Top
Best
New

Posted by crorella 12 hours ago

GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price(openai.com)
847 points | 773 commentspage 11
prometheus1992 11 hours ago|
where's the pelican?
IshKebab 12 hours ago||
I wish they'd list the environmental cost. My employer has an unlimited AI budget so I don't care about using Astra if it's just more profit for OpenAI. I care more if it actually uses 5x more energy.
otterley 12 hours ago||
Given that the number one cost of inference is memory and compute, and the incremental cost of each is energy, cost per inference is roughly proportional to energy consumption.
lp92 12 hours ago|||
Did you care about this when it came to your other computing needs? What PC/laptop are ypu running and how efficient is that?
IshKebab 11 hours ago|||
Yes I do. I've got a spare desktop that isn't too efficient (probably ~100W idle but annoyingly I've lost my power meter) so I don't leave it on even though I would like to use it as a server.

Laptops use very minimal power - you don't need to worry about them. If they didn't their battery life would suck.

codehorses 12 hours ago|||
Likely orders of magnitude different, this is a weak whataboutism.
Bolwin 12 hours ago|||
Look at Neuralwatt. They report energy usage with every call as well as aggregate statistics.
sergiotapia 12 hours ago|||
I want to energymaxx. Every home should have a nuclear generator for free limitless clean energy. Do not energysimp, we want prosperity for all we must energymaxx and invest heavily in solar/battery/nuclear.
rs_rs_rs_rs_rs 12 hours ago||
I don't understand the point of this, why just now when it comes to llms. Why wasn't anyone enraged with the environmental costs of kids playing video games. I would not be surprised the environmental cost of that is an order of magnitude bigger than what llms have.

Edit: for context, just Steam alone has ~200million monthly active users.

JDups 12 hours ago|||
Because people find video games fun, though I suppose there's some vocal people that think of them as bad for society. In contrast the AI companies are promising a torment nexus future.

I'd be curious as to how much of internet infrastructure is dedicated to gaming though.

paulryanrogers 12 hours ago||||
Considering Nvidia's hard shift to crypto and now AI, I doubt videogames are even in the same ballpark.

How many DCs are devoted solely to gaming?

lp92 12 hours ago|||
Considering many games make use of cloud computing for online play and similar functions, they probably make up a pretty goot bit of global cloud compute capacity. Likely quite a lot less than the big AI players, but not an insignificant amount.
HelloMcFly 12 hours ago||
The energy costs of the cloud computing required for gaming are substantially less in power - not to mention overall demand - than LLMs. Come on, we're not in the same energy ballpark here.
empthought 12 hours ago|||
You're not including the physical supply chain energy consumption of distributing video game equipment in this analysis. Nobody ships LLMs to big box stores and tries to sell them to consumers.
paulryanrogers 5 hours ago||
Historically all those boxes and discs were significant, at least if you sum them all for all time.

Today physical discs are rapidly becoming a niche that newer consoles just won't have. PC games are almost never sold physically anymore.

rs_rs_rs_rs_rs 12 hours ago|||
>The energy costs of the cloud computing required for gaming are substantially less in power

Yes but it adds up when you consider that just on Steam alone there are 200 million monthly active users.

paulryanrogers 5 hours ago|||
Steam games can run on less than 4GB of RAM, and per person are likely played only an hour or so a day. AI hyper scalars use far more to serve a single customer.
rs_rs_rs_rs_rs 12 hours ago|||
>How many DCs are devoted solely to gaming?

An entire planet. Just Steam alone has one or two hundres million monthly active users.

paulryanrogers 5 hours ago||
The whole planet's DCs are not solely serving games. Apparently games take about 300 TWh whilst AI is around 500 TWh.

