Additionally, I cannot find any obvious examples of what it does, or how it possible relates, and none of the sections seem to offer anything even resembling an example. The entire site seems like a marketing template.

FWIW— if it is similar, or there are examples, you should absolutely put them to the forefront. I literally have no idea what any of the statements/propositions in the pages actually mean.

In the US at least we’re gearing up for massive failure in the automotive industry solely because we’re avoiding competition. Yes, there will be margin compression, but without it domestic businesses become inefficient. It’s going to be “80s/90s Detroit” all over again with bigger bailouts because at some point it’ll be too politically popular to reduce prices. When that happens the public will be the ones footing the bill.

And all that says nothing of the fact cheap labor alone doesn’t make a better car. But the fact China can both make a better car (EV) and with lower labor costs really shows how dependent US automakers are on market inefficiencies. The US, and Europe, were massively ahead in quality but that lead been destroyed.

I’m not a fan of capitalism, but if the US is going to sell it and preach it— we might as fucking well embrace it. Otherwise we’re just subsidizing the rich without rhyme or reason (other than blatant corruption and exploitation). The cost of those subsidies will be stagnation, and the outright capitulation of quality long term.

It’s a near constant regression in my workflows. “Multiple agents” got destroyed recently, and the new interface for it some sort of command isn’t as good or reliable. Then you’ve got modals everywhere[1] and truncated bits (like long branch names) that make it insanely frustrating to use.

They’re constantly changing the UI without actually improving it at all. I’ll likely cancel it and use opencode for personal stuff with Deepseek and only use it at work because I have to. There was a time when I appreciated the harness but it’s becoming less useful, or at least noticeable, over time… all the while the actual UI becomes substantially more painful and awkward to use (like @ in the “agents” window being completely unable to find a file because it’s some sort of “global” scope).

One thing that surprises me about this whole segment is that JetBrains haven’t eaten these folks lunch. Their IDEs are leagues better than VSCode but their AI integration is awful by comparison (and the bar is low). I can’t even see how much of the context window I have left.

[1] it’s insane I have to answer questions in a tiny input box I cannot resize or adjust the size of. Let alone the fact the text area I input prompts into cannot be resized. Truly feels like the UI/UX is done by people without any experience.

If Optimus fails to impress, and gain traction, I’d seriously expect Tesla to end up a subsidiary of SpaceX within the next ten years as Elon tries to protect up his net worth.

I'm wondering what y'all are using for your daily driver these days and why?

I've found myself using GPT-5.5 more than Opus 4.7 for work; which, has been a pretty big reversal. Previously, I was using Opus 4.6 for everything, and GPT-5.4 was only ever in the picture to provide a second opinion (with Grok a distant 3rd only when I wanted to throw some "chaos" into the mix). The reason I've personally pivoted, is I've found GPT-5.5 to be a bit more consistent, predictable, and tends to write in a way I find less tiresome (even if the code isn't quite as good as Opus 4.7).

For personal projects, I've started experimenting with DeepSeek V4 and have been pretty blown away by it because of it's cost to quality and I've found the 1M token window to be incredibly helpful for long-running tasks. Though I may also have an over abundance of fear of compaction during tasks. DeepSeek isn't quite as good at one-shotting things as either GPT-5.5 or Opus-4.7, but with sufficient linter/static-analysis guardrails I've found it's really hard to complain or find faults (especially at the price).

Finally, if you're also making use of reranking and/or embedding models, or anything else, to augment or perform specific tasks please share those too!

Comments URL: https://news.ycombinator.com/item?id=48166147

Points: 11

# Comments: 16

GPT curses up a storm when I talk to it, and all I had to do was tell it I think it’s fucking weird when people don’t use profanity. Really makes it a lot more pleasant to interact with, IMHO.

I would honestly be more shocked if someone couldn’t just as easily coerce them into the opposite.

I honestly assumed that everyone had a rotten time with Gmail spam filtering but I guess it’s just a me problem. I suppose that means I’m up for an interesting time dealing with it as I move to a custom domain somewhere else.

Anyone have any recommendations for providers that have exceptionally good spam filtering? Hell I’d even just settle for ones that honor “mark as spam,” because Gmail absolutely does not.

It’s like someone negotiating by saying, “I’ll waste even MORE money to build something worse if you don’t give me a deal.”

I’m not discounting there may be other advantages to doing it. I just don’t think negotiating is one.

Now you’ve got me thinking I should give composer another go because speed can be pretty darn great for more generic, basic, tasks.

FWIW— the only sure fire way to win the trade is to buy time and assume both gross incompetence and negligence when it comes action. The only caveat is if the markets tank enough, this administration will signal capitulation before hand, e.g. Trump mildly capitulating on tariffs last April after the markets proceed to relentlessly defecate themselves.

0-DTE options are typically, and for good reason, stupid gambles. But, right now they can’t even be considered gambling, because there’s zero chance of winning. Not just bad odds, but no odds. Again just signaling how truly malicious this admin is and its disdain for anyone and everyone not close to them.

80% of the US population would disagree. It really seems like you’re applying what you like to the entire population and then assuming that anything else is rubbish.

Having grown up in a rural community, and small towns, I never really want to go back. Dense urban areas are wonderful, I find huge amounts of joy in multiculturalism. The plethora of ideas, language, food, and art is inspiring. I will never get that anywhere except dense urban areas.

