and is still hard and requires a lot of specific knowledge, but the amount of people appreciating well-made boots plummeted as factor-made, cheaper boots flooded the market.

And I know, rationally, that the line of "what is hard" is just moving. There are great industrially-made goods the same way there are great products whose code is mainly AI-generated. Like good boots, building safe, scalable, distributed systems is still hard.

But I can't help but have some nostalgia for how things were pre-AI. Work felt more honest, the skills I spent years building felt more valuable, and I was more satisfied at work.

Super interesting that every payment company now buys its own billing system as software monetization is genuinely hard again and complex billing is hard to build from scratch.

Yes, agreed. But it's not a binary. It's not like you can either wrestle with hard engineering problems until an eventual breakthrough or exclusively pretend to be working.

Sure, on one extreme is the purely value-extractive person whose only work product is calling recurring meetings to talk about how great things are going (while having contributed little). In that case, it's an error to be corrected.

But there are different types of useful work. Let's use engineering as an example.

You could build a set of new filter options on in-product analytics that customers have been asking for. Assuming this doesn't require net-new data sources or whatnot, this is usually not a complex engineering challenge, but will be loved by customers.

Then you have types who refactor an obscure caching function to reduce its memory use 15% when performance wasn't an issue, but it was a fun engineering puzzle that made the code more elegant.

Clearly both create value, but one will be more useful to customers, although it's not as fun to solve.

My point isn't that we should treat corporate bureaucracy, performative work, and freeloading as "the way things are" (they're toxic to any work culture). I'm saying that "deeper engineering challenge" doesn't equal "more important work".

Also, I don't necessarily like this! Deep problem-solving and focused work on tricky challenges are the most fulfilling things to me and to most people I know. But this is the dynamic we're in.

(and yes you can argue that having the most efficient code possible prevents future performance issues and whatnot and is thus long-term more valuable than analytics options, but at least in the startup world, limited resources mean you tend to solve things when they become problems, not because they could one day become problems)

I think this dynamic is not specific to SWE and as old as time. As organizations grow, so does the aspect of work that's more "seeing and being seen", and rightfully so.

There's definitely a ton of cruft that accumulates, and a lot of "work" being done that accomplishes little, just to satisfy a corporate bureaucracy.

But there is a reality where "good performance" is not just about the work you do, but also about your ability to get things done practically, e.g. not just your ability to write a specific microservice, but to make a compelling case for that architecture over another, and to get it reviewed and merged.

That's not to excuse wasting everyone's time on sycophantic vanity projects that don't help the business.

But I do think there's a tendency (especially on HN and Developer Twitter) to only respect complicated engineering work (e.g. optimizing Kubernetes deployments). To be fair, I'd love to almost never deal with company politics and performative work and am lucky to be at a company where effectively zero of that exists.

But as orgs grow, so does the share of work that's more political.

I think centralization (e.g. one player having almost all of the primary supply in the marketplace) is a big issue. The issue of venues/arenas etc. as mentioned, is another. I think an underrated issue most don't bring up here is that the demand is extremely concentrated too.

Most people want to see a tiny number of musicians/entertainers/shows (i.e. there's only one Taylor Swift, Beyonce, Kevin Hart), and there are extremely few venues that can accommodate those huge shows. That supply is further constrained by the artist's time and need for physical presence, meaning it's impossible to expand supply.

This makes it extremely hard to break into the market because you need to get one of extremely few, extremely demanding customers on board.

The dynamic almost approaches that of defense contractors, where your only potential customers are a few governments in the world.

Yeah, I think this is part of the issue. Job boards are the go-to interface for HR professionals to list jobs and for applicants to find them.

Those are much worse now, and most people aren't the level of standout talent that gets inbound jobs, nor scour careers pages for hours.

> And, from what I can gather, that has been increasingly true for a decade.

I think fake job openings, lying on your resume, and applying everywhere have been a thing as long as jobs have. AI may have supercharged this.

But I don't think the structural issue is that companies advertise nonexistent jobs. Even companies I know are hiring (I know because they're paying in-house recruiters) have a tough time filling those roles.

Neither do I think that lying on resumes is cause of the problem. Even if resume spam is a massive problem, you'd expect that adaptations either in applicant behavior or in talent sourcing would emerge that would remedy this.

I thought this was interesting because there seems to be a market in which:

-Job seekers find it harder than ever to get a job -Companies find it harder than ever to find the people they need

These two things shouldn't be true at the same time, at at least eventually balance out, but it only seems to be getting worse.

I agree, but sadly it’s kinda like telling someone the best way to have fun at the casino without losing too much.

It’s an environment designed for the exact opposite of your best intentions.