Wonder what this will do for the “human intelligence” of current young generations in the long run…
3 Likes
In my son’s defense 
he used ChatGPT mainly to nicely write up a report. The actual work he had to do by himself as it was centered about a proprietary service (which exists only at that company).
But your more general point is correct, of course. And I’m pretty sure my son would have used ChatGPT for the actual project work if ChatGPT was knowledgable in the area …
Poll time …
- I use AI (ChatGPT, Claude, etc).
- I pay for a subscription myself.
- My company pays for a subscription.
- Nether I nor my company pays for a subscription.
- I don’t even use AI directly (maybe Google or other services give me AI powered answers).
0
voters
I don’t use it, but I already tried though
1 Like
Yeah. To get a sense of the scale here, both Microsoft and Google have mentioned their token volumes recently.
Microsoft served 20T tokens in Q1 2024, 100T tokens in Q1 2025, 50T tokens in March 2025. (Source).
Google did 10T tokens in April 2024, 480T in April 2025 (Source), and 980T in June 2025 (Source).
So e.g. for Google’s models, it looks like OpenRouter is responsible for maybe 1/300th of the volume.
3 Likes
R1 was a frontier model when released this January. GPT-OSS-120b is only a couple of weeks old and I believe is on the cost/capability pareto frontier on some classes of benchmarks. (Though to be fair, GPT-OSS was trained almost entirely on synthetic data and heavily benchmaxxed for math and reasoning, so it seems that not a lot of people are finding it to be that good at real world tasks.)
It seems that at this point nobody except xAI is trying to brute-force the problem purely by adding more scale to the pretraining or by just adding more scale to the reinforcement learning. I’m sure they’re all spending all the compute they can spare on training, and will definitely be spending a lot more on it in 2025 than 2024, but it’s being spent more smartly than “spend 10x more to get 10% improvement”.
That seems incredibly implausible, given they were as well connected with SV VCs as is humanly possible, and had $1B in initial funding. They would have no need to find retail investors through some tiny Swiss private market. Crunchbase also knows nothing about these multiple seed rounds.
This is the first I’d heard of Moonshot, but from a little reading it seems that these were not deals for actual shares but at best some kinds of derivatives with non-credible counterparties. But more likely a ponzi scheme.
2 Likes
I didn’t say it was publicly available.
Moonshot had a tracker certificate (or at least on paper they had) for a lot of PE oppties (SpaceX, anthropic, openai, etc) on obscenely high fees. Could have been a win-win.
Maybe it was a scam, we will know in some months when the bankruptcy agency in Zug is done.
2 Likes
You are missing the: I bought a bunch of GPUs to host my own LLMs and am burning through electricity like a crypto farm that’s become self-aware.
A couple of years ago, I had no Nvidia GPUs and was firmly in the AMD camp. I now have 12 Nvidia GPUs. I need to go to GPUs Anonymous.
5 Likes
Does your company training their own LLM count as a subscription? 
1 Like
Technically, that’s a negative subscription (as they are paying you).
Please refrain from answering the survey, sir.
I think „are we in an AI bubble or AI bonanza „ deserve a separate thread
i feel it would be both 
AI will change our lives but we are also over estimating its impact and utility. I am not working in IT companies so I don’t have inside view. But I work in industry and I can say for now the use cases are not at scale except summarisation , content creation and task management.
And to be honest looking at all the AI slop, I hope the content creation via humans still have a place or else we will need to read through 1000 pages of garbage to get to 1 page of real information.
I was talking to someone from Google and they told me that when google says X% of the code is written by AI, it is a bit of exaggeration because it includes certain things like auto complete etc which was already in use for some time.
1 Like
Most of what we’re building out at this point is the inference […] We’re profitable on inference. If we didn’t pay for training, we’d be a very profitable company
A little bit like a car manufacturer saying
“We are profitable on sales and service of our cars.
If we didn’t pay for developing and manufacturing them, we’d be a very profitable company.” 
7 Likes
A more apt comparison would be to say “We’re profitable manufacturing and selling our current cars. If we didn’t develop new models we’d be profitable.”
5 Likes
“But also, if we did not develop new models, in one year we‘d be behind and our competition would take over.”
so yea…
1 Like
Looks like the US leader’s quest for „utmost loyalty“ doesn’t stop at government employees. Outreach to C-Suites: your well-behavior (as defined by us) will matter:
The West Wing has created a scorecard that rates 553 companies and trade associations on how hard they worked to support and promote President Trump’s “One Big Beautiful Bill,” a senior White House official tells Axios.
Why it matters: Trump works transactionally, and companies have rushed to pay demonstrative homage. Now, senior aides will have data to consult when considering corporate requests.
1 Like
Congrats! How is it possible that VWRL says +0.02% YTD in CHF when it’s ATH in CHF was ~130/share and today it’s at ~125.5?
The ATH probably wasn’t on 1.1.2025
1 Like
I’m clearly a dunce, yes you’re right, mental accounting at its best.
Heyyyy I just proved to myself I don’t look at the market all the time!
2 Likes