As IBM Pushes For More Automation, Its AI Simply Not Up To the Job of Replacing Staff (theregister.com) 38
An anonymous reader shares a report: IBM's plan to replace thousands of roles with AI presently looks more like outsourcing jobs to India, at the expense of organizational competency. That view of Big Blue was offered to The Register after our report on the IT giant's latest layoffs, which resonated so strongly with several IBM employees that they contacted The Register with thoughts on the job cuts. Our sources have asked not to be identified to protect their ongoing relationships with Big Blue. Suffice to say they were or are employed as senior technologists in business units that span multiple locations and were privy to company communications: These are not views from the narrow entrance to a single cubicle. We're going to refer to three by the pseudonyms Alex, Blake, and Casey.
"I always make this joke about IBM," said Alex. "It is: 'IBM doesn't want people to work for them.' Every six months or so they are doing rounds of [Resource Actions -- IBM-speak for layoffs] or forcing folks into impossible moves, which result in separation." That's consistent with CEO Arvind Krishna's commitment last year to replace around 7,800 jobs with AI. But our sources say Krishna's plan is on shaky ground: IBM's AI isn't up to the job of replacing people, and some of the people who could fix that have been let go. Alex observed that over the past four years, IBM management has constantly pushed for automation and the use of AI. "With AI tools writing that code for us ... why pay for senior-level staff when you can promote a youngster who doesn't really know any better at a much lower price?" he said. "Plus, once you have a seasoned programmer write code that is by law the company's IP and it is fed into an AI library, it basically learns it and the author is no longer needed." But our sources tell us that scenario has yet to be realized inside IBM.
"I always make this joke about IBM," said Alex. "It is: 'IBM doesn't want people to work for them.' Every six months or so they are doing rounds of [Resource Actions -- IBM-speak for layoffs] or forcing folks into impossible moves, which result in separation." That's consistent with CEO Arvind Krishna's commitment last year to replace around 7,800 jobs with AI. But our sources say Krishna's plan is on shaky ground: IBM's AI isn't up to the job of replacing people, and some of the people who could fix that have been let go. Alex observed that over the past four years, IBM management has constantly pushed for automation and the use of AI. "With AI tools writing that code for us ... why pay for senior-level staff when you can promote a youngster who doesn't really know any better at a much lower price?" he said. "Plus, once you have a seasoned programmer write code that is by law the company's IP and it is fed into an AI library, it basically learns it and the author is no longer needed." But our sources tell us that scenario has yet to be realized inside IBM.
That's concerning (Score:2)
I hope they aren't planning to have AI write code for their operating systems.
Re: (Score:3, Informative)
Everything will be sold. They actually sold off z/VSE to a third party company. I expect they'll eventually do the same with the mainframe division but still keep churning out patents because that's where the money's at, still.
IBM's been trying to get rid of VSE for 40 years (Score:2)
IBM wanted to get rid of DOS/VSE in the '80s and move customers using it to MVS (the predecessor of z/OS) so they'd only have to support one mainframe OS. Customers apparently pushed back, and IBM kept maintaining VSE and its successors. They finally managed to sell it off to 21st Century Software Inc in 2021. I don't think they want to get rid of z/OS.
merger please (Score:2)
The old tech companies IBM, Oracle, HP should merge then combine all the profitable products and businesses into a new company and spin off all of the barely profitable, almost profitable and money losing operations into a new company. This includes splitting by product and by geographic region.
I think it's been done before in the oil industry.
Re: (Score:2)
IBM doesn't write operating systems anymore, they write support contracts for PHB's.
Maybe they can train their salesbot on Dilbert.
PHBs outsource people who actually got work done (Score:5, Interesting)
and then replace them with people who just behave like they're getting work done. It's the ultimate evolution into a cargo cult from an organization that used to practice engineering. The problem? the people doing the measuring are morons.
If $1000 bolt and $10 bolt seem to be doing the same job, replace the $1000 bolt with the $10 bolt, and give yourself a bonus and a raise with the cost savings. The problem then is when your $1M machine starts breaking down because the $10 bolts are shearing off at inconvenient time, and the service contracts you had your customer buy (which made you money when they didn't have to call you for servicing) now cost you all of your profit and then some because you now need to dispatch technicians with backordered parts.
Replacing senior engineers with AI + junior coders is a shit move. When the AI breaks down, are the junior coders going to know how to fix it? Or are you going to have to outsource what you are claiming as your core competency now too? As a customer why don't I just hire the people you are going to for help, since you're just a useless collection of 1st and 2nd tier tech support, a bunch of underpaid technicians, and a useless mass of middle and upper management overhead?
I hate to say it, but at this point Oracle is starting to sound more competent than IBM...
Re: PHBs outsource people who actually got work do (Score:1, Troll)
Re: (Score:2)
IBM is pretty much busy shoveling its own grave at this time. They have been going in that direction for a long time, but now they have entered the hot phase. Competent engineers are critical to keeping any tech enterprise going. Forget that and there is no future. Obviously, their lawyers can do a lot of damage for decades to come. Best thing for society to do now would be to nuke them from orbit. They have no usefulness left.
