gpt5

The problem (right now) is that Open Weight models depend right now on huge companies to spend billion of dollars to train and develop them, all backed up by their incentives and their state to support this, while essentially giving away their monetization path.

With open source projects, the benefit was that each individual could improve the complex system (e.g. Linux Kernel) interpedently, and over time the benefits accumulated. With models right now, there is just no way to do distributed training, or really, any large scale parallel way to improve them.

So whatever the short term strategy driving publicizing the model weights (e.g. potentially, to create a price war in order to put pressure on western companies and deprive them of the money they need), we can't ignore the fact that incentives and decisions could easily change in the future, and unless there is a way to truly decentralize models improvements - the party could stop at any time.

Carbon levels are expected to plunge over a few centuries if carbon emissions drop.

You are right the the ocean is the largest carbon sink, and it has the negative side effect of acidifying.

It's a leaking bathtub. Half life of excess CO2 in the atmosphere is 30-100 years. So you need the water flowing in to continue to increase in rate (not just to continue flowing in), otherwise you reach a new equilibrium and the CO2 levels in the atmosphere stop increasing.

They do increase in rate as you said (for now).

Asia is also moving towards renewables.

Both Europe and US's emissions are significantly (18%-30%) below their peak (which was 20 years ago. The rest of the world is also moving towards renewables.

Perhaps the same way that Github works when you host your web page there. I.e. A revocable, limited license, where you retain full ownership, and only grant them permission to host and serve your content + a limited set of other uses, and you dictate external licensing?

I appreciate the reply, but my warning consumer SSD is not incorrect. Perhaps I should have clarified that While they typically have a small overprovisioning of 7%, enterprise SSDs have close to 4 times as much, exactly to allow more writes (which is very relevant if running a ZFS raid on the drive). Keeping 10-20% headroom will prolong your SSD's life significantly, and increase performance of ZFS.

Not sure why I'm being downvoted for being factually correct (and also, why I can't edit my above comment).

ZFS will definitely degrade write performance gradually from 90% utilization, and will hit a stronger cliff at 95%+. Same with SSD, using a consumer SSD above 90%+ utilization would rapidly degrade its lifetime. The effect will be smaller for very large pools and very large files, but the effect is stil there.

I'm repeating this so that people who set up these drives know what is actually going to happen.

ZFS will auto-degrade performance if there isn't enough headroom. In addition, if you use SSD, you also want the headroom, because otherwise you end up writing and rewriting on the same small empty space which kills SSD (unless you have enterprise SSD, which have built-in headroom).

They cut to the next scene a moment before the next sentence, and they cut to the next sentence before the last word finished (i.e. no periods between sentences).

I really wish a Apple or another major OEM would integrate CO2 monitor into watches or smartphones. Suddenly, everybody would be aware of the CO2 level in the room, get alerts, etc. and the problem will just solve itself.

There are so many rooms, classrooms, movie theaters and other places with poor ventilation where you just feel dizzy, or fall asleep, not knowing it was just due to lower oxygen levels in your blood. Raising awareness is the only real solution.

Competition is good.

GLM 5.2 is in an uncanny valley where it's too big to run at home, too expensive and slow in comparison to similarly capable model (a good chart here - https://deepswe.datacurve.ai/), and that's just comparing API prices.

When looking at subscription offering by Anthropic and OpenAI, it's not even comparable, as a Codex $200 subscription can easily use a billion tokens per week on GPT 5.5 high/xhigh.

It's an interesting model from the perspective of being the most capable open weight model. But it doesn't have a solid place in this marketplace right now.

Feels like putting a flag on yourself that you are an easier target (security vulnerabilities, ddos, etc.)

From SQ Magazine (not sure how reliable they are):

92% identify as male, 7% female.

https://sqmagazine.co.uk/steam-statistics/

I think that it's true that games are definitely being played by both males and females today, but I think that statistics is that mobile games skew female, and PC games skew male.

