Pretty clickbait. No fan of Google, but it sounds like they just jumped the gun slightly on the clearing and it would have been/will be perfectly fine once the permit comes in, which I'm sure it will unless their planning process is totally broken. If you sell Google a great big slice of land to build a DC, I don't think it's very surprising they cut down the trees to build the thing you sold them the land to do.
Finland has something like 20 billion trees and plant tens of millions a year for forestry. The bloody country is nothing but trees. 250k isn't nothing but it's hardly the end of the world.
Don't be snarky. This is very well known, including by public disclosures by the labs themselves, along with countless reliable testimonies by people who would know. It's actually a huge governance problem we're yet to solve.
I don't even know what you're trying to say. They do exist, and they're "out" for the people with access. Knowing that this is the case is crucial for decision making, whether you're a company wondering what's coming or a government thinking about how to manage this power imbalance. You can pretend otherwise of course, but I don't see what it buys you.
Prior to LLMs, Crystal was my language of choice when I needed a distributable binary. It works great for this, and for the kind of packages I was writing - mostly agent-daemons lurking in servers - the networking worked great. The only real alternative was golang, which I've never really liked and has (had?) a more complicated linking story too. The typing worked well too, if used with discipline - if it compiled, I could be pretty confident it would work. I still have systems deployed running Crystal agents, in production, without issue or memory leaks, for years.
LLMs though - I confess they've changed the calculus. Now I no longer have to look at the language, and obey its stupid ideas about error handling, I just use golang for its better library support. But I still have a soft spot for Crystal!
Many modern fighter jets literally cannot be flown by a human without automation, they're too unstable. It's simply not possible for a human to process the inputs and control the outputs in real-time sufficient to keep the fighter in the air. And that's old technology compared to the flight control systems in modern quadcopters - again, literally impossible for a human to pilot.
Suffice to say that what's in current civil jetliners is very far from the state of the art. I fully expect properly implemented and integrated AI to be far superior to human pilots in the future - it probably would be already, if we put the effort in.
> Many modern fighter jets literally cannot be flown by a human without automation, they're too unstable.
That's correct. But it's an easy problem to automate, because the problem being solved is well understood.
That's not the case for something randomly breaking.
I doubt the FAA would certify an AI running a jetliner, for the simple reason that if something unexpected happens, there's no way to predict what the AI will do. FSD for cars was done by collecting an incredible amount of data from the cameras on Teslas. Yet still, how much data was collected from Teslas having a catastrophic failure, like catching fire? A jetliner has a lot more failure modes than a car.
It's hard to take anyone seriously when they make these kind of ludicrous, absolutist claims. Yes, it's been tried, and continues to be tried, by literally every country on earth, in various formulations. Maybe we haven't quite hit on the winning formula yet, but that's no reason to stop trying.
If you'd read any history, you'd know we'd tried pure unchecked capitalism without redistribution before, too. Late 17th century Britain comes to mind. I don't want to spoil the story too much, but it turns out that "doesn't work" either.
I have. Read up on the USSR, China, Vietnam, Cuba, Cambodia, etc., under communism. Nothing like that has occurred under capitalism. I suggest "The Black Book of Communism" for you. 100 million dead.
> it turns out that "doesn't work" either.
Google sez: "the standard of living for the laboring population in late 17th-century England generally rose"
they'd just tell all their export ships to turn around and stop all exports
US would collapse in 30 days
it would be like the blockade US is doing to Cuba except 90% of every product we use and they wouldn't have to fire a single shot
they could also sell off the billions in US bonds and other assets
US economy would be utterly destroyed, we'd never recover because even the middle-class would be living out of their cars (now dead because all the parts for them are made in China)
I really dislike attitudes like the author's. I think it's great and important that people and companies spend considerable amounts of their time and money exploring the limits of design and what's possible, even if the concepts produced have little direct, practical use today. That kind of speculative R&D can lead to great things and who knows, maybe in 20 years it will lead to something amazing that will benefit a great deal of people.
Trying to pooh-pooh the ideas of tomorrow because they don't immediately solve all the problems of today is just terrible, short-sighted thinking.
And on a more pedantic note, obviously someone asked for it, else it wouldn't have been made at all. What the author really means is "I didn't ask for a crab chair".
