FinnLobsien

I agree with parts of this in the sense that I find autoplay on Spotify and many machine recommendations quite bad.

However:

a) I don't want better-targeted ads or more convenient infinite recommendation algorithms because they create friction that preserves some intentionality in my media consumption.

b) there's some survivorship bias here. We notice the bad things because they stand out, but not all of the machine-recommended things that we unquestioningly accept.

This is an interesting concept around becoming a magnet for your kind of people. I think it works because the economics of it are such that they're deliberately niche.

They're not trying to maximize viewers and deliberately filter for the types of people they want to attract.

I've also noticed that the most wholesome, supportive communities are usually in activities/interests that promise zero status or money.

This is similar to The Oval at Ohio State University. Instead of planning walkways across a field of grass, they looked where students actually walked and then paved those routes.

I think this applies to so many things. Centrally planned things rarely work exactly as planned, especially if the people doing the planning are separate from those executing the plan.

Much of the time, it's better to observe what's working in practice (even if in a messy way) and systemize it rather than have a perfect concept that disintegrates upon contact with reality.

That's great! I think anything that's not passively consuming a screen will be known as "good for your brain", the same way that regular food has become known as "health food".

Eschewing eating ultraprocessed, factory-made food is now seen as a deliberate health choice, even though it used to be the default. I think the same is happening with the brain.

Reading a book is now seen as some erudite choice. Having a hobby like knitting, birdwatching, etc. will be some kind of "good for your brain" activity. Things that used to be the only option are now "healthy choices".

I believe that we'll see a lot of "x is good for your brain" news (with X just being any activity that doesn't involve a screen), now that attention spans are so short and everyone feels like their brain isn't the same anymore.

The same way we got gyms when people realized physical inactivity is super unhealthy.

Did Jev take off briefly and stopped getting talked about a few days later? Yes.

But that doesn't make it a failure. In the era where we have:

a) an AI hype cycle where the benchmark of "successful product" is ChatGPT, Claude Code, etc. b) a social media environment that only rewards the most-engaged content

It's easy to declare something "fell off" because it didn't immediately revolutionize how everyone does their daily work.

But by that measure, nearly everything is a failure, even if it creates great businesses. That's especially true for infrastructure that runs in the background.

Yeah exactly. Reimburse $20 for whatever AI paid plan if applicants don’t have one. Many will anyway, and one can’t expect Claude Max and Free users to compete for the same job.

Yeah I agree. The exact mechanics are fungible, but the point is that it’s definitely possible to give applicants AI access without giving them a seat on the enterprise plan.

I think there are a few interesting things here:

-If the AI part is so important, why doesn't the company use OpenCode/Kilo etc. and fund it via OpenRouter? That way you could even level the playing field by giving everyone the same access to the same models.

-I believe some of this is temporary. Just about everyone is now constantly using AI, so I believe we'll soon care more about someone actually being able to explain how they'd go about something and why rather than just "can you produce code that does this"

-The idea of "applications are a numbers game" and spewing hundreds of resumes everywhere is dying. I know extremely highly qualified people whose resumes never get looked at and recruiters who never look at applications. Everyone is making application materials with AI and HR teams review them with AI. I believe we'll enter an era that's much more manual and high touch.

I feel that beyond the proper social events (e.g. dinner parties, hangouts etc.), another important part that has declined is the little social stuff.

Having a chat with the butcher, barista, etc. or just bumping into people you know and talking for a few minutes is pretty uplifting when it happens regularly.

I feel like the containerization of life erodes this.

though I guess that's still mainly an iPhone pitch, not them acknowledging a specific competing product.

I find this page super interesting because it's what you'd expect from a B2B SaaS vendor, not from Apple. You'd never see them build an "iPhone vs Samsung" page, for instance.

I think we’re past the point where making the best model guarantees company success.

At this point a lot of stuff happens at the product level.

OpenAI and anthropic have distribution and data advantages now that are hard to replicate.

A smartphone is also a tool. If you could have a smartphone without all the compulsive checking, endless scrolling, etc. it'd be incredible.

But the smartphone bundles the productive (but not addictive) capabilities with addictive, anxiety-causing wastes of time. The latter consume the vast majority of time spent on phones.

LLMs have the same problem: they bundle so many capabilities that they enable our worst tendencies. I can use it in a precise manner and reserve the decision-making, cognitively demanding work for myself. But realistically that's not what most people are going to do, and even those with the best intentions will frequently produce AI slop because they're in a rush or whatever.

I think it’s a downward spiral:

-fewer people are asking questions because they’re asking AI instead or AI obviates the question by solving the challenge immediately.

-Asking a question is less likely to get you good answers because fewer people know the answers (or give you an AI response you could’ve generated yourself)

This could be great for documentation. Screenshots in docs are frequently useless because they show me a screen and say "click X" where I still have to search X visually. And I could just to dhat in the other tab I have open.

I believe the “vibe-code an xyz replacement using AI” works well for personal or small internal projects because you can build the 20% of features you actually use.

I’m much more skeptical about a tool being replaced at scale because AI code tends to be way more brittle as you accommodate more use cases.

The data model you can vibe code for the 2 integrations you need won’t work when you need to support 200 integrations.

