AI vs. software companies

[cdogstu99]

On 6/23/2026 at 9:56 AM, whatstheofficerproblem said:

I highly suggest trying GLM 5.2 - pretty much on par with Opus, some argue it's better even. Atlas Knowledge's stock has 15x'd YTD. The question is how far that rubber band can stretch... If frontier intelligence can increasingly be developed in the East at a fraction of the cost incurred in the West ( GLM-5.2, was trained entirely on 100,000 Ascend 910B processors with zero Nvidia silicon), then the largest capital allocators are also the ones most exposed to over-investment risk.

 

The breaking point was always likely to be when one of the major spenders concludes that shareholder returns are better served by spending slightly less. The problem is that “slightly less” is not embedded in anyone’s assumptions. The entire AI complex is priced for ever rising capex as inference demand grows.

 

This set up is becoming increasingly precarious.  The louder this message gets the more momo bottleneck bros inch toward the exit which creates neg momentum in its own right. Now we might be inclined to say who cares about GLM because it's distilled from Claude & GPT, but distillation is/was/ and always will be only one part of the equation. It's at best a bootstrapping method, the training, harness and weights over that existing distillation is still differentiated and great. If distillation is all that is required to make a model on par with the frontier labs, I don't think GLM would be the only option. But thus far, it's the only one that has come close.

 

Then there's the seemingly endless debate on open-source being good for hyperscalers. First, the premise is inconvertible: Jevons on volume, i.e. free ish weights collapse the price of intelligence + current elasticity = massively expanded token consumption. And in parallel, margins migrate somewhat away from frontier labs as "good-enough" OS eats the middle of the task distribution (albeit pricing will be pareto distributed anyway so wouldn't overstate this margin point). The labs currently act like demand aggregators, and this proliferates outward, ok, no debate thus far.

 

Second, OS weights can be served by anyone: neoclouds, sovereigns, on-prem, and maybe (probably not) the edge. So not just Hyperscalers. OS explodes TAM for everyone not just the Hyperscalers. So, for me, the shape of compute demand changes from prepaid capex commitments (from two behemoths primarily) to fragmented opex amidst a price war. So, from the Hyperscalers perspective what's changed assuming OS demand make wholes (and then some) frontier lab demand? Well, the demand risk now moves onto their balance sheet.

 

Earlier it was the VCs carrying (via financing the frontier labs) that demand risk. Now, that demand risk shifts onto hyperscaler b/sheet. Third, a question that's less clear to me: in an OS/OW inference dominated world, doesn't the cheapest integrated stack (Google) win? Doesn't the margin migrate from frontier labs to the Hyperscalers with lowest servicing cost to compete? So the OS = Hyperscaler bull case is really a cost leadership bet again. Thoughts?

i really like reading your stuff, but either Goldman ripped you off, or you ripped them off....not sure which is which but should at least disclose sources....
 

[whatstheofficerproblem]

1 minute ago, cdogstu99 said:

i really like reading your stuff, but either Goldman ripped you off, or you ripped them off....not sure which is which but should at least disclose sources....

 

this is news to me. Can you share the excerpt?

[cdogstu99]

[whatstheofficerproblem]

Holy shit. When was this out?

[cdogstu99]

i think yesterday?  and given your knowledge on the space i was having a hard time believing YOU copied them....so wow, that's pretty messed up but at the very least a nod to you my friend.

[thowed]

That's completely crazy.  Where is the lurker?!?

[whatstheofficerproblem]

17 minutes ago, cdogstu99 said:

i think yesterday?  and given your knowledge on the space i was having a hard time believing YOU copied them....so wow, that's pretty messed up but at the very least a nod to you my friend.

 

Looks like the bolded part IS from Privo. We have this debate every other day at work and keep throwing points out to one up each other, if it turns out to be productive then I combine all of it and post here. 

 

He did send.

