I’m hanging on to my 2003 BMW sedan with a maniacal grip — as every car I’ve considered buying today would jam me into the auto industry’s version of IofT—and never mind all of my driving habits being sold to the insurance companies. Thank god I live on an island, I’ve only 70 thousand miles logged on a 20+ year car. They’ll bury me in it.
The central mistake is the propeller hats keep putting plumbing as the main feature. IoT works fine if automation and networking is wanted,just as blockchain is a clever key exchange mechanism, but these are not user facing. These should make the system more efficient but they keep ramming them into everything and telling us it’s better than the invention of refrigeration (which it’s not). AI should be a feature not the whole show. But alas, propeller hats.
Glad to deliver. FWIW,my full stance. If we spoke of GEN AI as not some new fangled, confusing tech black box but as an automation tool then I believe the average person would be better equipped to have a nuanced discussion. AI automates tasks that would take time - see the kerfuffle about improving accents in the movie the Brutalist. A human could manually enhance an accent (either by the actor getting better 🤣🤣 or by manually tweaking knobs on a device) but it currently takes too much time. There is a quality/character difference between automated output and handmade, let’s have that discussion and let people put their values on the table.
I still remember reading the book Abundance ten years ago that told us of all these promising IoT things that would be happening in the next 5 years. I’ll never forget every soda cup in the world was supposed to have an IP address so if you didn’t pay for soda the machine just wouldn’t turn on when you put an off network cup beneath it. I think about it all the time and how believable it was then and how insane it sounds now.
Really great, succinct. This is the kind of sharp and short anaylsis you can send to friends who are not yet entertaining what a disaster it could all be.. LLTM
Another flaw with these high-tech Internet of Things objects is that when the company stops supporting the apps or networks they depend on, the product becomes much less useful, or even useless. And your local appliance repairman is not going to fix the wifi on your fridge. So even though you've purchased a "thing" you don't fully own it. More often than not, you've just subscribed to it.
What the IoT has in common with AI (and crypto, too) is that it's often a form of overengineering. A complex solution applied to simple problems. A sledge hammer used on a finishing nail. Too many AI use cases being proposed are for things that can be done almost as well but much more efficiently without AI. Ex: Google search used to be great, then they slowly made it terrible, then they applied AI to "fix" it, and now it's even more terrible, but also helps to boil the oceans.
AI makes good sense for some uses (let it analyze a billion x-rays and learn to detect tumors, sure!) but for many things it does not make sense. Yet we are having it forced upon us anyway, because someone needs to profit from us at our expense..
I read somewhere they are doing this on purpose to stop people buying appliances for life. They basically want things like fridges etc to go on update cycles like they fucking iPhones – utterly pointless and wasteful.
Thank you, this made me smile for a change RE this boring topic. Great to see AN lover of tech taking on this Glorious AInimal and kicking it's ass.
While the world is worshipping at the Golden Calve some of us, albeit a minority, can see the horrific feet of clay. It will be so nice when this jaggernaut implodes, which we know, it will.
Probably three decades ago I was doing some interesting stuff with this AInimal at IBM. At least we could help some physically challenged people in the process.
So nice to be part of your club, you should actually create a undercover UG around this, it could be fun to help constructively derailing the dreams of all these mad scientists!
The hands down absolute best IoT thingy had to be (drum roll, please) …..
“The Cellmate Chastity Cage, built by Chinese firm Qiui, lets users hand over access to their genitals to a partner who can lock and unlock the cage remotely using an app. But multiple flaws in the app’s design mean “anyone could remotely lock all devices and prevent users from releasing themselves,” according to UK security firm Pen Test Partners. Breaking open the chastity cage by hand would require bolt cutters or an angle grinder.” https://www.theverge.com/2020/10/6/21504019/internet-enabled-male-chastity-cage-cellmate-qiui-security-flaw-remotely-locked
I think that, this time, it will depend. A toaster you can talk to, sure, that's useless. But whilst a vague "connected-to-the-internet" IOT feature is not actually much of a feature, the capabilities of LLMs definitely are a feature that can be meaningfully deployed. A physical chessboard that you can talk with about the layout of the pieces? A colour changing light bulb that might actually understand a voice instruction like "the green from the Matrix movies"? A physical globe toy that you can interact with and discuss the geography and history of the places? I just made these three up on the spot and I'm definitely not a creative person. This is not a repeat of 3D television
Your article makes the point that IOT devices never had any utility, designers couldn't figure out what the connection was for. Essentially you say, no one can think of any ideas for products with gen AI either, hence gen AI is equivalently useless. But it's much, much easier to think of meaningful and varied use cases for AI in products and services, e.g. my three trivial examples, whereas in IOT it was always "you can control it from your phone", which begged the question, "why?".
"LLMs are not revolutionary" is a poor choice for a hill to die on, I think.