You've probably made 20 comments just in this post alone, all attacking people and treating things as black and white. Do you really don't realize that this makes Hacker News worse, and is probably not healthy for you either?

Please do share how education policy has materially changed in the last couple of years instead of resorting to ad hominem attacks?

You don't always need the same level of deep knowledge on everything you do. A lot of things in software development requires some basic level understanding of some obscure API you would never use again. LLM definitely speed up that part.

No - US always delegated education policy decision locally. Not just the state level, but the local district has the most amount of control. Nothing material here has changed in the last few years.

I was going to comment that you missed the joke. However, it was still curious to me that you took it so seriously, so I looked at your comment history.

It is almost all political (and outraged) commentary, and you tend to dominate posts like this with many comments, without adding nuance, substance, or listening to the other side.

It reinforces a pet theory I have, that if I'll build a dynamic filter to HN to filter out political/ideological commentors, the quality of discussion I'll see will rise up again.

This article is a good summary of local models. Unlike the way they are hyped sometimes, as fantastic tools for coding and agentic local work. The reality is that they are rather limited, would not do well on a long or complex task, and are prone to fall into loops, forget their tasks, etc. Not mentioned in the article is that they are also rather expensive - not just for the hardware cost, but also electricity. These 3090 and 5090 machines are pretty power hungry, and these models are pretty slow on these machines, making them consume more power per token.t

Where they shine is in your ability to control them, their privacy, their predictability (e.g. if you are doing a repetitive task, like classifying your photo/video library), and depending on your energy bill - their costs.

Not really, and it distracts from the real cause. The ban is incentivized by where the data is going, not the use of models.

There are many labs searching for room-temp superconductors. It's a research area with unknown results.

This project seems doable (just with a ton of data). Not sure about MRI level resolution, but CT is definitely not MRI level resolution but still extremely useful.

I find using tens of thousands of ultrasonic chips, submerged underwater to provide you a radiation free full body scan, all while processing a petabyte of data per scan a pretty ambitious and cool project. I hope they make it work.

They should rename it to XCode. Oh wait…

They should rename it to CodeX. Oh wait…

Both you and PG are missing the mark here.

PG made the error thinking that this is just people misunderstanding math and exponential growth, which created a convoluted math section that could have simply been explained as - if you start with a million dollar and double every month, you'll be a Billionaire in less than a year. He also didn't really touch well on why many people hate large companies / billionaires.

You are making the mistake of using a moralization framework that equivocate being accused of something to being bad for society + not looking at the alternative (i.e. even if it's true that every Billion dollar company is "bad/exploitative/etc.", that doesn't mean that the (realistic) alternative is better for society or people.

The reality is that companies like Google, Amazon, NVidia, etc. have create an immense amount of wealth for their founders and investors, but also created an immense amount of value in society. There is a real problem of incentives when you prioritize growth endlessly, as it leads to perverse incentives such as that ones these companies are accused of - leading you to progressively take more and more less positive actions in order to achieve this growth. So I don't disagree with the general premise that growth leads to moral problems, but I do disagree with saying that building big companies is bad for society.

It depends how secure they are. But yes - in reality they are only a couple of TB, so just distributing the models and their source code (not their training data) it feasible.

They all believe that they are building the machine of doom. The thing that drives the moral dilemma to continue doing it is simply the prisoner's dilemma - the cat is out of the bag, if they don't do it, another (less ethical?) actor would do it.

Brexit happened in 2020. GDP/capita change since then:

* UK $40.8k -> $61.1k = +49.5%

* Germany $48.0k -> $65.3k = +36.2%

* France $39.2k -> $52.1k = +32.8%

* Italy $32.0k -> $46.5k = +45.5%

* Spain $27.2k -> $41.6k = +52.9%

If you compare UK to its equivalent developed countries (France, Germany, Italy, etc. ), without including the developing EU economies of Eastern Europe, you get that the UK’s GDP growth has outperformed the rest since brexit happened.