 

Quote

Privo:

The question is how far that rubber band can stretch... If frontier intelligence can increasingly be developed in the East at a fraction of the cost incurred in the West ( GLM-5.2, was trained entirely on 100,000 Ascend 910B processors
with zero Nvidia silicon), then the largest capital allocators are also the ones most exposed to over-investment risk. Thebreaking point was always likely to be when one of the major spenders concludes that shareholder returns are better served by
spending slightly less. The problem is that “slightly less” is not embedded in anyone’s assumptions. The entire AI complex is priced
for ever rising capex as inference demand grows.

 

Didn't read the Privo part. That's what I was thinking, when did my stupid colleague become so smart. I have been tinkering with GLM in the office, so natural conclusion was that this guy saw what it's capable of and had a Eureka. LOL.

Edited Wednesday at 03:50 PM by whatstheofficerproblem

[Longnose]

LoL who used AI to write it? is AI quote you and goldman used AI to help? or are you using AI and AI is quoting goldman? or are both of you using AI and its quoting another source?

 

or is it the biggest wtf moment of all time lol. 

 

Or hell maybe hes lurking on the forum quoting your ass haha

[whatstheofficerproblem]

26 minutes ago, Longnose said:

LoL who used AI to write it? is AI quote you and goldman used AI to help? or are you using AI and AI is quoting goldman? or are both of you using AI and its quoting another source?

 

or is it the biggest wtf moment of all time lol. 

 

No jokes on me. I was simply putting fwd what I discussed at work and my colleague put this fwd. He did say it was from Privo, I didn't see Privo and posted that part here.

[cdogstu99]

what exactly is Privo ---he didn't quote Privo or anyone for that matter, so he seemed to be passing off as his own writing. 

[whatstheofficerproblem]

19 minutes ago, cdogstu99 said:

what exactly is Privo ---he didn't quote Privo or anyone for that matter, so he seemed to be passing off as his own writing. 

 

That's our nickname for Rich Privorotsky, the analyst who put out the note. .

[cdogstu99]

ahhh...ok....now i get it, thanks for being upfront about this....

[Paarslaars]

Based on MU earnings it doesn't seem like the flow down the AI valuestream is stopping anytime soon... like I hinted before, if this is a bubble, we are not in '99 yet.

Edited yesterday at 05:56 AM by Paarslaars

[Spekulatius]

LOL:

I don’t think the IPO of Anthrophic

and OpenAI will happen anytime soon.

[frommi]

54 minutes ago, Spekulatius said:

LOL:

I don’t think the IPO of Anthrophic

and OpenAI will happen anytime soon.

RIP

[rogermunibond]

Is there a strain of techno-paranoia in the West?  Does something similar exist in China?

[Spekulatius]

On 6/25/2026 at 1:56 AM, Paarslaars said:

Based on MU earnings it doesn't seem like the flow down the AI valuestream is stopping anytime soon... like I hinted before, if this is a bubble, we are not in '99 yet.

It’s not MU deciding this, it’s their customers and I think we are getting indications that they are getting antsy. I think shares prices will peak way before the memory business and the earnings do as Mr Market is good at smelling a rat in these high visibility situations,

[frommi]

6 hours ago, Spekulatius said:

LOL:

I don’t think the IPO of Anthrophic

and OpenAI will happen anytime soon.

If this ruling stays, they are both toast. Open models have catched up and theres no real reason to pay for the expensive models anymore. If they are not allowed to publish frontier models, how should they earn money? If both of them don't fund the creation of new models who should step in?
To me this is the clear bell ringing at the top for semis and the bottom for software. 

[Spekulatius]

46 minutes ago, frommi said:

If this ruling stays, they are both toast. Open models have catched up and theres no real reason to pay for the expensive models anymore. If they are not allowed to publish frontier models, how should they earn money? If both of them don't fund the creation of new models who should step in?
To me this is the clear bell ringing at the top for semis and the bottom for software. 

Agreed. It is not inconceivable that the AI bubble will be deflated by overzealous government regulation.

 

Anthrophic has been blacklisted by the government  and their best product is export contoured which being quite a bit of bureaucracy with it an limits sales into man markets. Now OpenAI looks like it’s in the same doghouse. I don’t see how this business can be IPO‘d with these risk factors.

Edited 21 minutes ago by Spekulatius