<div dir="ltr"><div>I am encouraged, however, by the growing number of cities and towns that have passed bylaws against any new data centres. The electrical and water usage is outrageous.</div><div><br></div></div><br><div class="gmail_quote gmail_quote_container"><div dir="ltr" class="gmail_attr">On Wed, 27 May 2026 at 12:31, Mikalai Birukou via kwlug-disc <<a href="mailto:kwlug-disc@kwlug.org">kwlug-disc@kwlug.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">May I have a rant.<br>
<br>
- It is an insult to my intelligence to call this next token predictor <br>
an intelligence. It ain't there.<br>
<br>
- I had a tremendous effort to start to peel one person from "it thinks" <br>
into, "token predictor over patterns", and "those are processes allow <br>
you to change flow of prediction". Just peeling human into more correct <br>
mental model, just a little bit, has immediately allowed that person to <br>
use this next token predictor based on large language (implicit) model, <br>
allowed to use it efficiently and with success. Just a modicum of <br>
understanding that it ain't an intelligence, allows one to move <br>
further/faster/better.<br>
<br>
- Is better name Large Language Condensate, not LL Model ?<br>
Model implies some explicit thing. Some structure. Some actionable <br>
structure. Actionable on level of actor, not on a level of automaton, <br>
which current code is.<br>
Condensate is random. There is an overarching process of putting meat <br>
into meat-grinder, but grinding is random, like condensation.<br>
And then you can distill condensate further, or, quantize it further. <br>
Right? All of a sudden word condensate makes inroads into quantization <br>
that let's distilled essence run in smaller memory sizes. Right?<br>
<br>
Okay. This was a productive rant.<br>
Large Language Condensate: LLC.<br>
And then an accidental analogy in social, Delaware LLC's that make this <br>
economic f****y ?<br>
<br>
<br>
On 2026-05-27 01:49, Chris Frey wrote:<br>
> I hope we're not all influenced by it.<br>
><br>
> I compare the site and the "analysis" and end up wondering where all<br>
> that money is coming from? Is it all debt? Wild eyed investment?<br>
> Stock market gambling? Government loans or gifts? CIA black budgets?<br>
><br>
> Real people are losing their jobs while these companies scramble and<br>
> gamble with other people's money trying to win a race to push even<br>
> more people out of work. And while it happens, all this inflated<br>
> money is distorting the economy, making it harder on the poor.<br>
><br>
> I mean, the inflation caused by AI companies have distorted the economy<br>
> so badly that huge cloud companies like OVH have to recalculate their<br>
> costs and prices.<br>
><br>
> My gut reaction is that AI is a bubble, and it will be a glorious<br>
> day once it pops. The sooner the better. I'm not strictly against<br>
> the tech, I'm against the economic distortion that favours the<br>
> insider, the billionaire, and the politician at the expense of<br>
> everyone else.<br>
><br>
> If/When this bubble pops, how many of the CEOs are going to go broke?<br>
> We all know it's not them who will bear the risk of their actions,<br>
> nor the cost. It would be bad enough if AI was a disrupting tech<br>
> like the automobile that sustained its own growth, but it is not.<br>
> AI literally robs from everyone (in data, compute, priority, and debt)<br>
> and yields underwhelming results, partial truths, and lies.<br>
><br>
> Just my $0.02.<br>
> - Chris<br>
><br>
><br>
> On Tue, May 26, 2026 at 10:02:55PM -0400, Remi Gauvin wrote:<br>
>> I kind of hate how accurately that AI echoed my thoughts, even those I<br>
>> hadn't really solidified yet. There's just no way I could read<br>
>> something like that and not be influenced by it.<br>