Re: (Score:1)
All too often we think "senior" as old and "junior" as young. There are some young bucks out there that are real problem solvers, and some old codgers that are just seat warmers.
But management can't tell the difference between real problem solvers and seat warmers, or AI.
Re: (Score:2)
If $1000 bolt and $10 bolt seem to be doing the same job, replace the $1000 bolt with the $10 bolt, and give yourself a bonus and a raise with the cost savings. The problem then is when your $1M machine starts breaking down because the $10 bolts are shearing off at inconvenient time, and the service contracts you had your customer buy (which made you money when they didn't have to call you for servicing) now cost you all of your profit and then some because you now need to dispatch technicians with backordered parts.
The problem is PHBs are smart, the PHB that got the bonus for using $10 bolt would be long gone with the money by the time the machines break down. It is a different PHB that faced the problem, who will do something similar, pocket the bonus and leaving problems behind for the next guy.
There is no downside for PHBs that do this (pocket money while leaving problems for the next guy), as long as they move to the next job fast enough.
Maybe? (Score:2)
Maybe IBM needs more data centers [slashdot.org].
For the last 20 years (Score:5, Insightful)
Re: (Score:2)
Back in the day, I worked for an Austin dot.com that IBM made a huge offer for 50%. Our CEO did not take it. Hubris.
In any case, at that time I worked out the average $ of IP per IBM employee and it was on the order of $400K. Now I suspect it is much lower, I think Watson was the beginning of the end. Can't think of any thing of intellectual value out of IBM.
Re: For the last 20 years (Score:3)
Re: (Score:2)
Indeed. And now they do not have people left that do useful stuff. Hence this desperate move to "AI". For obvious reasons, that will not work and cannot work.
Re: (Score:1)
Enshitify RedHat.
Re: (Score:1)
What do they actually do now?
I believe they accumulate patents [greyb.com]
Computer Science (Score:5, Insightful)
The absolute state of compsci, compeng and related fields has got to be unique in its idiotic and schizophrenic capacity for bludgeoning people with facts and math on the one hand, yet simultaneously engaging in levels of magical thinking that would make a scientology acolyte blush.
I am a big fan of Lisp, Scheme and functional languages in general, read up on Symbolics Lisp Machines or Knight machines if you ever want to get depressed about the direction the industry COULD have taken.
It always makes me sad that the AI winter more or less put the final nail in the coffins of these cool ideas, but now I am eagerly looking forward for the next AI winter.
At least this time it will only he LLMs that get decimated and not legitimate AI research.
Re:Computer Science (Score:4, Informative)
The absolute state of compsci, compeng and related fields has got to be unique in its idiotic and schizophrenic capacity for bludgeoning people with facts and math on the one hand, yet simultaneously engaging in levels of magical thinking that would make a scientology acolyte blush.
Having observed the field and having been part of it for almost 40 years now, I completely agree. There are too many people in IT and CS that do not understand context, cannot see surrounding factors and, essentially, cannot see reality, The number of highly intelligent, big-ego idiots you find in IT and CS is staggering. For a simple test, try to explain risk management to an IT or CS person. My experience is that about 20% get it, which is pretty much the number of people that can understand and be convinced by rational argument in the general population. The rest just fails. They are so hypnotized by their ability to handle complexity in a narrow field that they simply cannot see the rest of the picture. It is really pathetic.
Of course, it helps that I am mostly in IT Security, where magical thinking has a tendency to backfire relatively fast. Not that the average IT security "expert" is any good either. They tend even more to magical thinking and often have even less skills.
I am a big fan of Lisp, Scheme and functional languages in general, read up on Symbolics Lisp Machines or Knight machines if you ever want to get depressed about the direction the industry COULD have taken.
It always makes me sad that the AI winter more or less put the final nail in the coffins of these cool ideas, but now I am eagerly looking forward for the next AI winter.
At least this time it will only he LLMs that get decimated and not legitimate AI research.
Indeed. That said, functional concepts have made it into some modern languages like Python or Rust. Obviously far too late, but at least the time where the average CS grad has not even seen anything functional is slowly coming to an end. It does help that OO has mostly failed to deliver on its promises. Incidentally, look at what otherwise smart people expected from OO for a ton of examples of magical thinking. Sure, it has its place, but it is not the overriding one true paradigm at all that it was regarded (and sometimes still is regarded) as by many people.
An overloaded acronym? (Score:3, Funny)
As in "IBM's plan to replace thousands of roles with assorted Indians"?
Re: (Score:1)
Or "Actual Indians"
Re: (Score:2)
IBM started standing for Indian Business Machines a long time ago. Most of their US presence shifted overseas over a decade ago.
The only difference now is that we just came off of a work from home revolution. Upper management is now more convinced than ever that if their remaining software engineers can be productive working 50 miles away from an office, they should somehow be equally productive if their job was replaced by an Indian subcontractor working 10,000 miles away from the office. Sure, the new fol
International Business Machines (Score:3)
IBM is not people.
