wyre

The two things that torrents are best at are downloading Linux ISOs and pirating content. So using qbittorrent to download Linux ISOs is usually code for piracy, since torrenting Linux ISOs is legal and piracy is not :)

Was reality tv the start of it? People get to turn their brains off to watch the stories media companies want to tell. I’d be surprised if it wasn’t correlated with the decline of the American presidential debates.

I don't understand how this question can be asked in good faith?

If you have to defend smart glasses based on a technicality, that doesn't reflect very well on the glasses.

Thanks, I got a good laugh out of this comment, but therapy is just a tool to mostly learn how to deal with things that are outside of your control.

What? So 10 years ago you though 1+1=Firefox is using Google Analytics, something you can't prove, but is "obvious" and "recognizing patterns"

I get wanting to be conspiratorial, but its not cool to go after others that challenge your conspiracy, even if its "obvious"

The page says their hardware team builds the fastest trading systems on the planet. Jane Street does HFT.

This is about Old-School Runescape, not RS3.

It's special because it has been active for 25 years. There are no skill trees, instead every character has access to all 24 skills giving user's complete agency over what their character can do and has access too.

The game's rhythm is based off of a tick instead of click-and-wait. This can let skilled players abuse click rates for more xp or win PvP or high-leveled PvM fights.

The game takes a lot of time. Each skill needs over 13 million experience to get to max level, this requires significant investment for each skill, letting user's grow attached to their characters. I haven't made it this far, but there is a very rich and expansive endgame to reward the players that have made the grind to those high levels.

It can require less attention than a lot of other video games, so multi-tasking is possible and common.

I think for a lot of people it is the nostalgia around playing the same game when they were a kid.

What's with this resurgence in MMO's? WoW Forever, Runeterra, and now this?

Technically, you can pay your Cloudflare bill with it...

Pricing is dropping quick. Inference is so cheap, I think they are losing a lot less money selling subscriptions than you think. It might even be more expensive managing the load, than actually selling the tokens at subscription prices.

We are seeing with OpenAI, allegedly through their new pricing scheme, as intelligence and model efficiency increases they offer the same throughput while advertising 1/2 as much usage, letting Astra consume more usage, essentially only being available to those wealthy enough to afford it while still offering essentially unlimited Sol and Luna to their subscription tiers.

Also if you're cache hit rate is high enough a billion tokens tokens from Deepseek 4.1 Flash costs less than $15.

Because Anthropic releases their different model level's at very different points in time, they seem to always have one model that is by far an away the best to use for everything. Haiku 4.5 is almost a year old. Sonnet is fine, but idk if it has any real benefits over Opus. It's only been since Fable has been released that you get to choose between Fable and Opus, but not with 5.5 there is no reason to use Fable.

I feel like instead of releasing fable, they should have released it as Opus 5, then their next Opus release they would call Sonnet, and their next Sonnet release they would have called Haiku. I don't know if their pricing structure would have been able to support that, but Anthropic has always been the least competitive regarding token pricing.

Such a crude take. How deep in the sand does your head have to be to think that HN is anti-ai because of jobs? AI topics are constantly on the front page and is a frequently discussed topic here. The US has 12x as many data centers as China with only a fourth of the population. There is so much more to the American hyperscaler roll out beyond replacing coders. Not to mention, there are very few reasons to be rooting for Oracle.

Define regular folk then. I bet your definition isn't "the top 60% of a population".

There is overlap, but there is also a large portion of regular people that do not hold stock in oil companies.

I don't think the limitation is the price of inference of extraordinary private models, but rather scaling to the demand of the model and if that's the case the labs would keep the models a secret, unless they need it for marketing purposes like we saw Anthropic marking Mythos.

Thats for such a great explanation!

So to bring this back to the scale of an OpenAI experiment; they can have their inference compute, host compute (I don't imagine they are hosting the compute for 12,000 agents locally in their SF office) and the office computer to interface with the experiments with the experiment connected via VLAN thru IPMI. Then the rist is agents hacking the IPMI and gaining control of the data center? I certainly don't know data center security, but those headlines would be so much worse than anything we have seen from HuggingFace. It really seems like if these AI's want to get out, they will. I want to say sandboxing autonomous agents is a difficult problem, but seemingly it is only OpenAI having these major incidents, but I don't think it's a problem that they care to truly solve. They get marketing and they get to learn a lot more about aligning their models.

Thats a percentage, not statistics. Statistics would guide you that "regular folk" wouldn't describe the 60% most well-off, but rather the 60% in the middle. Also 'any retirement savings at all' ≠ mutual funds or stock market ownership.

When I first learned about Uber ~14 years ago they were offering free rides in my city.

My question with this point is that OpenAI’s office (or any office doing agentic research, really) is not in the same building as the DC that powers the models, so isn’t the only way to access the models over the internet?

Seeing how cheaply Xiaomi was able to train Mimo 2.6 I am starting to wonder if they are greatly over-exaggerating the real training costs to increase their valuations and investments.

Probably referencing how when Astra came out it was only 1 point ahead of 5.6 sol.

What do you mean the way rates are going? Homosexuality has existed for as long as there have been humans.

What models are good at is so subjective it isn't OpenAI's place to really say "Use Sol for X and Luna for Y". They are publishing benchmarks so you can figure out how to best utilize each model. I get that it sucks to have to do this yourself, but eventually there will probably be some type of benchmark that help with discovering a model's strengths and weaknesses

The issue that OpenAI had when they had mini and nano models is that ambiguous the differences between those and everyone just used the base model anyway. I have no idea what type of job mini can do that nano couldn't or vis-à-vis.

I do wonder if it would just be better if they were named 6-small, 6, and 6-big?

Fractal tool discovery is a fascinating idea! Have you worked on implimenting this into any agent harnesses already to any success? My first impression is allowing the agent to fork itself, not unlike launching a subagent, then returning it's response back to the main agent.

Wasn't SpaceX's insane valuation largely based off of Grok, because their rocket and satellite businesses could never be valued at over a trillion $$?

I don't see your metaphor to iphones and flip phones. This new Luna model is cheaper than deepseek 4.1 flash, except for cache reads. OpenAI having to compete with China is a much larger economic-political issue that is far larger than just our AI labs.

Ya, are LLM's not a great example of Jevon's paradox? I don't think Jevon's needs all else being equal. The paradox being that we should be able to use things less because they are more efficient, when instead they get used more.

Surely, a large part of the increase of the demand in LLMs is in their intelligence, but to hit the demand models needed to be made more efficient, and labs found that more efficient models, still demanded more usage.

Didn't SpaceX already set a precedence for garbage fire S-1s? I don't think OpenAI has to worry about that?

they said quota so i would imagine the $200 subscription. Probably through Codex or Pi coding agents.

Any business would charge more if they could. Jevon's paradox would mean that they can make more money by charging less because demand is going to keep growing.

I think their target audience are Apple Music subscribers who are going to watch the recordings of the live videos; they could probably care less who attend's their concerts. It doesn't seem like they want a music venue to support London's live music scene, but a music venue to produce content for their platform.

Frankly, that just makes it worse.

>They could have donated ten million quid to the Music Venue Trust instead.

oh, what the world could be like if music streaming platforms actually supported the industry that power them.

Since local models are largest constrained by context window you want to have a tiny system prompt. I know Hax: https://github.com/OleksandrChekhovskyi/hax was designed with local models in mind, but I haven't used it.