Demand vs supply is the crux of the affordability crisis, and the points outlined in the post you’re replying to are all valid and great ways to help increase supply.

And FWIW—- you’re absolutely welcome to enjoy and appreciate sparsely populated areas, but I really think you need to understand the vast majority of people disagree with you. Not because they’re “stuck” in some dense urban area but because they want to be there.

The existing megacorps have huge swaths of infrastructure, expenses, and requirements that require massive amounts of capex to maintain. Even if performative, Meta, Google, OpenAI, Anthropic, et. Al cannot simply layoff their entire engineering, accounting, HR, sales, and support infrastructure. Those orgs are large for “good” (historically necessary) reasons.

Now fast forward to today, and this is where I differ in opinion, it is our megacorps are the civilizations who should be scared of being discovered. Minus infrastructure providers, they are the large advanced entities which can be annihilated by someone with a decent budget and a good local model.

For ~$30k-$50k (primarily buying RTX 6000 pro GPUs and a CPU with enough PCIE lanes) “anyone” can build a system using open weight models that, and let me truly emphasize this: autonomously create functionality to compete. Previously it would take me months, or years, of immense dedication to show up after work and produce something of value. Now I can do it using excess compute on my existing workstation. No existing corporation can afford to undercut every possible idea. If I only gave 1000, 10,000, or a 100,000 users they cannot compete. That may, and I believe it will, provide more than enough capital to attack that megacorps X or Y. If I’m making $100k a month, I can afford multiple autonomous systems per month. After that initial capex, I can then hire other people to help manage them. At no point will a company with billions upon billions of dollars in quarterly capex be able to compete.

Maybe they can compete with one, two, ten, or a hundred but they cannot compete with the absolute onslaught on thousands of possible frontlines. They can cut costs, by reducing their workforce, but they’ll only be increasing their competition to save their earnings report.

And yes, I realize that the open weight models are created via obscene amounts of capital, but we’re lucky that competing nation states, and cultures, like China have immense incentive to do so. Good enough, is still good enough.

The forest may be dark, but it won’t be for much longer.

tldr; call the an ambulance, but not for me. It’s going to be for the existing power structure.

If the OpenAI domino falls, and I'd be happy to admit if I'm wrong, we're going to see a near catastrophic drop in prices for RAM and demand by the hyperscalers to well... scale. That massive drop will be completely and utterly OpenAI's fault for attempting to bite off more than it can chew. In order to shore up demand, we'll see NVidia and AMD start selling directly to consumers. We, developers, are consumers and drive demand at the enterprises we work for based on what keeps us both engaged and productive... the end result being: the ol' profit flywheel spinning.

Both NVidia and AMD are capable of building GPUs that absolutely wreck Apple's best. A huge reason for this is Apple needs unified memory to keep their money maker (laptops) profitable and performant; and while, it helps their profitability it also forces them into less performant solutions. If NVidia dropped a 128GB GPU with GDDR7 at $4k-- absolutely no one would be looking for a Mac for inference. My 5090 is unbelievably fast at inference even if it can't load gigantic models, and quite frankly the 6-bit quantized versions of Qwen 3.5 are fantastic, but if it could load larger open weight models I wouldn't even bother checking Apple's pricing page.

tldr; competition is as stiff as it is vicious-- Apple's "lead" in inference is only because NVidia and AMD are raking in cash selling to hyperscalers. If that cash cow goes tits up, there's no reason to assume NVidia and AMD won't definitively pull the the rug out from Apple.

Gross margin of zero would be mean you sell at exactly the cost to produce. Net margin of zero means you cover all your expenses including COGS. The only really difficult, practically impossible, thing would be doing both at the same time. Though, I could also see a case where you drive down net margins once sunk costs are paid and achieve both.

Doing so practically, or sustainably, in most circumstances would be uhh crazy… but it’s not impossible. Even then I think aiming for zero margin is a pretty credible tactic in eliminating competition if you can out sustain them.

TLDR; Weird? Sure. But not impossible. And even sort of likely if you’re trying to atrophy your competition out of existence.

Current frontier models are really good at generating boiler plate, and really good at summarizing, but really lack the ability to actually comprehend and reason about what’s going on. I think this sort of test really highlights that. And is a nice reminder that, the LLMs, are only as good as their training data.

When an LLM or some other kind of model does start to score well on tests like this, I’d expect to see better them discovering new results, solutions, and approaches to questions/problems. Compared to how they work now, where they generally only seem to uncover answers that have been obfuscated but are present.

FWIW— I have also not heard anything even remotely close to this at all from anyone using either service.

The more I use the cloud based frontier models, the more virtue I find in using local, open source/weights, models because they tend to create much simpler code. They require more direct interaction from me, but the end result tends to be less buggy, easier to refactor/clean up, and more precisely what I wanted. I am personally excited to try this new model out here shortly on my 5090. If read the article correctly, it sounds like even the quantized versions have a “million”[1] token context window.

And to note, I’m sure I could use the same interaction loop for Claude or GPT, but the local models are free (minus the power) to run.

[1] I’m a dubious it won’t shite itself at even 50% of that. But even 250k would be amazing for a local model when I “only” have 32GB of VRAM.