surgical_fire

You think that Google throwing its weight around to prevent competition to even exist (however small it may be) is fine because otherwise Apple may gain share and yet you have the boldness to call anything harebrained.

This is rich even for the very low standards of HN.

Moronic take of the day. Quite a feat.

> They have infinite money

More like they have infinite debt

Just to drive the point home, in the real world, if the rebar fell because of lack of quality control, the company that made it would be responsible. If it was for lack of maintenance, the driver would be responsible.

The language of a rogue rebar is as absurd as the language of rogue agents. OpenAI is horribly negligent, and in a sane world its administrators should be facing legal consequences.

No, just lead an assault on companies that abuse their market position to kneecap any possibility of OSS being deployed meaningfully.

This actually is a brilliant idea.

This is where governments should slam them with regulations.

Absolute anti-competitive behavior. "Android is open, but not really"

I bought a refurbished Thinkpad from 2021 to be my daily driver. 32GB RAM. Paid a few hundred bucks for it. I didn't do it for environmental purposes, I just figured that the specs were more than good enough for general use and software development. Has the added benefit of not becoming e-waste.

Works like a charm, does everything I need, just not very good for more demanding games due to lack of a dedicated GPU.

The old tech thing is maddening. In no small part that is consumerism pure and simple, but there's also ridiculous planned obsolescence at play.

My cellphone is inching towards 5 years, and it is perfectly serviceable. I could likely keep using it for years more. I am having to look at a replacement in the near future because there will be mo more updates for it.

And for computers, most people could probably make do with machines that are many years old and not feel a difference in their capabilities.

I lived in the Netherlands for a bit, where I used to do a bike ride to the grocery shop.

Now I live in a more car-centric place where the only practical way to do grocery shopping is by car (unless I am to walk back to my house looking like a pack mule). I am not brave enough to ride a bike on the road.

If I were to choose, I vastly prefer the bike (not even for environmental concerns, it was just more pleasant). But things just aren't built that way.

> far left

What the fuck are you talking about?

The UK was governed by the Tories for 14 years, and then by Labour for a couple of years.

And if you followed anything during the time Kier Starmer was the PM, it was more or less a tory-light government.

On my small personal projects I seldom find any mistakes. I use Chinese models on those. GLM, DeepSeek, MiMo.

At work I use Claude, but it is a considerably larger and quite complex codebase. I use LLMs a lot, but it is a common weekly occurrence for me to correct misconceptions, bugs, or overengineering from Opus/Sonnet (Opus plans and reviews, Sonnet implements).

This is the same sort of fake safety concern from the previous bullshit whistleblower that plays on "AI is super dangerous" from last time.

Sorry that I don't take it seriously when the whistleblower parrots the narrative the CEOs of those companies are already espousing in the desire to amp up hype for an IPO.

This person should be shamed.

Correct. I customized my harness so agents communicate with each other through markdown files.

I find that it makes coding and review at the same time less prone to errors and cheaper

I was almost replying to your previous reply by saying something akin that with the text-to-speech you can be illiterate, and that should be a reader's utopia.

Then I got this feeling that there was an odd chance of you being sarcastic.

It's difficult to tell on HN these days. Some replies I read should be sarcasm, but I those bad takes are stone faced.

Not sure I see the problem there.

Even in projects where I use LLM liberally, I would rather not read or have to engage with your LLM output.

I use LLMs already, I can make do without yours.

One day I may get there, but even for minor PRs I constantly find important issues that I need to intervene.

Neither, I use different sessions for each step.

Username is adequate.

GLM-5.3-flash is my implementation model after GLM-5.3 writes the plan.

It's an excellent workhorse. When I am running out of my GLM quota I switch GLM-5.3-flash to DS-4.1-flash.

> Why? Why is that good?

It solidifies further that the US is an enemy to Europe, a hostile country that has a desire to undermine and destroy us.

> What if supply stays low and prices stay high?

This will have the long term benefit of making EU countries rely less and less on a commodity that it has no control over.

> and Russia takes the chance to invade while Europe has nothing to power their industry and tanks?

Russia can't even fight a respectable war against Ukraine. Now that Ukraine gets to actually blow refineries in Russia and explode shit in Moscow, even less so.

> Why must the energy transition be "move fast and break everything" while the industry and households suffer under the strain of high prices?

That's how changes happen. Without challenges everyone stays complacent.

Pi is a great harness because you can customize it with the building blocks that make sense for you.

I think that harnesses for LLMs are as important as IDEs for languages.

I hope this time around European leaders grow cojones, however small they can be, and start treating the US as the enemy hostile country it is.

It will inflict short-term pain in Europe, but in the long-term it will be better. It's like ripping out a bandaid.

That's a fair point, but US companies do that all the time, especially startups.

What annoys me is that when this discussion is about US companies it is taken as just "that's how business is done" and accepted as a perfectly voable strategy but when it is companies elsewhere it is suddenly evil, icky, and predatory.

Yes, and apparently they need several trillions of revenue to make the investments make any sense.

I'm not the one out of my depth here.

Feel free to have the last word. I prefer to read idiocy in homeopatic doses.

rofl, and I am the one that doesn't know what he is talking about.

> Deepseek has low prices and despite that their profit margin at the beginning of this year was a whooping 82.9%. Since then, they have significantly raised prices.

DeepSeek increased prices substantially not long ago. I find their profit margins hard to inspect considering I have very little idea what sort of environment they may get in China (from cheaper energy to government subsidies). I honestly doubt you have any insight here as well.

> You can actually check the approx. financials of OpenAI and Anthropic.

No you can't. They are not publicly traded, and they constantly and selectively leak bullshit metrics, from extremely unclear ARR, to extremely deceiving EBITDA. You willingly eat their bullshit and call me a picky eater in return.

> There is no reason to believe why OAI/Anthropic wouldn't have a much better profit margin than DS, taking into account a much higher prices.

I see no reason to believe (much less any actual evidence) that OAI or Anthropic have any path to profitability.

If inference (particularly for subscriptions) was in anyway as profitable as you claim today, they wouldn't need private investment rounds like crazy nor they would be desperate to offload this hot potato in an IPO.

82% margins lol. Are you telling me that if you created a machine that turns 1 dollar in 5 what you would do is dillute your ownership of the machine instead of using these fabulous profits to expand the business?

Piracy.

I have been experimenting with self hosting Navidrome for that purpose.

Almost correct, separate sessions can communicate with one another. In my case, planning and coding communicate by writing files locally.

Er, no?

I mean, not in Pi at least. In Claude Code it is a chore.

Eh, I find managing sessions in Pi extremely easy. In fact, I customized how it mages session as part of my workflow, and how agents in different sessions communicate with one another.

It's part of why I really dislike to work in Claude Code, and found it too unwieldy. There I have to keep dancing around it to manage the context in a sensible way.

> You don't have cases where the main session has important planning stuff but the actual work to execute has so much crap in it that context compaction will probably dig into the important plan stuff too much and make it too lossy?

For that is it not better to have separate sessions for planning stuff and doing actual work? Pi is super flexible with session management, and a lot of that can be automated by its extension system.

It's actually the main reason I chose Pi.

I did create some extensions where it spawns sub agents for specific tasks, especially when I want to keep the context clean or when I really want to offload a piece of work to a cheaper model. And for that I have a high degree of control over, I know which model is being used for each subtask.

I find Claude Code too unwieldy for my tastes. Pi's philosophy of being very light on features nut highly flexible for customization, clicked very well for the way I work.