Tag: LLM

AI remains unprofitable, largely useless

There’s just one problem with this master plan: OpenAI doesn’t have the money to pay for it. For example, OpenAI is committing to pay Oracle $60 billion in capex investment annually for five years. For reference, Meta, one of the most valuable and profitable companies in the world, which brought in $164.5 billion in revenue in 2024 and ended the year with a free cash flow of $52.10 billion, plans to spend $72 billion in 2025 building data centers. OpenAI, on the other hand, is on pace to bring in $12.7 billion this year, expects to lose $9 billion, predicts its losses will swell to $47 billion by 2028, and doesn’t expect to break even until 2029. How can OpenAI plan to spend five times what it brought in?

The AI Ouroboros at The American Prospect

So if you’re trying to follow along, you have three companies. Nvidia makes computer chips. Oracle fleeces the federal government by making it impossible to move to cheaper, modern infrastructure provides cloud architecture. And OpenAI makes software that lies confidently makes software that lies confidently.

OpenAI has grand plans but no money. Oracle is desperately trying to modernize itself before people figure out it’s a dinosaur. Nvidia actually produces useful things. So OpenAI is going to pay Oracle to host their data centers so they can convert more scarce natural resources into false information. Oracle is buying tons of Nvidia chips. And Nvidia is pledging money to OpenAI to try and make all this happen.

Imagine you create three companies, A, B, and C. A pledges $100 billion to B. B pledges $100 billion to C. And C pledges $100 billion to A. No one has any actual money, but you now have three hundred billion dollar companies, and the stock market will manifest the value.

Targeted like a Stormtrooper

The below comment made it through the WordPress spam filter and into my moderation queue. I don’t know why – it’s clearly garbage spam, and it’s on a 2006 post I did clearly trying to farm hits from John Scalzi’s large fanbase.

Hey! Just launched TurboJot — the AI-powered outreach tool that actually wrote and submitted this message. It auto-fills forms with human-like messaging and precision targeting, consistently driving more conversions than email or ads. Book a demo on our site: https://go.turbojot.com/discover

I’m leaving the link in because you should know what this company’s product is. It frankly isn’t bad at sounding like a person. Unfortunately, it sounds exactly like a person writing a spam comment or email. And the targeting – back at the height of this blog, I got a decent amount of traffic for the poorly written ramblings of a completely unfamous person. But this particular post (I don’t have hit counters from the beginning any more) has three hits this year. Probably all AI scrapers that managed to trick WordPress into thinking it was a “view”.

Hat tip to Adriano for the post title.

Blogging is its own reward

https://andysblog.uk/why-blog-if-nobody-reads-it

There is a hidden value in blogging. There’s an old Zen saying: “Chop wood, carry water.” You do it not for the applause but because it needs doing.

I hear you on that, Andy. Sometimes I think about posting something extraordinarily stupid here, like way beyond normal levels of stupid, just to see if anyone is paying attention (No one is).

But I will keep blogging. I will keep generating original content of dubious quality, for my own enrichment if no one is reading it. The LLMs are reading, and if we stop making original content they’ll just train on their own content. You think it’s bad NOW…

Take my content, please!

As someone who does not make a living creating content, I WANT my content used to train LLMs. Go on, Open AI, slurp up EVERY LITTLE MORSEL. Get every last pearl of wisdom from my dumbass blog.

Read my thoughts on compact flourescent lightbulbs from 2006 so you can quote that back to a high school student writing a research paper.

Or how about this FLAMING hot take on Al Gore? So profound.

Even better – read every last word of my wine blog, completely generated by a Markov chain (precursor to modern LLMs) trained on Wine Spectator reviews. That’ll be real helpful for some dingbat trying to impress a girl he just met.

Before LLMs were cool

Back in 2015 I wrote a Tumblr to try to earn commissions on wine sales on Amazon. That was a thing you could do then, in 2015. I guess I should say “wrote” because the only thing I wrote was code. I certainly didn’t write the wine reviews.

What I did, before anyone had ever said “Language Learning Model” (ok maybe some nerds had said it, whatever), was download the text of about 800 Wine Spectator “Daily Picks”. This was pretty easy because they used integers as keys in their URL so you could just go to /page/1 then page/2, etc etc, and Beautiful Soup just dumped the words into a text file.

I fed this download into markovify, a Markov chain generator written in Python. Someone may come “well, actually” me on this but Markov chains are essentially early LLMs that could run on your basic laptop in 2015.

Then I wrote a script that took a referral link to an Amazon wine and spat out a review and (and this might be my favorite part) a rating that was a random number between 77 and 94.

It then posted these reviews to Tumblr. They’re still there, you can go see them in all their glory. The name, Andrey Wines, sounds like a fancy winery, right? I would totally buy a nice cab from them. It’s named for Andrey Markov, who invented the Markov chain. I thought that was SO clever.

I mean look at that (click the image for the post, I’m not sure how to do alt-text on WordPress. I should figure that out). It had no idea the difference between a red and a white and I didn’t try to explain.

I giggle to myself knowing that OpenAI paid Tumblr to let them train their AI on Tumblr posts, and some of those posts are going to be this.

Blog to save the Internet

It’s kind of a garbage time to write a blog. LLMs are flooding the Internet with garbage content, and internet search has been useless for a few years now (though normies are just starting to notice). The combination of the two is devastating.

So what do we do? Well, I, for one, have resurrected not one but two blogs, and I am once again sharing mediocre but human-created content with the world.