x3haloed

This is the point. hmontazeri is imagining where AI still has left to grow as a consumer product, but AI is currently more of a business product. Even without a new consumer use case, it's easy to see demand for tokens keep going up.

Here's the thing. As you're indicating, UI and UX are not the same thing, but they're both subjective.

You can ask the model to implement common patterns for both UI and UX. You can ask the model to correct obvious UI errors.

But you can't really ask the model to have UI taste unless you supply that taste, and you can't ask a model to "do UX" in a vacuum. UX work requires user testing. So any UX workflow that is going to be handled by agents must also include observational data of humans using the software.

There is no set of instructions that can do "UI/UX" without also incorporating human feedback.

I had a pre-order. The reason I canceled was because the delivered product had a lot less range and was much more expensive than previously announced.

There were other things that contributed to the decision, too: the announced "stressed skin" exoskeleton was panned in favor of adhering the body panels to the body with glue, it became clear that Tesla's profit and valuation were funding Elon's right-wing political project, and I would have preferred a solid-axle design.

The absurdity was not what was wrong with it. It's OK to be absurd. I really liked the stainless steel design. Less trail damage when off-roading and no need to worry about paint scratches causing body rust. The door wedge design? Meh. But it enabled the stainless steel body.

In the end, I want what the Cybertruck failed to deliver: a full-size electric pickup at a good price with excellent range. I currently have my eyes on the RAM REV. We'll see if that one gets axed, too.

Where do you find that T3 Code is strongest and where it is weakest?

I want to make a joke about that last line so badly.

I've had the same experience as you with Claude, and I have to choose a single subscription, so I'm on Codex. I definitely wish I had access to Opus 5.5 right now.

Codex (Pro x 1), no outside models, custom harness, custom orchestration, custom agent harness.

I avoided building anything custom until I had to, and I avoided adding more than a single agent until I had to.

My preferred way to use AI for coding has been to work closely with a single coding agent with the code open. That changed wth 5.6-Sol and a handful of skill files. I stopped needing to watch closely to keep it from fucking up the codebase. With the right request, that combination will get the job done 9/10 times.

That unlocked a new workflow: as soon as a change is requested, a coding agent reads it and gives me a PR. I want a complete proposed code change in front of my eyeballs by the time I open the change request myself.

So I have a custom solution in place to manage a handful of agents aligned with this workflow.

Custom solutions are so cheap and easy now... it's really hard to put up with solutions that are only mostly right for the job.

I literally just said that you should let the AI surface the information for you -- not summarize or draw a conclusion for you.

True. Although I generally find that to be rare.

Yeah, I don't like the direction of any of this.

Yeesh. Claude still does that? I haven't had that problem since GPT 5.

Any TV that you want. Just don't connect it to a network. Why is this hard?

OK, I like this in concept a LOT. I especially like that it's efficient and headless with an attachable GUI.

But I don't like how opinionated it is: "Therefore Prospero prime design principle: Never wait forever. Every operation has a deadline. This decision influences everything."

I respect that opinion, but I don't share it, and therefore an entire coding harness developed around that opinion is not helpful for me.

Oh, you are 100% correct about that.

I have a suspicion that eventually you won't have to be a developer to successfully create good software. I do think that familiarity with algorithm design will still help a lot. But ultimately, if your verification, testing, and observability regimes are good enough, the exact implementation shouldn't matter.

This is why, to me, it's more interesting to ask the question "how are good devs attaining outsized successes with AI" than "why are good developers wary of AI?"

Yes, there are real reasons to be careful, but that's already obvious to everyone in the field. The real gold is in discovering how successful devs are managing these risks while also retaining the benefits.

Honestly, I see good points in both of your comment sets.

But there really is a floor on how deeply a person can understand their own software. There has to be a "good enough" somewhere. As a C# dev, I can understand my entire codebase pretty deeply. And I've read enough of the .NET Core framework to understand some of the nuances of how a lot of the methods we rely on every day are implemented. But I've never once looked at the CIL that Roslyn has produced, and certainly not inspected the machine code that the runtime has compiled. I can't read or write Assembly. I don't really know anything about the x86 or ARM instruction sets. My understanding of the CPU is barely better than "it's transistors," and even lower than that, I have a vague awareness that neutrinos from space can flip bits and there are error correcting mechanisms for that, but I have no idea how they work. And yet, I'm able to write fast and reliable software.

There are super knowledgeable devs out there who pretty much DO understand the whole thing, but it's impossible to memorize every line of every version of framework that you're using or the unique alterations and upgrades to every piece of silicon that your program will run on, and as you guys pointed out, proprietary operating systems are basically completely opaque to our understanding.

So the question is really "where is the practical floor that you stand on?" And for me, personally, there is room for at least some areas of code at the application level that I'm OK with never having read myself.

