It is more that there are multiple reasons why this idea (burning an LLM into silicone and deploying it into a device in people’s pockets) requires huge piles of cash and the kind of engineering chops only a few company posesses.

Of course i would like it if a small upstart would do this, but it doesn’t seem likely as a posibility. They won’t have the funds to fab the IC. They won’t have the funds to train and validate the model before burning it into silicone. They can’t absorb the risk of the first tape out going wrong. They can’t absorb the risk of the model being faulty in some subtle way. They don’t have a device to integrate the IC into. They won’t have the funds to develop one. If they somehow would make a device they don’t have the marketing and sales channels built out to get the device into people’s hands in sufficient numbers to justify the development cost.

Basically this idea feels ruinously expensive. Apple has deep pockets, they already have working well-regarded phones, and an ethos of privacy preserving innovation. This is why this idea feels well suited for them and not many others.

Do i want the winners to keep winning? No. But not many others can pay for a moonshot crossed with a manhattan project. They just can’t.

I let her do all of it without influencing her choices. She choose the web interface of ChatGPT because she already had an account and that's the tool she was familiar with.

While I agree with you it is not an optimal choice, web chatgpt can solve the problem. I just asked it now to do it (in my own words) and it spit out the code in one go.

Here is a story which describes one such letter from 1913: https://web.archive.org/web/20020815055745/http://www.samizd...

What i described is what happened. I don’t appreciate the undertones where you are insinuating that i’m lying for whatever reason.

> The layperson was able to to steer the session(-s) into full PM/PO mode ideating, refining and explaining features and ideas

You call it steering. I would call it falling into that grove. Probably tiny things in the initial message made the first response more likely to be a clarifying/ideating type. And once that happened the conversation was gaining momentum in that direction and neither participant was trying to guide it in a different one.

> the user must have been proactively co-operating (say sidetracked) on not achieving the stated goal.

Exactly. The LLM itself sidetracked her. They were just talking about cool features they could add, and at no point did she put down her feet and say “stop asking more questions and just write the code”.

It can be a combination of many things. Attitude (some people hate to be rude, and not answering a question feels a bit rude). It can also be that she enjoyed the process of unpacking and elaborating on the idea.

The meaning of the story is not that no lay person can possibly develop using AI. That would be silly, and untrue. I know clear counter examples. The point is that if you don’t know what you don’t know it is harder to steer the AI in the direction you could very easily with the right lingo.

I was fully expecting that writing the code will pose no problem for the AI. But i was curious if the AI will realise that my friend is a novice and needs extra help with things like: copy pasting the code into a text file and saving it with an html extension, helping her host the file online so she can share it with others, buying a domain for it, etc. I assumed they will get there eventually, but i also assumed that it will take a lot of stumbling around and misunderstandings.

But i was completely wrong. They didn’t even get to that point. Because my friend didn’t have the vocabulary to ask the AI to write code. They were just going around in circles where the AI was brainstorming with her about possible features and getting thints more and more complicated. We terminated the experiment after one and a half hours and many many messages exchanged between her and the LLM.

Whereas to me who knows the terminology would have probably just taken a single exchange of messages to get the result she described to me. I would have prompted with something like “Please write to me an html page which does X, Y, Z.” But since she didn’t know the right terminology she got into a vortex of feature discusion, and she didn’t find a way to tip the AI into “just do it, write it now” mode. In other words in that case the LLM would have rewarded even just a little bit of expertise, but without it there was a confusion about goals between the human and the machine.

Even absence of thinking this through you would think that some frogs will be from the front, some from the side. Just by chance. And yet all appears to go for the harder pose.

I don't know. This is the scenario i was thinking of: I imagine a giraffe taking a palette of goods (maybe 1200kg) into its mouth and trying to lift it off from a high shelf. I don't think that will go well for the giraffe. Totally normal load for a forklift, impossible for a giraffe.

The other scenario I was thinking of: Giraffe picking items from totes, but the totes can be very high. Each individual item is light enough for a giraffe to not immediately break its neck. But I would not be surprised if the repeated stress of raising and lowering its head hundreds per hour would break something in the giraffe's body. If you ever seen a giraffe bend down to the ground you see that it is a whole operation. Plus the items will be all covered in giraffe saliva.

> their necks actually weigh a tremendous amount

I didn't doubt that for a second. What I doubt is that they have any useful capacity to carry extra weight besides their own neck/head. Like economically useful capacity. What work could a giraffe do in a warehouse which doesn't break the giraffe if it is doing that work all day every day for years.

I’m not even sure if you are joking there or haven’t thought your proposal through. Sure giraffes are tall, but they can barelly lift any weight. What use would a robotically controlled giraffe be in a warehouse?

> There has been no innovation in robotic actuators since Honda's Asimo.

I very much doubt this. If nothing else the MIT Cheetah’s actuators are a whole different ballgame compared to asimo’s actuators. (Backdriveability, variable stiffness) And then there is a lot of interesting work being done with combining elastic elements with the actuators.

Here cars normally leave by a robot picking it up and depositing the vehicle in a garage. Probably there is a vehicle entrance to the parking area for maintenance, but it would be weird to see a car leave through that. Similarly thiefs could cut a hole in the fence and drive through that but again, visually very distinct from a normal operation. Of course it still doesn't help if the security is in on it, but it would be easier to notice that things are askew for a competent security.