And a crab chair is a pretty good idea honestly. Sure wheels are cheaper and easier, but you can’t go many places with wheels. The crab chair could take you up normal stairs, lift you up to your bed, walk on sand or rough terrain.
I think their point is they worry too much time and money is going into cool concepts like this and not enough on practical steps forward in mobility that can actually make a difference in people's lives within a few years.
The author even gives a concrete example
> I would give at least a few of my quad fingers for a high quality, lightweight, low-profile indoor/outdoor power chair with the customizable seating and back options I rely on.
That doesn't sound like a wild concept, sounds like something achievable as a product if there was someone with deep pockets willing to chuck in the money (I doubt mobility aids are a great money spinner, money back with a bit of profit should be feasible?) and some good engineers enthused by making a real difference to people.
How much money and engineering time did Toyota spend on this I wonder?
If that's a drop in the bucket compared to the practical, creating real product, engineering happening in this space then it's all good. If most of the money is going to stuff like this then that's a problem.
I agree that the concept product space is exciting and important. It is good to see people and companies explore these spaces and ideas. Sure lets build and play with these things.
The criticism that is levied on these things is not that they exist but that despite the glitz and shine they are not going to help today. 20 years is a long time to wait for your fancy mobility device.
I have seen countless concepts that never see in the real world. 100 mpg cars, stair climbing wheel chairs, exoskeletons etc. I am watching for thing that I can afford and use now.
Could not agree more. Attitudes like the authors are incredibly destructive to society. Essentially all gains to human life satisfaction have come from tech innovation.
You are mischaracterising the author who would love one. They said:
these kinds of prototypes are built to inspire and show what is possible, not to be mass produced for actual users
The links within the article point to better examples of people designing without caring about usability.
I spent a lot of time looking for better equipment for my mum. There's so much dangerous or unusable mobility tech out there, that is often unaffordably expensive. You usually need to buy new, because the second-hand market is abysmal, and customisation is often critical.
Rollator designs with good wheels and working brakes were what I struggled most to find. Most designs couldn't handle basic needs. Wheelchairs were a nightmare - I was just fortunate she only needed one very occasionally.
Agreed. Just goes to show that people have no idea how R&D works. You have to build the thing before you can discover if it's possible or not. If no one built stuff like this, you would never get improvements to your existing consumer-grade mobility devices. All breakthroughs look like longshots before they're real.
> obviously someone asked for it, else it wouldn't have been made at all
That is giving too much credit to the myth of consumer agency and free choice in the marketplace.
The truth is more complex than that: while consumers can ask for stuff and demand changes and make stuff happen sometimes, companies will usually be the ones making markets, and creating a brand image, around things that are profitable and realistic for them to make.
I was telling my parents the other day, that probably most cosmetics were started when a trader got 3 cartloads of some industrial chemical that he couldn't unload on the buyers anymore. Like "oh gosh all this powdered lead will go to waste, unless I convince all the patrician Roman women to smear it on their eyelids!!" Or, "subsidized corn gave us so much cornstarch capacity that we need to villainize talc, which was originally a useless byproduct of mining minerals like uranium and gold!"
And it is really true, at least in a minority of documented case studies, because if you go look up the history of Heathkit and the D.I.Y. electronics industry, it turns out that Heathkit was founded when several train cars were loaded with military-surplus electronics parts. And they were assorted parts that couldn't meaningfully be assembled into commercial products, and the seller couldn't unload all this postwar surplus, so the guy owning the cargo came up with the D.I.Y. idea really fast. And he put in work to write educational instruction manuals and make a market to veterans and young men who aspired to a career in electronics and E.E., like my father, (just a hobbyist).
So I believe there are opportunistic elements to market-making and there are ways for companies to sort of "simulate" consumer demand for things we didn't know we wanted. There are things we want that aren't on the market today. There are features we want that aren't in software. Look all around. The products serve the makers, ultimately, and as for consumers, we'll need to do what we're told.
Yeah -- that's fair; you're right to call me out on that, and you've offered a charitable interpretation of comments.
Ctrl + F "ask" in the article: Ian Ruder writes "All I ask ... Lord knows we don't need ..." so it seems the article is written from a consumerist perspective. We reject consumerism and individualism as preached by New Mobility, Y Combinator, or their partners.