All of gaming is in the 300B USD range while just the CapEx of hyper scalars is already over 200B.

lbrito 12 hours ago||||
If you recall history past the last 5 minutes, you will remember that people have indeed been enraged with the environmental costs of things for a long time. Its just that AI seems to have induced a mass amnesia, and people tend to forget about what happened pre 2024.
rs_rs_rs_rs_rs 12 hours ago||
> you will remember that people have indeed been enraged with the environmental costs of things for a long time

Yeah? Show me the big movements against computer gaming.

lbrito 11 hours ago||
There are movements against consumerism and the environmental impacts of industry in general. Greenpeace is over half a century old.

The differences with AI are: 1) we are starting off (mid 2020s) from a baseline point of already being in a hopelessly shitty situation, past the 1.5C warming target; and 2) Electronics, chips, data centers etc were already a thing for a long time, but industry took _decades_ to ramp up production to pre-AI levels, and these things are used everywhere for a huge number of things. Now we're consuming electronics/data centers/water/power at an unheard-of rate, and for a single purpose (AI) with questionable benefits, besides the private interests of a handful of people.

IshKebab 11 hours ago|||
I don't think video games consume nearly as much power. A PS5's power consumption is apparently around 200W. That's not enough to run even one GPU, let alone the armada it presumably takes to run Astra.

Even then people do care about the power consumption of non-AI things. Look at the energy label on your TV or tumble drier for example.

rs_rs_rs_rs_rs 10 hours ago||
> non-AI things

But this is not that, the same gpus you play games with are used to run llms. How was energy consumation by gpu not a topic before llms?

> I don't think video games consume nearly as much power. A PS5's power consumption is apparently around 200W. That's not enough to run even one GPU, let alone the armada it presumably takes to run Astra.

Just Steam has 200 million monthly active users. Add Steam, PS, Xbox, and whole other devices having gpus and I'm pretty sure you at least 10x the energy consumption of all ai companies.

IshKebab 8 hours ago||
> But this is not that, the same gpus you play games with are used to run llms. How was energy consumation by gpu not a topic before llms?

I dunno what you're not getting but a GPU to run games is like 200-500W. A GPU cluster to run Astra is probably more like 10kW.

Also gamers tend not to spin up dozens of other machines to also game for them.

joduplessis 12 hours ago||
Sol is so good, honestly - the sweet spot for me. I've only ever found it stumbles when you don't give enough direction. But for idea execution - Sol is the GOAT.
formvoltron 12 hours ago||
didn't 6 sol just come out a couple weeks ago?
Alifatisk 12 hours ago|
Other comments have already addressed this.
lynx97 12 hours ago||
Came here only to check if the pelican spam has made it to the top again.
slopinthebag 12 hours ago||
$2/10 is pretty cheap for a frontier model...
dcchambers 12 hours ago||
GPT-6 Sol released a week ago. Shortest model life ever?
cmrdporcupine 12 hours ago|
Taking GPT-6 "Sol" outside behind the shed and giving it a merciful end is about the best outcome possible.

Huge misstep releasing it.

tandr 12 hours ago|||
The misstep was naming it Sol - it was Terra-level all along.
algoth1 12 hours ago|||
Bro, I'm a visual thinker
cmrdporcupine 12 hours ago||
Sorry, I'm GenX. Growing up they showed us "Old Yeller" in the school gym every year like that was some kind of treat.
cmrdporcupine 12 hours ago||
Aka "we made an oopsie last week and released what should have been called GPT 6 Terra with the name GPT 6 Sol"
jdw64 12 hours ago||
GPT 6.0 Sol was so terrible—I wonder if 6.1 Sol will be good?
gxs 11 hours ago|
I have used both extensively and I don’t care what the bench marks say - Claude has been way, way better and most importantly, predictable

I can handle issues much better if they are predictable even if the model makes mistakes — much more frustrating when the model is erratic

I find codex wanders off road more often and fails to see the “bigger picture” (as much as LLMs can see the bigger picture at least)

And tbh when it was first released Astral felt even worse

I’m being forced to use it right now and at the end of the day I’m making do so it’s fine, but Claude makes for a smoother experience

More comments...