More like Death Star levels of wasted resources
How much does it cost to build something the size of a moon, only for it to explode.. Twice!
Popcorn icon sorely needed
Oracle is betting that AI coding will deliver "Star Wars" levels of software engineering productivity, CEO Mike Sicilia told investors last week. With its shares sharply down over the past year, Big Red claims its engineers' productivity will skyrocket as the company mainlines the AI wonder drug it prescribes to customers. …
Customers, however, complained that the AI was "not working that well." Support note numbers had disappeared, automatic service requests could no longer be created, and users reported wasting time on broken or internal links and limited patch searches and downloads.
As a company we've never been so productive! Now pay your Oracle licence for this great service or we'll send the auditors around.
> we're able to disrupt ourselves and become so productive is because we have decades of experience
Only someone who doesn't have the slightest clue about the reality of working in a company could say such a thing.
Companies do not have experience. The employees are the sole holders of all that value. And when they leave, get fired or retire, that experience goes with them. Sure, some things will be written down, but that tends to be the "what", not how, why or when. Information that might look impressive to a bean-counter (behold! the weight of our documentation) but even if someone does find what looks like the correct action in a given situation, there is no guarantee they will know how to apply it.
And you can have another situation. Where they get in replacements for you/me who have less experience and appear to have a weird mentality. Where, when given a problem to solve at a department where I made sure, because it was important, to take good notes and document them. So the next engineer didn't spend hours looking for what I've already spent hours on and solved. So the replacement visits, can't work out the fix (its in my documentation), gets another 2 engineers to take a look, none of them work it out. Then the staff in that department say "The last engineer we had fixed that and its in his documentation", and get the reply "Oh, him, we don't look at his documentation".
Fucking idiots.
The employees are the sole holders of all that value
Up to a point. If you want to build a long term sustainable business that evolves its products to market needs and customer requirements then it is 100% the case. On the other hand, for a product with a large installed customer base, selling mainly to deep-pocketed and risk averse corporates, has high costs and time to switch, and limited competition, then sod the employees and the future, all the value is the future income stream of a largely locked in customer base, out to a say seven year time horizon. On a discounted cash flow basis, that value is then easily maximised and commercial risks reduced by sacking almost all your staff to slash costs now, and jacking up fees until customers start to leave. Then just bathe in the excess profits because the death of the product may be inevitable, but it will be slow.
Ladies and Gentlemen, I introduce to you Broadcom and VMWare!
If that was the case surely a company like Oracle would have lost a lot of value with these shenanigans. For example, if their share price hit a peak this exactly a year ago, we might expect them to have lost 60% of that by now.
*Takes a look at the graph of Oracle's share price over the last 12 months*
Well.
Once upon a time we built frameworks and code libraries to improve productivity. Solve common difficult problems once and do it right.
Now AI blasts out massive amounts of buggy regurgitated code to solve problems that it has already seen the solution for. It this point, AI is practically an extremely unreliable compiler.
I think it's even worse.
Star Wars isn't even proper SF (S=Science), it is more like Fantasy or a faiy tale decorated with space ships and robots.
This naming (much like "Stargate") shows, how much they misunderstand/misrepresent AI output as factual results and assume/pretend they can build engineering on it. Honestly, these days, I can no longer distinguish between naive mis-applications of AI and intentional marketing stunts. One half is trying to pretend AI is working great using well-controlled environments and specific settings, the other half stumbles over reality, trying to actually apply AI as advertised...
If this is true, Oracle seems to be putting all its anergy into reaching the productivity oasis shown by a the AI-mirage in the IT-desert... with predictable results.
Star Wars is indeed Space Opera, but in case anyone misreads your comment as meaning "Soace Opera was coined in order to describe Star Wars" instead of being a term that predates SW by - quite a bit - I recommend the writings of, say, E.E. "Doc" Smith.
The key to Opera being that each aria is bigger and better and louder than in the previous Act (WARNING: Spoilers ahead):
* Skylark moving from a sphere that fits in the backyard to a craft that contains the backyard - and the entire house - as the obvious solution to the question "what do our quarters look like?"
* Lensmen starting with "coruscating beams" (the coruscation is vital to its operation) to making a slightly ludicrous thermionic valve ("What are we using for a heater? Well, you are quite familiar with it already...")
* Death Star 1, half a Death Star 2, Death Planet (which can now shoot superluminal and to multiple targets at once)
Appears so. Not just software… the whole techbro loony bin..
Definitely not "Spice" — these clowns would be able to perceive the future if it were and desist from their lunacy, but then perhaps not.
I recently read that one favourite amongst that circus, ketamine actually binds opioid receptors† which I guess makes the whole special-K crew little better than your common garden variety smackhead.
"We're able to disrupt ourselves and become so productive" — "Disrupt ourselves" ? Wakey, wakey, hand off snakey ! Productive of what ?
† from https://www.nature.com/articles/s41594-026-01874-9, https://www.nature.com/articles/s41594-026-01839-y
Big Red has invested so much money on data centres that they literally can't afford for AI not to be the golden goose.
Unfortunately saying something doesn't make it true (else we'd all be in real trouble, just like Oracle). You'd think with a name like that they'd have seen this coming...
"Our periodic workforce restructurings and reorganizations can be disruptive," Oracle UNDERstated.
Disruptive is probably not the word that those being laid off would use...
To follow up my joke post with a serious one:
There's something inherently masturbatory about training a generative model on your own code, documentation, internal procedures and contracts. It's like you have resigned to the fact that you as a Chairman/CEO, and your company itself, will never produce anything genuinely innovative ever again. In that regard it's little wonder why Oracle Corp is the world's biggest AI cheerleader!
It's also a bit like a smack, crack or meth addiction. These drugs (so I am told) feel like Heaven when measured on the scale of hours, and Hell-on-Earth on the scale of weeks. As long as you can keep on increasing the dose it feels great, until you can't, because you suddenly have a choice between death by overdose, or Hell on Earth in the form of a massive crash.
So-called "AI" is like a narcotic addiction only on an organisational scale rather than a personal one. It seems great for the first few months, but then you hit its inherent limitations and need to spend double, quadruple, octouple what you budgeted for, in order to eke out any more magic juice (which it turns out, was obtained by slurping the souls of others and is inherently limited by what Humans have done before. Its own output is poisonous to itself).
To sack -all- of your staff and hand the whole company over to the bullshit generator would obviously mean death, but to back off means acknowledging its limitations and looking at the enormous mountain of technical debt (and financial debt!) you have just accumulated, having laid off most of your staff.
>> It cited risks involving review workloads, safety, security, and intellectual property.
The risk is heavily weighted towards intellectual property. Safety and security does (or should) form part of the review workloads anyway. But IP? That requires checking third parties, which is time-consuming.