people might joke but this is where the IT industry is headed and nobody is safe. it doesn't matter if you are good or bad, you are just another commodity. just look at how many engineering jobs were cut in the past 18 months. Oracle is now reporting it will cut 30k jobs. there is no regulation in place and by the time our politicians wake up there will be nothing left. once it was a thriving industry and now it is on the brink of collapse. the change is real whether you like it or not

@nixCraft I completely agree with this article. AI doesn't replace tech jobs as much as provides convinient excuse to bully expensive engineers into submission.

Fire many, force others to do their job or else

Previously these actions led to sharp stock price decline, now CEOs tell tales of AI replacing...

https://doctorow.medium.com/https-pluralistic-net-2025-04-25-some-animals-are-more-equal-than-others-9acd84d46742

The enshittification of tech jobs

Our last line of defense has fallen.

Medium

@iurii @nixCraft

I saw this earlier:

‪@bytevagabond.com‬

10 years of OpenAI:

2015: $0 profit
2025: Still $0 profit (but in a blazer)
Lost ~$8B in 2025.
Projected $14B loss in 2026.

Deutsche Bank estimates $143B in negative cash flow before profitability.

@krystyna @nixCraft I expect the hardware giants (Amazon, Microsoft, Google, NVidia, Apple) to either sideline or acquire software-first companies like OpenAI, Anthropic, and Cursor. They’re closing the tech gap and have the capital. When training a single LLM generation costs hundreds of millions in compute, and every model could be a flop, a waste or money

@iurii @krystyna @nixCraft

Problem is, these giants also can't make a profit with it.

Hardly anyone actually wants to pay anything for AI, and even less are willing to pay the real price.

All the expensive hardware will be obsolete in two or three years and to this day I have not seen a single actually "disruptive" and functional use case for LLMs. Small improvements here and there, sure, but providing free coffee would have produced the same productivity boost.

@AdmSnackbar @krystyna @nixCraft you may be right.

One disruptive case for LLM that I see is a shift away from LLM to less hungry models.
Speech to text model can be connected to intent recognition to summarisation, etc.
Small, specialised models, also domain specific, for DNA, semantic video compression, etc. LLMs become enabler of these specialised models.
Also LLMs could train LLM much cheaper. Basically cost efficiency gap could be closed.