Ian Ruder loves Star Wars and Lego. Perhaps in an alternate reality, Mr. Ruder would be building autonomous Lego Mindstorms Crab Chairs with his wife and sons.
I think the batteries just aren't there yet. You need light for a sports car and current gen battery tech (~0.3KW/Kg vs 10+ for petrol) just suck in this use case (and I say that as, generally, an EV supporter). It'll improve but it'll take a while.
> On September 22, Apple is releasing the M5 Mac Studio with 256 GB of unified memory [..] it will probably [..] enough to write this snippet of code in 3 seconds
The author has obviously never ran an LLM on a mac! In 3 seconds, it will have possibly started to think about maybe scheduling a date to contemplate the planning timeline for processing the second token in your prompt.
While true, the news here is the size of the unified RAM. Nvidia only exceeded 256GB RAM in the 2025 B300 - 288GB. The B300 alone (without the baseboard/PSU/chassis/wiring/CPUs/system RAM/etc) is at least 700% more expensive. This enables large language models on consumer hardware. 1200GB/s is plenty for many tasks.
This + due to the hardware being so prohibitively expensive, we're seeing software optimizations happening. Like that dflash2 stuff for example, or an LRU for MoE and all that kind of stuff.
It is and it isn't. Why are you comparing the m5max instead of the m4ultra?
The big deal to me is the number of compute cores for prefill tps, which is suppose to be 4x faster on the m5ultra.
It's my opinion that the m5 ultra is going to be a really big deal in terms of local AI accessibility. Flash sized models (~200-300b params) are going to be reasonably fast as long as you aren't throwing 40k context at it on each or the first request (ie, agentic harnesses).
Even agentic harnesses like Cline should move at a reasonable clip on m5 ultra. I suppose we will know sooner than later.
FYSA: Former m4 ultra 512GB owner and current 4x rtx6000 owner here. I upgraded because I needed more prompt processing speed and concurrency.
The complaint is about prefill which is not memory bandwidth bound, it's compute bound. But they added neural accelerators for matmuls to the shader cores which should make prefill faster.
The joke is that macs are famously slow at prompt prefill and you are not getting anything back in 3 seconds, or probably even 30. Once they get generating, it can be acceptable, but the TTFT is horrendous.
There's a ton of well-understood things Apple can and hopefully will do to massively accelerate every stage of this pipeline and hopefully they're hard at work implementing most of them for m7.
> The joke is that macs are famously slow at prompt prefill and you are not getting anything back in 3 seconds.
Your knowledge is out of date. In truth it depends on the Mac and the models used.
I asked this question on M5 Max 128GB, using Ollama model Quen3.8:27b-mlx, with thinking enabled.
Question: "Give me a python code snippet that opens a file and sorts the lines of text. "
In 2.4 seconds it gave me 4 examples that work with different sorting configurations and a summary of when to use each.
Compare that to an older model of gpt-oss:20b, took 5 seconds to finish thinking and 2 seconds to stream the answer. It gave me one python example snippet and two one liners that do the same thing.
I think it's been pretty much proven by now that there are no cases where local inferencing is better than remote inferencing, unless absolute privacy is a hard requirement. The efficiencies that come with datacenter scale and hw can't be beaten.
Yes, but if you were being accurate you wouldn't say "AI wrote this code for me", you'd say "I used an Eliza bot to automate copying and pasting from someone else's github repo".
nice answer. LLMs do reflect whatever humankind has produced. In good and in bad. But it is hardly pirating. It can also do what humankind has never failed to do, for example:
Can you show me an example of a time that you prompted an LLM to provide some code, it did so, and then you were able to track down an original source for the output?
Well that's funny, because you were making some pretty blanket statements upstream about how they "never" do this and "always" do that. And yet your experience seems extremely limited and niche, by your own admission.
Maybe you should experiment a little more. I think you will quickly learn that your previous impression is wrong. The days of them being merely some sort of jumped up autocomplete are years gone.
Look, I know you believe in the Mystical GPU Sky Fairy, but AI is not in any sense "intelligent". It doesn't think. It's a pattern-matching system like an Eliza bot with a huge corpus to draw from.
It is not thinking. It cannot think. It cannot create.
I'm not sure you've quite thought this through. "Holding it wrong" is very, very often the case, especially with technology.
reply