I just read about a study on Nigerian school children, introducing AI into their educational experience had the effect of "compressing two years worth of educational attainment into six weeks" (paraphrased). There's a raw power here, behind this technology, of a fundamentally different quality to "let's connect a device to the internet".
From the perspective of your point in your article, about products, it will feed through I think. The consumer-facing internet was around for many, many years before IOT came along. ChatGPT was launched November 2022, it's still early days.
Sorry, I mean that it’s always touted as revolutionary by those selling it. I’m not convinced. It will have utility, and smart people will figure out good use cases, but I’m not sure it’s going to come in the form of stuffing into our software or products.
I’m waiting for some McKinsey scum to say the quiet part out loud, fold IoT and generative AI into one (useless) product, and sell it to executives as “Bright Shiny Objects In Primary Colors.”
I work in home building, specifically Offgrid, and the IOT everything was stupid. Unless you’re building an Offgrid house.
The reason is, you actually needed applicances to talk and work together for legitimate cost and value reasons.
For instance, in an Offgrid house you need to not have everything running at once, you need the water heater, the hvac system and Tesla charger to ramp up and down based on what else is running for a savings of $7000~
If you also can have those items run preferentially during the day, you can save another $10,000-$13,000
In lithium batteries
The problem with IOT is that there wasn’t a real use case, but there is for Offgrid.
Ai is world class at online search, try perplexity pro
I use it to read manuals and configure settings for my solar and battery systems
Ai can do math faster and better than 99% of the population, we use it to analyze our building plans, measure square footage, generate a materials list and estimate the cost to build a house
We use it to calculate manual J calculations
I’ve used it recently to engineer my concrete slabs
IOT replaced nothing.
Ai is replacing engineeers, accountants, sales reps, customer support people, teachers, researchers and personal assistants.
They are not the same as IOT, this article is true to the extent that some things don’t need AI
But you’re missing the fundamental part that AI is a BETTER engineer already than any engineer I’ve ever worked with. And a better personal assistant than you could ever hope to hire
I’m hanging on to my 2003 BMW sedan with a maniacal grip — as every car I’ve considered buying today would jam me into the auto industry’s version of IofT—and never mind all of my driving habits being sold to the insurance companies. Thank god I live on an island, I’ve only 70 thousand miles logged on a 20+ year car. They’ll bury me in it.
Love it! They can float you into the ocean in it like some kind of Viking-like tribute
The central mistake is the propeller hats keep putting plumbing as the main feature. IoT works fine if automation and networking is wanted,just as blockchain is a clever key exchange mechanism, but these are not user facing. These should make the system more efficient but they keep ramming them into everything and telling us it’s better than the invention of refrigeration (which it’s not). AI should be a feature not the whole show. But alas, propeller hats.
Propeller hats. I like that.
Best reply so far
Glad to deliver. FWIW,my full stance. If we spoke of GEN AI as not some new fangled, confusing tech black box but as an automation tool then I believe the average person would be better equipped to have a nuanced discussion. AI automates tasks that would take time - see the kerfuffle about improving accents in the movie the Brutalist. A human could manually enhance an accent (either by the actor getting better 🤣🤣 or by manually tweaking knobs on a device) but it currently takes too much time. There is a quality/character difference between automated output and handmade, let’s have that discussion and let people put their values on the table.
I still remember reading the book Abundance ten years ago that told us of all these promising IoT things that would be happening in the next 5 years. I’ll never forget every soda cup in the world was supposed to have an IP address so if you didn’t pay for soda the machine just wouldn’t turn on when you put an off network cup beneath it. I think about it all the time and how believable it was then and how insane it sounds now.
And completely pointless? Stolen soda hardly a world-ending problem. A great example of finding a problem for the solution.
Really great, succinct. This is the kind of sharp and short anaylsis you can send to friends who are not yet entertaining what a disaster it could all be.. LLTM
Feel free to share! ;)
Already done (x3!) my friend : )
Another flaw with these high-tech Internet of Things objects is that when the company stops supporting the apps or networks they depend on, the product becomes much less useful, or even useless. And your local appliance repairman is not going to fix the wifi on your fridge. So even though you've purchased a "thing" you don't fully own it. More often than not, you've just subscribed to it.
What the IoT has in common with AI (and crypto, too) is that it's often a form of overengineering. A complex solution applied to simple problems. A sledge hammer used on a finishing nail. Too many AI use cases being proposed are for things that can be done almost as well but much more efficiently without AI. Ex: Google search used to be great, then they slowly made it terrible, then they applied AI to "fix" it, and now it's even more terrible, but also helps to boil the oceans.
AI makes good sense for some uses (let it analyze a billion x-rays and learn to detect tumors, sure!) but for many things it does not make sense. Yet we are having it forced upon us anyway, because someone needs to profit from us at our expense..