Soylent Green is people
Interesting read (Score:2)
This post me wondering what IBM actually makes anymore, or is it just all services?
Sad, that for a company so visible and important as they once were, that even if they are still busy and profitable, that they are largely irrelevant or invisible to the average non-corporate person.
So, I searched "does ibm manufacture anything anymore".
A link pointed to a Quora thread. Many or most responses are by IBM'ers, many waving the Big Blue flag.
But, one I found particularly interesting. Follow this link :
https://w [quora.com]
Re: (Score:3)
Many years ago I was told to look into the Domino server and mail product line of IBM, as a possible alternative to Exchange. Then I found out that they charged 800 USD per mail client connecting to the Domino server. Way back then it was a pretty poor mail client functionality-wise and felt shame if I needed to subject anyone to its dreadful graphical interface.
IBM I considered to be a very deluded company for trying to sell that kind of monstrosity to any business for revolting prices. Guess there have be
Such a surprise (Score:3)
And once again, demented "management" could not separate solid vendor assurances from vendor hallucinations.
Although, having met some IBM consultants, I am surprised that AI is even dumber than they are...
Outsourcing is one way to prevent the AI lawsuits. (Score:3)
The wave of none technical executives that bet big on AI is cresting for the moment.
AI is simply not ready to take on most roles. Augment roles sure. But replace not really.
Fundamentally AI lacks precision, AI best achievement is mimicry. For some things this is good enough. But for tasks such as coding it is a long way off. Coding is a task that requires the creator to understand the context of the problem. Even when requirements are stated with high precision what is not stated in most requirements is the greater context of the problem/domain/business. AI with all it's information gathering capabilities will still lack this context for years to come.
AI's lack of precision is already starting to impact contracts and obligations of organisations. Since AI is a mimic at this point it excels at producing content that appears to be correct on the surface. Collectively at the moment these are largely called AI hallucinations. And this has been getting companies in a lot of hot water lately. With AI promising actions to customers that simply are not possible or even on offer. Like reduced prices, dates, agreements etc.
In code this lack of precision results in bugs, missing features, security lapses, and generally loose code that is wide open to a range of issues. I would expect to see laws enacted soon that ban AI from contributing code to transportation, and direct medical treatment fields ( EG control the radiation beam in Chemo therapy. As apposed to identify a tumour in an image. ). Well at least I hope to see these laws come about.
Now all this is said based on the current state of AI. Will this change? Of course it will. In domains that are more self contained with fewer external dependencies of knowledge then I see AI making fast progress. Example: Design an optimal heating and cooling system for large multi purpose complex. This has a relatively finite set of requirements with low external dependencies.
So there will be a second wave of AI proliferation. It's not too far off. My bet is it will be in very targeted domains. AI started in targeted domains. Then there was a push for more general capabilities. This is the current wave. Now that massive funding and huge datasets are available going back to the targeted domains will yield better results. This will be the next wave. For the wave to be successful the knowledge of what AI can do and it's maturity will need to be infused in the leadership of organisation. Right now AI is purely a speculative gamble. And some companies like IBM gambled big.
As predicted by Rod Sterling (Score:1)
There are many bromides applicable here: 'too much of a good thing', 'tiger by the tail', 'as you sow so shall you reap'. The point is that, too often, Man becomes clever instead of becoming wise; he becomes inventive and not thoughtful; and sometimes, as in the case of Mr. Whipple, he can create himself right out of existence. As in tonight's tale of oddness and obsolescence, in the Twilight Zone.
- "The Brain Center at Whipple's" [wikipedia.org], The Twilight Zone, 1964
IBM is largely outsourced (Score:3)
I worked at IBM for quite a stretch of years beginning in the 90's, and a bad experience it was for most of those years. There was always a persistent strong push to outsource overseas. Frequent rounds of layoffs in the US that often seemed to be mysteriously targeting the older and more experienced people. And they were better paid too of course, which is probably why.
Eventually the company transformed from something based in the USA to a different thing that was about 80% residing in "low cost geographies" to use their term. There is just a relatively small shim of native US people, enough to be the face of the company there. Meet with the clients and be the interface with the vast bulk of the company which is mainly in India and China..
Now an even more attractive concept for them is to just eliminate the people insomuch as this is possible. Tech talent was always disposable for them, now the idea is that maybe they can do without most of it completely. Seems be have been a little premature but I think they will keep pushing for it.
IBM=HAL (Score:2)
Indian CEO (Score:2)
He's turning the company into an Indian business- falling apart at the seams while pretending to get work done.
Surprise, m*********er! (Score:2)
What a surprise that a bunch of IFs, no matter how extensive, can't replace competent developers. I'm shocked, mate, shocked..
The CEO is Arvind Krishna? (Score:2)
Perhaps increasing employment in India was his real goal all along...
IBM is an Indian company (Score:2)
The direct competitors of IBM are TCS and Infosys. It doesn't produce anything, but instead has a zillion consultants running around the Fortune 500 companies. As most of its technicians and managemeht are now Indian, IBM should be considered the Indian Business Machines company.
'IBM doesn't want people to work for them.' (Score:2)