Most of the time the companies say the layoffs are because AI is doing people jobs, but the reality is that making things that use AI is just a lot more expensive. And enabling employees to use AI more than the free tier is really expensive too. So, they need to cut cost elsewhere to balance it out.
Another company slowly shutting themselves down.
Oracle putting itself out of misery would be the most philanthropic act of an organization in history.
Thou shalt not make a machine in the likeness of the human mind.
Clay Magouyrk (CEO):
Larry, in order to pay for these datacenters we might need to cut up to 20,000 jobs.Larry Ellison:
Cut 30,000.Clay Magouyrk:
But Larry, 20,000 would be more than…Larry Ellison:
As long as we’re going with Gary Oldman:
“Fire everyone.”
“What do you mean everyone?”
“EVERYONE!!!”
I’m going to watch this movie, right now, but the RiffTrax version
A really good reference
“Zorg, you’re a monster.”
“…I know.”
Tens of thousands of employees and not one of them knows how to make a good error message
I was once told by Oracle engineers that it would take 18-24 months to add a drop down with auto complete to their ticketing system.
They don’t have good engineers because Oracle is a law firm pretending to be a tech company.
They don’t have good engineers because Oracle is a law firm pretending to be a tech company.
We call this the IBM business plan.
Engineer? Definitely not near Oracle.
Not sure the engineers are entirely at fault for how long things take there. I’m guessing they must have some insane review and release processes with significant bottlenecks and it’s all because of their structure
I recall reading an ex oracle engineer say that the code base is basically spaghetti. Not only that, but you basically have to be lucky to get your pr in, as due to said spaghetti, there is a high chance that it will be broken by the pr merged before yours.
Which code base? All of them? They have a lot of products.
But of course it could also be all of them
Not sure if this is it, but this sounds similar to what I remember: https://news.ycombinator.com/item?id=18442941
I worked with a guy that used to work at Oracle and he pretty much said the same thing. Essentially if you sneezed anywhere within a 10km radius of Oracles code base (and it didn’t matter which product) you ran the risk of it all crashing down. dudes spent more time talking about how to theoretically fix something as opposed to actually fixing it.
These guys were blown away when I showed them the same feature in a competitor’s product.
The company I worked for had just been bought by Oracle and we were under a directive to switch to all Oracle software, so I guess they weren’t motivated because there wasn’t extra money in it.
Okay, yikes. I didn’t think it could be THAT bad, but I guess Oracle never fails to surprise.
The worst part is that it was all down to our existing system being owned by Salesforce, and Larry Ellison hating the CEO of Salesforce.
We could still use Amazon and Microsoft products but nothing from Saleforce because Uncle Larry was butthurt.
Maybe they should consider spinning off Paramount or not buying WB instead firing people.
That’s the son’s company (Skydance). Maybe Larry could just ask him for a loan I guess.
What, the only parts of the company that actually produce anything of value?
Bbbbut, they also sell a payroll & HR system that is somehow even worse than SAP’s! The main goal of all their Fusion Cloud junk seems to be producing consulting fees, because it’s impossible to get anything to work without them.
Oracle Employment Numbers Trend (Approximate) 2025: 162,000 2024: 159,000 2023: 164,000 2022: 143,000 2021: 132,000 2020: 135,000 2015: 132,000 2011: 108,000
What do AI and toilet paper have in common? I wipe my ass with both.
Fuck AI.
And enabling employees to use AI more than the free tier is really expensive too.
The cost is upfront for training the LLM, then they have to sell the end product to us (to recuperate costs, start training the next 500B model, etc.), so giving their employees access to a higher tier of LLM is relatively minimal cost, inference isn’t exactly cheap, but I think they could afford to give their employees access to higher models.
But, this is Oracle I guess.