>><br>
>><br>
>> On 2026-05-26 9:33 p.m., Raymond Chen wrote:<br>
>>> I sent the link to AI and it responded:<br>
>>><br>
>>> I\u2019d be cautious with <a href="http://isaiprofitable.com" rel="noreferrer" target="_blank">isaiprofitable.com</a> <<a href="http://isaiprofitable.com" rel="noreferrer" target="_blank">http://isaiprofitable.com</a>>.<br>
>>><br>
>>> From what I can tell, it\u2019s more of a provocative \u201cAI economics<br>
>>> tracker\u201d than a rigorous financial analysis platform. The core message<br>
>>> \u2014 that most AI companies are spending huge amounts while only a few<br>
>>> infrastructure players like NVIDIA are making large profits \u2014 is<br>
>>> directionally reasonable. Even mainstream coverage has echoed that<br>
>>> framing.<br>
>>><br>
>>> That said, there are a few things I\u2019d watch for:<br>
>>><br>
>>> * The site doesn\u2019t appear to provide transparent methodology,<br>
>>> audited data sources, or detailed assumptions.<br>
>>> * The numbers are likely estimates stitched together from public<br>
>>> reports, capex disclosures, and speculation.<br>
>>> * \u201cAI profitability\u201d is extremely hard to measure because:<br>
>>> o AI revenue is often bundled into cloud/business units<br>
>>> o companies capitalize infrastructure differently<br>
>>> o R&D accounting varies<br>
>>> o many firms are intentionally in growth mode, not profit mode<br>
>>><br>
>>> So the site may be useful as:<br>
>>><br>
>>> * a conversation starter<br>
>>> * a macro sentiment indicator<br>
>>> * a simplified visualization of AI spending vs returns<br>
>>><br>
>>> But I would not use it for:<br>
>>><br>
>>> * investment decisions<br>
>>> * precise financial conclusions<br>
>>> * technical market analysis<br>
>>><br>
>>> One thing the site /does/ get right conceptually: there\u2019s a real<br>
>>> distinction between:<br>
>>><br>
>>> 1. Infrastructure winners<br>
>>> Example: GPU vendors, cloud providers<br>
>>> 2. Application-layer AI companies<br>
>>> Many are still subsidizing users heavily<br>
>>><br>
>>> A good real-world analogy:<br>
>>><br>
>>> * During the gold rush, shovel sellers often made steadier money<br>
>>> than prospectors.<br>
>>> * In AI, NVIDIA selling GPUs has often been more profitable than AI<br>
>>> apps trying to monetize end users.<br>
>>><br>
>>> I also couldn\u2019t find much independent reputation history or strong<br>
>>> third-party validation around the site itself. The only broader<br>
>>> mentions I found were articles discussing its claims, not validating<br>
>>> its accuracy.<br>
>>><br>
>><br>
>> _______________________________________________<br>
>> kwlug-disc mailing list<br>
>> To unsubscribe, send an email to <a href="mailto:kwlug-disc-leave@kwlug.org" target="_blank">kwlug-disc-leave@kwlug.org</a><br>
>> with the subject "unsubscribe", or email<br>
>> <a href="mailto:kwlug-disc-owner@kwlug.org" target="_blank">kwlug-disc-owner@kwlug.org</a> to contact a human being.<br>
><br>
> _______________________________________________<br>
> kwlug-disc mailing list<br>
> To unsubscribe, send an email to <a href="mailto:kwlug-disc-leave@kwlug.org" target="_blank">kwlug-disc-leave@kwlug.org</a><br>
> with the subject "unsubscribe", or email<br>
> <a href="mailto:kwlug-disc-owner@kwlug.org" target="_blank">kwlug-disc-owner@kwlug.org</a> to contact a human being.<br>
<br>
<br>
_______________________________________________<br>
kwlug-disc mailing list<br>
To unsubscribe, send an email to <a href="mailto:kwlug-disc-leave@kwlug.org" target="_blank">kwlug-disc-leave@kwlug.org</a><br>
with the subject "unsubscribe", or email<br>
<a href="mailto:kwlug-disc-owner@kwlug.org" target="_blank">kwlug-disc-owner@kwlug.org</a> to contact a human being.<br>
</blockquote></div>