Google seems to have some desire to enter the gaming world, with Stadia before this.

Perhaps it's their culture of letting engineers experiment with stuff during work hours that also led to Gmail, Maps, etc. and the mobile game ad market is certainly big.

I guess this mainly competes with mobile games, most of which are mediocre at best and aim to get a few "whale users" to spend inordinate amounts of money on coins, gems, or whatever in-game currency.

Maybe the theory is that those games are so easy to vibe-code that completely democratizing it will lead to enough of a flood of games?

And a few will stand out and become Candy Crush-sized successes?

Or perhaps it's the idea that if you can have millions of extremely niche games, each may have a few hardcore fans because it caters to their specific interests (the same way Kindle Direct Publishing enabled hyper-niche literary genres)

No, but I think this is wrong on three levels:

1. Truck drivers are not deciding to take a Ferrari instead of their truck. The company knows roughly how much gas a given delivery is going to consume and how much that is.

With AI, that definitely happens if you default to Claude Opus/GPT Sol.

2. A truck driver's value scales with gas usage. The further they drive, the more valuable.

Not true for tokens.

3. Gas doesn't create extreme outliers. One developer can easily consume 100x the tokens of another, which makes AI economics much more extreme.

> This is what happened with cloud computing. People with expertise wrote software around the software creating machine because just giving naked compute to people en-masse didn’t actually achieve anything valuable. Likewise all of the talk of enterprise agents etc, they’re all custom software that wraps around a reasoning model.

Yeah exactly. Plus, most people don't want to configure MCPs and write skills about how to correctly use CLIs. The type of person who uses Hacker News may do that to customize whatever harness they use.

A sales manager at a sheet metal firm in Wisconsin doesn't. They want AI tools that make their lives easier, but they don't know or care whether that happens with an MCP.

I guess the cloud is a great analogy because that same sales manager may have loved Salesforce because it was always available and updated compared to a manual spreadsheet.

True, I guess I was mostly thinking of people using AI in their workflow, not just AI code review and things of that nature.

> I do agree that home (or at least independently) grown harnesses might be the next logical step. Harness vendors who charge by the token have an inescapable conflict of interest here. Moderate, well-governed LLM usage isn't good for their revenue.

It's true that moderate usage harms current harness providers, but I think that's because we currently conceptualize them as AI services. I think in the near future, companies will have harnesses that centrally configure MCPs, CLIs, model routing, etc.

and they'll be much closer to an auth/permission service than to an AI service.

Yeah I feel like token usage is probably power law distributed as well.

Also this feels like a good way to detect runaway jobs. I’ve seen something similar implemented for data plans where you had unlimited data, but every 50 GB had to request another free 50 GB to prevent people from turning their phone plan into their home WiFi

You can set spending limits, but I don't feel like that helps much because everyone's accustomed to AI. Nobody would accept "We're out of usage so we have to wait until Monday" and do all of their work manually.

I believe that we're in a scenario where usage is unlikely to go down and neither are frontier AI costs.

I believe we'll see a shift to more organizations building their own harnesses with model routing logic to get central control over who can use what AI and for what.

A marketer doesn't need to default to Opus 5.5 to upload a blog article with MCP, which could be done by a model 10% of the price.

Modern product selection is so fascinating to me. We can somehow cater to super niche tastes at reasonable prices. This has somehow also led to a decline in well-made, basic products.

I wonder if that's due to social media having made the visual aspect of things more important?

This kind of AI-written article often does more harm than good. Rejection Sensitive Dysphoria is not a diagnosis (as the article admits).

That means confidently stating who this applies to, what it feels like, and what to do about it is not a great idea because it could teach someone the wrong things.

I do have ADHD, and inability to control my attention is a big issue for me.

What you describe sounds like you have the condition managed extremely well and perhaps don't tend too far into the hyperactive dimension of it.

This is interesting framing, and I tend to agree with it. I think there's a difference when it comes to smart (home) devices:

1. Something desirable but inaccessible for most suddenly becomes accessible by technology (e.g. security cameras via Nest or health data via Apple Watch).

These types of products tend to work.

2. Something you already do, but faster/better. This is where we get smart ovens, smart toaster, smart light switches etc.

In this category, we see what you describe: little to no benefit over the "normal way". Some step function changes exist (Roomba), but most are ephemeral junk that either becomes worse as cloud services deprecate or are the same thing with a tablet instead of buttons.

I don't think so. I think the use case for these smart devices here may be much more in the fast food/automation world.

> There are some models of addiction medicine that say that people with hope and connection have a much easier time giving up physical addictions than people who don’t. There isn’t enough stick in the consequences of addiction to get you out of the cycle. You need carrot to transition, and the bigger the carrots the easier the journey.

I totally agree with this. And I think this is the part that's easy to miss in the narrative around how trauma and ADHD connect. Wherever ADHD traits originate, someone's desire to change matters.

If you love your job, but inability to concentrate is hampering performance, you're much more likely to WANT to change than if your job is merely a nuisance required for economic sustenance.

Exactly. Even if we assume the distribution of traits on the ADHD spectrum in human brains has been constant over the last 100 years, today's media and work enablements have massively increased the share of people (and part of the spectrum) for whom ADHD traits inhibit their way to live the way they'd like to.