The AI industry has made major promises about its tech boosting the productivity of developers, allowing them to generate copious amounts of code with simple text prompts.
But those claims appear to be massively overblown, as The Register reports, with researchers finding that productivity gains are modest at best — and at worst, that AI can actually slow down human developers.
In a new report, management consultants Bain & Company found that despite being "one of the first areas to deploy generative AI," the "savings have been unremarkable" in programming.
"Generative AI arrived on the scene with sky-high expectations, and many companies rushed into pilot projects," the report reads. "Yet the results haven't lived up to the hype."
First off, "developer adoption is low" even among the companies that rolled out AI tools, the management consultancy found.
Worse yet, while some assistants saw "ten to 15 percent productivity boosts," the savings most of the time "don't translate into positive returns."
It's yet another warning shot, highlighting concerns that even in one of the most promising areas, the AI industry is struggling to live up to its enormous hype. That's despite companies pouring untold billions of dollars into its development, with analysts openly fretting about an enormous AI bubble getting closer to popping.
Related Stories
Jacobin has an interview with Cory Doctorow about the pending implosion of the AI bubble:
Worrying about whether AI can do your job is a blind alley, Cory Doctorow argues. The real danger is AI's bubble: a speculative fantasy built on convincing bosses to replace workers with systems that can't actually do what their salesmen promise.
As artificial intelligence continues its inexorable march through human institutions, its popularity appears to be reaching an early nadir. So far, the sector's behavior almost seems tailor-made to provoke a negative response. In San Francisco, billboards and bus stop ads exhort employers to STOP HIRING HUMANS. Workers across the country brace for layoffs blamed on AI, and AI companies spend hundreds of billions of dollars on environmentally destructive data centers. You can't talk to a customer service rep anymore, only a chatbot that tells you lies. AI slop is filling up social media feeds, Spotify playlists, and even academic journals and newspapers.
One additional factor is that tech workers failed to unionize while they had the upper hand in the early decades of the WWW. Now those chickens are coming home to roost.
Previously:
(2026) Anthropic Eyes an IPO as Big Tech's AI Cash Crunch Comes for Wall Street
(2026) OpenAI Secures Record $110 Billion Funding Round Backed By Amazon, Nvidia, and SoftBank
(2025) AI Coding is Massively Overhyped, Report Finds
(2024) US Stock Plunge: Could The AI Bubble Burst?
... and many more.
(Score: 4, Funny) by Billy the Mountain on Friday October 03 2025, @04:51AM (1 child)
If you are a decent developer with prompting skills and not getting at least a 30% productivity boost using AI, you're doing it wrong. Now, there are some specialized areas where AI is lagging like FPGA coding but by and large AI lets me do in hours what would take several days.
Just follow these tricks. Save the complete context of your prompts to a text file. The moment AI gives you a bad result or seems to get tired, switch models and in the new model paste in your current context and current goal and continue from there.
That's it! Now go knock it out of the park!
(Score: 5, Insightful) by ikanreed on Friday October 03 2025, @04:57AM
This reads more of admission of you being a shitty developer with no useful skills than meaningful advice for "how to use AI well"
I don't know a single serious developer who spends 30% of their time coding.
Chasing important requirement details, and understanding what the impacts are
Identifying the proximate cause of bugs, and especially what assumptions led to them so you don't introduce new bugs fixing them
Planning a maintainable architecture
Documenting well
Once these huge, slow barriers are overcome, coding takes minutes. And AI sucks shit at dealing all of them.
(Score: 5, Insightful) by jb on Friday October 03 2025, @08:48AM (1 child)
...except of course for the nomenclature, I must take issue with that: when they say "AI" what they really mean is LLMs ... and whilst it's debatable whether LLMs form part of the field of AI at all, even if they do they most certainly do not constitute the whole field of AI, so its altogether wrong and very misleading to pretend that the two terms are synonyms (like most of the mainstream media do). Okay, that rant aside, on to my real point:
It is hard to imagine a real-world field more dominated by logic than software engineering. So of course no self-respecting SE is going to allow his pristine code to be polluted with the ramblings of a "stochastic parrot".
That does not mean than LLMs are completely useless though. There are clearly is at least one profession which LLMs could eventually replace entirely: politicians.
After all, the main optimisation goal of an LLM is to produce output that sounds plausible, even if completely devoid of merit. That's the same main optimisation goal of almost every modern politician. A perfect match!
(Score: 2) by corey on Sunday October 05 2025, @08:45PM
While you speak in jest, using AI to replace politicians is actually planned, by Silicon Valley execs. That’s according to guys in Sam Harris’s latest podcast episode. Scary, I think.
https://www.samharris.org/podcasts/making-sense-episodes/435-the-last-invention [samharris.org]
(Score: 2) by Mojibake Tengu on Friday October 03 2025, @02:03PM
Well, dear LLM, now you know where you can suck the money you need from...
The kinder you are, the easier it is for wicked people to morally coerce you.