I was just informed from my boss and HR that my entire profession is being automated away.

[o7___o7]

145. lol, no

I submit to you that the entire AI grift is running on para-religious mania and fumes. Ed Zitron lays out the case perfectly here:

https://www.wheresyoured.at/subprimeai/

In summary, OpenAI et al believe that LLM (that is, large language model) performance scales linearly with training data and processing power,. Therefore, all that they need to do is continue building data centers and scraping all possible text sources to improve their AI models to a "god-like" capacity, after which any bills coming due (such as loans or violent social upheaval or climate change) will be rendered moot by the apotheosis of our new robot overlords.

This is dumb:

1) Turns out LLMs don't scale like that, and even if they did, the industry has burned through almost all of the human-generated texts in the world.
2) As an example of the hole that is being dug by the industry, OpenAI, by far the dominant firm in "AI", is spending money 5x faster than it can make it, and is making up the deficit with fund raising.
3) OpenAI raised 5 billion this year, giving it a nine-figure market cap.
4) OpenAI made 1 billion this year.
5) OpenAI loses money on every API request.
6) Large institutional investors such as Goldman Sachs are openly skeptical.
7) OpenAI has released several not-quite-GPT5's with huge fanfare that have all failed to improve over their predecessors.
8) OpenAI has projected that it must spend trillions (With a t) of dollars to build the physical infrastructure to power the next generation of AI models.
9) But LLMs just don't work. No matter how much money is poured into hardware and training, these LLM architectures will never be capable of more than being "spicy autocorrect."

In the next few months, barring a huge infusion of capital from a nation-state level angel investor, these companies will start falling like dominoes. The only question is how much human carnage credulous idiots like this guy's boss will cause in the meantime by preparing for a magic robot future that will never come, because people need to eat in the meantime.

For a more rigorous, edifying, deeper dive, I highly recommend "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?" by Drs. Bender, Gebru, Mitchell, et al.

https://dl.acm.org/doi/10.1145/3442188.3445922