AI FROM EACH END

LLM AI seems to be the one prop still keeping the USA's stock 
market, and perhaps the global economic order, up these days. At 
the same time it seems to be getting forced down everyone's throats 
with increasing intensity, although it's possible that the majority 
of people are blindly willing to accept this while ignoring its 
glaring flaws. Here are some snippets of my own experience in this 
new world of artificial thought and investment.

THE USER END

I won't linger on this because it should be easy for anyone to ask 
these AI/LLM chatbots a difficult question, get a completely wrong 
answer presented in an exceedingly misleading way, and conclude 
it's completely untrustworthy. Besides that, in outsourcing your 
own thought processes, you're surrendering yourself to the bias of 
the AI/LLM companies, who will inevitably use that to exploit you 
to their own ends. Evidently the average Joe doesn't actually care 
about these things, which has badly harmed my already critical 
perception of human society's relationship to knowledge.

However with these AI/LLM code generators I'm aware I might have 
become an inadvertent user, bogging my computer down with 
inefficient, buggy, and inexplicable AI-generated "slopcode" 
inserted into programs I've long trusted. The list of AI-infected 
software projects Ben Collver linked to in a recent phlog post 
(below) was insightful. I knew Firefox was embracing it, with 
Mozilla even labelling themselves an AI company, but I only use 
that as a begrudging last resort when Javascript is unavoidable 
anyway. Rsync was more of a surprise, and of course I already knew 
Linux was quickly heading down the AI rabbit hole but I didn't 
realise many of the BSDs were also following. I've alrady been 
thinking of ditching Linux for other reasons, but this might make 
picking alternatives even trickier.

https://codeberg.org/small-hack/open-slopware
gopher://tilde.pink/1/~bencollver/log/2026-06-01-on-avoiding-open-slopware/

THE WEB SERVER END

Since these AI/LLMs took off, reading Web access logs has become
incredibly distracting. The amount of nonsense crawler requests,
either those identifying as bots or trying every trick to avoid
bot-blockers, has gone off the charts. Here's a recent one I
experienced:

I have lots of source code archive files mirrored on one website. 
Last week I saw Claude's Web crawler had been making thousands of
requests for individual files inside them (at a rate of up to ten
requests per second), like:

http://www.example.com/gtk2/gtk+-2.24.30.tar.xz:gtk+-2.24.30/docs/reference/gtk/html/gtk2-High-level-Printing-API.html

I thought I'd accidentally enabled something to allow browsing tar
archives, until I noted all these requests received a 404 response.
The bot downloaded the archive file first, then tried to access
all the HTML and image files inside from the server. Over 11,000
requests from that IP address in 12 hours, so it isn't taking "no"
for an answer. I'm glad I wasn't actually running anything to serve 
the contents of tar files or it could have caused tons of load.

You can imagine how much of this computer power that's apparantly 
destroying the planet to fuel these AI/LLMs actually just goes into 
this sort of useless idiotic behaviour. Just to create an 
end-product that's equally idiotic, which they'll probably use to 
generate more Web crawler code to make even more pointless Web 
requests, requiring even more computer power...

THE INVESTOR END

Well here finally is a side of AI I can get behind. A few days ago 
I sold another $1000 worth of shares in the semiconductor-company 
ETF I bought $2000 worth of shares in less than 18 months ago. Just 
like six months earlier, the value of the shares had reached $3000, 
so again I returned my investment back to $2000, having now made 
back my full initial investment and retained about half the shares 
I originally purchased. That this happened again in only half the 
time is quite notable, and now I've made back my money I think I'll 
get greedy and wait for a 5x or 10x increase in value before 
selling another third. If the inevitable crash happens before that, 
there's nothing lost, and unless all the world's major 
semiconductor manufacturers go completely broke the shares will 
still be worth _something_.

Of course given the above points this is a bit ethically dubious, 
though at least I'm not pushing up the value of the AI/LLM 
companies themselves, just the companies those mugs are ordering 
stupidly large numbers of chips from. Either way, I think the blame 
really lies with AI's users. Yes, companies are ramming AI down 
everyone's throats to try and make their share prices skyrocket, 
but if all their users/customers really said "no" to it like I try 
to then it wouldn't get them anywhere.

 - The Free Thinker