In Big Tech's rush to add AI to everything, they are breaking a lot of things. From The Verge, we get an update on Smart Homes. How AI broke the smart home in 2025: The arrival of generative AI assistants in our smart homes held such promise; instead, they struggle to turn on the lights.
This morning, I asked my Alexa-enabled Bosch coffee machine to make me a coffee. Instead of running my routine, it told me it couldn’t do that. Ever since I upgraded to Alexa Plus, Amazon’s generative-AI-powered voice assistant, it has failed to reliably run my coffee routine, coming up with a different excuse almost every time I ask.
It’s 2025, and AI still can’t reliably control my smart home. I’m beginning to wonder if it ever will.
The promise of AI in the Smart Home, was that it was going to make it easier to set things up, and easier for people to get things done, like run that "Make the coffee" routine. It hasn't worked out that way.
Click thru for the details of how Amazon made things worse with AI. The old systems were dependent on you using precise vocabulary, but it could actually do things.
Hat tip to Pixy Misa at Ambient Irony and Daily News Stuff 24 December 2025: Zero Shopping Days Edition
How AI broke the smart home in 2025. (The Verge) (archive site)
When all you have is an LLM, everything looks like a chat room.
Contrary to what they seem to believe in C-suites across corporate America, AI is not the answer to everything.
You know what this reminds me of?Tech Enthusiasts: Everything in my house is wired to the Internet of Things! I control it all from my smartphone! My smart-house is Bluetooth enabled and I can give it voice commands via Alexa! I love the future!
Programmers / Engineers: The most recent piece of technology I own is a printer from 2004 and I keep a loaded gun ready to shoot it if it ever makes an unexpected noise.
This also reminds me of the scene from 2001: A Space Odyssey where HAL says he can't open the pod-bay doors.
ReplyDeleteI think that AI is the biggest con job in human history. It just doesn't work well and the stories of it lying or making up stuff (the charitable version of "lying") are everywhere. The stories of how much power it's going to suck up are so absurd, I don't see how it can be done. They're talking of growing data centers that will require an additional 30 gigawatts of power by 2027, next freakin' year, and nearly 70 more gigawatts by 2030. That has never been done. I doubt a design to build a 30 Gigawatt power plant started now couldn't even get to starting building by next year. A single ChatGPT query consumes approximately 10 times more energy than a Google search does. To get an answer that you then have to spend power to determine if it lied to you or just gave you a PFA answer.
ReplyDeleteWhat AI needs is an effort to teach it an ethical system. It can be argued that's what Asimov's Three Laws of Robotics were, but that's far too primitive. AI needs something more like the 10 commandments.
I tend to think we're around the same age, and early in adulthood both of us went to computer rooms in school that were dark, full of monochrome monitors and teletype machines, and the programs we wrote were stacks of punch cards. I still remember the limerick someone had printed out and taped to the glass walls. It ended with
It never does do what I want
Only what I tell it.
BTW, I kept a copy of that "I work in IT..." meme (or whatever it's called) at some point in the past. Windows tells me it was saved it on 2/23/2020.
The AI we have today lies, because, for the most part, what we have are Large Language Models. They know what sounds good, but not what is true. ChatGPT (? - one of them) has lost the ability to count from x to y. Most of them can't do simple arithmetic, and none of them (the last I checked) can tell you if a number is prime or not.
DeletePixy Misa at Ambient Irony (check sidebar) regularly links to AI insanity, had a link not too long ago to a cost/break-even analysis that highlighted the insanity. If I find, I will ad a comment. It basically said, that these people will never recoup the costs on electricity, let alone make a profit under the current models. Snake oil. Feels a lot like Pets.com from the late 1990s.
Yeah, I started in high school with paper tape, and boxes of punched cards. Then PDP 8s and PDP 11s in college. And more. I was tempted by a PhD program in CompSci, but I got a close look at the reality of post-doc insanity through some friends, so I went into industry. It wasn't all that interesting but the money was good, and I could live pretty much in any city I wanted to.
As for Asimov's 3 laws... one of them was "Thou shalt not kill." One of the first things we are doing with AI is building military drones, whose function is to kill. Asimov told some good stories, but this makes me believe that he fundamentally didn't understand humanity.
DeleteHad to dust off my search-fu, (DuckDuckGo not Google - which stopped working a while back) to find this all the way back in Sept. Daily News Stuff 24 September 2025: Uphill Climb Edition
ReplyDelete"The pace of AI companies' spending on new datacenters is expected to require annual revenues of $2 trillion by 2030. Even the most optimistic (read: horseshit) estimates put the industry 40% short of that number. (Tom's Hardware)
"Is it still a bubble when even the people in the bubble know it's a bubble?"
Also from Ambient Irony - Daily News Stuff 21 September 2025: Inexpensive Edition
ReplyDelete"There isn't an AI bubble. There are three separate and distinct AI bubbles all happening at the same time. (Fast Company)
"As the article explains, there's an asset bubble, an infrastructure bubble, and a hype bubble, and while they all interact and reinforce each other, you need to examine them separately to try to figure out what is really going on."
Excellent additions to the source later in the day. I meant to get back to this later Saturday and don't even remember what sucked up my attention. Besides the over abundance of paperwork in the narrow corner of ham radio I'm playing in these days.
ReplyDeleteThis being a bubble or three shouldn't be a surprise: the tech sector has always thrived on hype over what's coming for as long as I can remember. Pump up the stock price, or product price, until the competition shows up and then make another bubble. We've seen Eric Schmidt (former Google CEO) buy Relativity space ostensibly to put solar powered data centers in orbit. How does he account for the other half or two-thirds of the orbit when the solar power plant is in the Earth's shadow? More techno-hype BS?