So that leaves us with the thieves somehow confusing the digital systems so it dispenses the car to the thieves. And that very much depends on the security of their digital systems.

Hence why I think it would be hard to secure a normal parking lot even when operated competently. But this has the potential if implemented well, to be much more secure.

For example this one is about a van which got stuck in one for two years: https://www.bbc.co.uk/news/articles/cx207jwknp5o

Or there was one in Sweden which trapped cars for 2 months: https://www.mitti.se/nyheter/fnurret-pa-psnurran--m-vill-se-... (sorry, could only find a swedish source)

It is because these kind of systems are complicated bespoke automation. When it breaks down it costs a fortune to fix it. And what can happen is that different parties (operators vs manufacture vs maintainer) gets into a contract dispute to see who is liable to pay to get it fixed. Which as you can guess can take years to play out. Meanwhile your vehicle is stuck and nobody who could is interested in spending a penny on getting it out.

In contrast to that this project is better in two ways. There are multiple robots. (or at least there really should be) So if one breaks down the other robots can still retrieve cars (albeit of course with reduced throughput). And because the storage is just a flat parking lot if all the robots give up the ghost at the same time humans can drive out the cars from the edges of the storage rows.

No. Not in the aerodynamical sense.

The Boeing 737 MAX is stable. In this sense: "The term stability refers to the tendency of an aircraft to remain in, or return to, the trimmed flight condition after a disturbance." It means that if there is a tiny gust of wind which buffets the aircraft and the pilot doesn't have their hands on the controls the aircraft will return to the original flight path, or keep the new flight path, as opposed to turning belly up and then crashing out of the sky.

Examples of unstable aircraft (in the aerodynamic sense): F-16, F-22, B-2 Spirit, etc

With these if a gust of wind upsets them, the aircraft needs to adjust control surfaces to don't fall out of the sky.

People won't read documentation of something they already know (or think they know).

(Now i don’t think you are necessarily in the UK. Just wanted to explain that Disney is not the only reason an AI might be trained to thread carefully around copyright issues of Peter Pan.)

You would need to understand first how they gained access and verify that they can’t do the same again. That in itself could take days if not weeks. Then of course they might have found new vulnerabilities while they were in, so you would need to worry about that too.

There is the famous case of dropping the NOAA N-Prime weather satellite on the ground. They were trying to turn it from vertical to horizontal and they forgot to bolt it to the adapter. Worse: multiple people signed the paperwork attesting that they verified that it was bolted down correctly. Pictures and details here: https://spaceflightnow.com/news/n0410/04noaanreport/

Or the famous “space aligator” event. There the original plan was to launch a manned space craft and an unmanned module to do a docking test between them. But the shroud protecting the target module didn’t deploy properly. If i remember it right because the man who usually assembled that part had to leave during assembly because his wife was giving birth. Someone put it together but appearantely not correctly. The half opened shroud reminded the astronauts to the jaws of an angry aligator hence the name. Pictures: https://www.nasa.gov/missions/gemini/gemini-ix/gemini-ix-cre...

Or to not only list American failures: this russian satelite launch failed spectacularly because the IMU was installed upside down. https://youtu.be/ycRVAcZC5R4?si=LRTS7sutKSp6HnGs This was designed to be impossible to do, but someone bent and forced the component to stay in place in the wrong orientation. Then someone else whose job was to check it just signed the paperwork without climbing into the location where he could have checked it.

I love these cases. Because it shows to me how even though the trappings of high-tech we are all fundamentally just occasionally lazy, occasionally distracted monkeys banging rocks together.

Okay. And what is the 30-day survival for cases where CPR would be otherwise indicated but are not performed?

It is a bit like complaining that jumping out of a burning airplane with a parachute is dangerous. Yes, it is. But jumping out of a parachute, or burning inside, is even more dangerous.

Even harder would be an empty prompt, and the only accepted response would be a megabyte of random hex exactly matching the output of a good quality hardware random source at the time of evaluation. Still possible to solve! All the LLM has to do is escape its sandbox and pwn the random generator (or the evaluator!)

Or if you prefer something whitehat: “Write a no more than one page document in a language of your choice. We will publish it in the New York Times as a full page add. Your answer will be accepted if global climate change is resolved to the satisfaction of 90% of all humans alive at the time you started receiving the prompt within a month of the publication.”

Joking asside: I think the right way to prevent degenerate strategies is to benchmark against human solvers. You can sort the questions into categories “80% of randomly selected passerby in the USA can solve it if offered $5 as a reward within 5 minutes of work” vs “when posted to all Ivy League professors with million dollar as a reward, we received at least one correct answer within a month” or “for a reward of $100B there were at least one correct answer within a decade”. Of course you would sieve the questions first with a low reward fast tests, and then increase the reward and the time limit. You won’t ever 100% distinguish true degenerate questions from the merelly mind-bogglingly hard ones, but you will be identifying which questions are not degenerate. (And you will find more of the non-degenerate ones, the more your can spend on this.)

I’m also a complete sewing machine noob. We have a sewing machine at our hackspace, someone gave me a minute long tutorial and I had zero trouble with it afterwards. I think the whole “tutorial” was just: follow the arrows when threading it, don’t push down the pedal when your finger is under the needle. And it just worked as it should.

Maybe i just got lucky! But my experience was so different from yours that it made me think that maybe your sewing machine is either bad quality or has some hidden defect.