I read somewhere they are doing this on purpose to stop people buying appliances for life. They basically want things like fridges etc to go on update cycles like they fucking iPhones – utterly pointless and wasteful.
Thank you, this made me smile for a change RE this boring topic. Great to see AN lover of tech taking on this Glorious AInimal and kicking it's ass.
While the world is worshipping at the Golden Calve some of us, albeit a minority, can see the horrific feet of clay. It will be so nice when this jaggernaut implodes, which we know, it will.
Probably three decades ago I was doing some interesting stuff with this AInimal at IBM. At least we could help some physically challenged people in the process.
So nice to be part of your club, you should actually create a undercover UG around this, it could be fun to help constructively derailing the dreams of all these mad scientists!
The hands down absolute best IoT thingy had to be (drum roll, please) …..
“The Cellmate Chastity Cage, built by Chinese firm Qiui, lets users hand over access to their genitals to a partner who can lock and unlock the cage remotely using an app. But multiple flaws in the app’s design mean “anyone could remotely lock all devices and prevent users from releasing themselves,” according to UK security firm Pen Test Partners. Breaking open the chastity cage by hand would require bolt cutters or an angle grinder.” https://www.theverge.com/2020/10/6/21504019/internet-enabled-male-chastity-cage-cellmate-qiui-security-flaw-remotely-locked
I can’t wait for the AI upgrade.
What could possibly go wrong ???
Haha. That's a hall of famer right there.
I think that, this time, it will depend. A toaster you can talk to, sure, that's useless. But whilst a vague "connected-to-the-internet" IOT feature is not actually much of a feature, the capabilities of LLMs definitely are a feature that can be meaningfully deployed. A physical chessboard that you can talk with about the layout of the pieces? A colour changing light bulb that might actually understand a voice instruction like "the green from the Matrix movies"? A physical globe toy that you can interact with and discuss the geography and history of the places? I just made these three up on the spot and I'm definitely not a creative person. This is not a repeat of 3D television
I mean... they are use cases, for sure. Hardly revolutionary though.
Your article makes the point that IOT devices never had any utility, designers couldn't figure out what the connection was for. Essentially you say, no one can think of any ideas for products with gen AI either, hence gen AI is equivalently useless. But it's much, much easier to think of meaningful and varied use cases for AI in products and services, e.g. my three trivial examples, whereas in IOT it was always "you can control it from your phone", which begged the question, "why?".
"LLMs are not revolutionary" is a poor choice for a hill to die on, I think.
I just read about a study on Nigerian school children, introducing AI into their educational experience had the effect of "compressing two years worth of educational attainment into six weeks" (paraphrased). There's a raw power here, behind this technology, of a fundamentally different quality to "let's connect a device to the internet".
From the perspective of your point in your article, about products, it will feed through I think. The consumer-facing internet was around for many, many years before IOT came along. ChatGPT was launched November 2022, it's still early days.
Sorry, I mean that it’s always touted as revolutionary by those selling it. I’m not convinced. It will have utility, and smart people will figure out good use cases, but I’m not sure it’s going to come in the form of stuffing into our software or products.
I’m waiting for some McKinsey scum to say the quiet part out loud, fold IoT and generative AI into one (useless) product, and sell it to executives as “Bright Shiny Objects In Primary Colors.”
Oh, I'm sure they will, all for the small consulting fee of millions of dollars.
Unfortunately, the tech bros just got a promise of $500 billion in government aid from the Trump administration.
That would be real troublesome if not for the fact that Trump’s promises often fall short when it comes to actually paying for stuff.
Richard had some good points
Here are a few of my thoughts
IOT didn’t have a real use case
Ai has at least 15
I work in home building, specifically Offgrid, and the IOT everything was stupid. Unless you’re building an Offgrid house.
The reason is, you actually needed applicances to talk and work together for legitimate cost and value reasons.
For instance, in an Offgrid house you need to not have everything running at once, you need the water heater, the hvac system and Tesla charger to ramp up and down based on what else is running for a savings of $7000~
If you also can have those items run preferentially during the day, you can save another $10,000-$13,000
In lithium batteries
The problem with IOT is that there wasn’t a real use case, but there is for Offgrid.
Ai is world class at online search, try perplexity pro
I use it to read manuals and configure settings for my solar and battery systems
Ai can do math faster and better than 99% of the population, we use it to analyze our building plans, measure square footage, generate a materials list and estimate the cost to build a house
We use it to calculate manual J calculations
I’ve used it recently to engineer my concrete slabs
IOT replaced nothing.
Ai is replacing engineeers, accountants, sales reps, customer support people, teachers, researchers and personal assistants.
They are not the same as IOT, this article is true to the extent that some things don’t need AI
But you’re missing the fundamental part that AI is a BETTER engineer already than any engineer I’ve ever worked with. And a better personal assistant than you could ever hope to hire
Maybe the data harvesting was the main point after all, rather than the actual utility.