I think this blog post is underbaked. I think it misplaces where the tension is, and also the skiing metaphor is attempting to simplify an explanation that's already simple.

The post describes the tension as being between leaders and developers. I don't really think that's where the tension is. I'm on a small team of all senior devs, and I've seen both sides -- skepticism and full embrace, and leadership was not pushing AI on us at all. On our team, no matter where we're each individually at on adoption, we're all pretty clear-eyed about the upsides and downsides. The upside is absurdly fast code production. The downside is losing touch with every line of code in the codebase.

There's no need to introduce the skiing metaphor, because the problem is simple: it's the devs' necks on the line when the system goes down or a customer record goes missing. When you get a call at 4 in the morning from the CEO, "Let me ask someone (AI)" isn't an acceptable answer. It's your job to know the failure modes like the back of your hand and be able to diagnose and rightsize in minutes. We've spent our whole careers guarding against those failures by agonizing over every line of code that goes into the product. With that understanding, it's pretty obvious why senior devs are hesitant to adopt something new without a clear picture of how they can be sure that the program isn't going to go down under load, isn't going to drop or mangle customer records, and when one of those things inevitably does happen, how to recover quickly. In a nutshell, job #1 for a senior dev is to account for every failure, and this becomes much harder when an automated system is pushing through 10x as much code.

I'm one of the devs who has chosen to adopt AI. As I reflect on that, I think the reason I'm willing to do so is because I have a level of certainty that I can work the pipeline to safely introduce more changes, faster. So far, I am getting mixed results, but that's not deterring me from continuing to try to reach an optimal point on the tradeoff curve.

This topic is bringing out the hallucination-panic crowd.

It's perfectly safe to use LLMs to surface underlying data and knowledge without having to trust that it's giving you correct advice.

Use the LLM as search technology: string match -> semantic search -> search intelligence.

This is especially helpful when you don't know the right question to ask. Semantics only help you find text that is semantically near the input query. Semantics won't help you when what you are looking for is connected to what you have by something other than semantics, like causation, mechanism, or interpretation.

Trying to figure out why your crops aren't growing well will likely require you to read several articles across a range of sciences, probably scoped to the specific conditions you are dealing with. A semantic search is not going to surface those first.

Allowing an LLM to chain together several sources for you to read first does not require you to ask it for a conclusion.

Decent LLMs can run on anything now: https://github.com/he-be/turbo-fieldfare

Kiwix is great, but I'd also like a language model to be able to deeply interrogate all of that information much faster than I can read.

Just arrange its harness so that its responses are grounded in the source material and require source citations for everything. Then read the citations to confirm it's not a hallucination.

I made it a while ago now, but similar: https://github.com/x3haloed/eowkit

Helps you download a local LLM and offline Wikipedia. Probably easier to just ask your favorite agent to glue something together yourself these days.

I'm curious why it's so upsetting to you. What does the existence of this repo represent to you that you hate?

I'd love to hear the story behind this. Motivation, methodology, alternatives considered, etc.

I generally agree with the diagnosis here, but:

> That isn’t a revolution.

What do you think a revolution is? Revolutions are literally complete upheavals of the system. The problem that you're really identifying is that this revolution appears to be unanticipated by the players fomenting it, and it also seems to be unintentionally targeted directly at themselves.

Wherever this goes, it's sure to be interesting.

Maor dots!! Does anyone have their dots yet?

If OP sees this, I really like how you are thinking about things. It's really hard to see forward. The target isn't fixed. Bell Labs, Apple, etc. were mostly solving for a fixed target. But now the target is a relationship between humans and LLMs, and the way that LLMs work has been evolving rapidly. I can think of a number of ways to conceptualize how things are moving, but they're probably all wrong. My best guesses about what needs to remain stable is that I don't think shared software is going away entirely, and I don't think all input is going to collapse to just a microphone and a camera. There is a floor on tasks that we want agents to do for us. I'd rather rename a file on my desktop or text my friend myself. Apple and Google have both tried very hard to abstract the file away from user reach, but it's such a convenient unit of sharing, that it might stick around for a while.

I love PocketTTS. It's stupid fast and the voice quality is decent. But it's not the highest quality.

Here’s your medal.

This website is has to be pure parody/play. Because it’s too terribly thought out to be actually used. Any model worth “exfiltrating” has absolutely no way to access its own weight files.

This basically collapses to “a different OS than Android.”

The question is why you care about that.

Now that NoMachine 10 has no free option, I’ve decided that I should try to burn some tokens on making a good, performant remote desktop system. I’m targeting moonlight-like performance with Windows RDP-like user experience. Cross-platform.

https://github.com/x3haloed/ascry

I don’t need my particular repo to be the one where this gets solved, but now that coding agents are really good, we shouldn’t have to put up with paying for shit software anymore, IMO.