I agree. The same thing happened with hardware engineering through the 2000s. You used to be able to have an incredible career designing fun products and earning good money if you could do the hardware and embedded software. Then they started outsourcing everything. Over that time I looked at the situation in Shenzhen and sometimes wish I'd been born in China - they got what we used to have. Malls full of electronics part manufacturers and a rapidly developing eco-system able to crank out amazing products.

It was way more exciting than doing yet another iPhone app, but at least the app stuff was pretty easy and paid well. Now the writing is on the wall for app stuff as well. The whole tech thing is a mess.

One thing I have learnt though, is that there is very little connection between how hard you work or how difficult your job is vs the pay. You can still work as an electronics engineer in the west, but you'll work at least twice as hard as you would have back 20 years ago, for the same pay without inflation adjustment. You'll also be the lone voice at the company lamenting corner cutting and duck taped solutions for everything, yet be blamed and have to pull the weekend shift when the product starts catching fire because of the dodgy temu batteries. It's not worth it when a good EE can do pretty much anything else.

Ultimately I don't know what the advice will be for my kids. I think we're just moving to a new sort of feudalism, where pretty much all jobs won't pay enough to have much of a life, so it will all come down to your family's wealth. I see that I have an opportunity to provide them with a bit of a financial base, though I think the window is closing. So really the key is have rich parents and then just do whatever you want since none of the jobs are really going to pay very much unless you have the right connections.

I don't deny that the financial sector has important uses, but a large part of it now is just a make for work scheme funded by government bailouts. Outside of things like keeping your money safe, facilitating trade, and insurance etc, its main purpose is supposed to be the efficient allocation of capital, and this dominates the sector. Yet we repeatedly see situations such as the GFC where it essentially just perpetuated a gigantic ponzi scheme, and rather ironically, it was so bad at allocating capital to growth that even when the price of capital was negative (zero interest rates) for over a decade, it still kept trying to feed money into various ponzi scheme, rather than invest in actual real economic output.

It is laughable in how useless all the modern derivative, high-frequency-trading and other innovations are at generating real economic growth. The victorians built the first industrial society with log tables and quill pens.

Anyway, my point is that we already know that we don't need much of the financial system, yet it continues to propagate and grow like a cancer, so I have no doubt that the good folks at Goldman Sachs will find yet more ways to harness the talents of our best and brightest in the pursuit of pointless financial innovations, even if this AI replaces a lot of their existing work.

For the engineers out there - small gas turbines are fundamentally inefficient because the boundary layer (the layer of air next to the blades/stators that is not moving) does not scale down as you make the machine smaller. So the boundary layer becomes a larger percentage of the flow cross section through the engine. The boundary layer is a large loss component, so this means that the smaller you go the more losses you get as a percentage of mass flow rate.

There is no way around this problem. It is why all attempts at making small but efficient gas turbines have not been commercially viable.

Turbochargers work because they capture waste heat energy from the exhaust. But the additional losses they introduce mean that they rarely result in a net efficiency gain for the engine (just more power from a smaller engine).

For efficient range extending you want something like a normally aspirated engine running the Atkinson cycle which is what the Priuses uses. The Prius has about the most efficient range extender you could come up with, and the power splitter gives a slight reduction (I think it's about 30-40%) in electrical machine size compared to a pure series hybrid. In other words, the people who designed the Prius did a good job, and that is really where you end up with you want an optimal hybrid design right now. I'm sure if it could be done better with a gas turbine, the people at Toyota would have looked into this.

They wanted to use Model E but that was already copyrighted by Ford (now used for the Mach-E). Hence the Model 3.

I get where you're coming from, but I think we also have to accept that we have only an extremely basic understanding of how the brain works, and still don't really understand what consciousness is at all. Our present situation reminds me of the early experiments with 'evolution' where people thought you could evolve bacteria by putting some chemicals in a glass jar. If they had had access to an electron microscope to see how complex even a single cell is, they would have realised how absurd it was to believe that you could just mix some organic compounds, wait a few days and get mulit-cellular life.

The 'neural nets' we have are a weird proxy for how we understood the human brain to work in the 1980s. We know that the way neurons work is way more complicated (chemical signalling, much more interconnections, each neuron being like a city in complexity etc) than our simplistic models (simple non-linearing gating and single dimensional interconnects etc). It's actually impressive that they are as useful as they have turned out to be.

Of course we may be close to stumbling upon the secrets to AGI, but if we did it would be just blind luck. I think it's much more likely we will continue to develop a suite of algorithms that can do things that humans can do more and more efficiently, though this will be a long slog of incremental improvement.

At some point we might be able to build a robot that can essentially emulate a basic human, and I guess at that point the philosophical question is whether that machine is conscious or not if it pretends to believe it is.

I think this idea of a super intelligence is really dumb though. For all we know (and there is some evidence with autism etc), there is a inherent tradeoff between pure intellectual capability and stability in other areas of the brain. Or there might not be. But we have no idea, so just guessing that once we can create an AI it will be able to improve itself in a pretty naive position to take (though understandable if your goal is to become a mega billionaire).

Ives is an artist, not a business person, nor an engineer. I think you've got to realise that, while I'm sure it was pretty exciting working at Apple on world changing work, the actual designs he was cranking out were pretty braindead. In most IT products, nobody really wants the actual device to be the focus, it's meant to facilitate the content, and that's why something like modernist principles is where all technology design has headed. I like modernism, but it hasn't been innovative for 80 odd years now.

Back when he started in industrial design in the 80s/90s/00s, it was quite a unique time (remember IDEO), because technology was long behind where our imaginations were. For example, everyone knew that a flat screen monitor would be the best format for a computer, but we couldn't make them. So you had to use design language to make that big bulky CRT screen look interesting. That is a good use for design. But once we had flat screens and compact CPUs etc, the need for all that evaporated.

I don't think he would have stayed at Apple if Steve hadn't been there. Steve probably just genuinely enjoyed having the 'cool artist friend' at work (in the jobs book there were stories about all their fun 'work' trips to Japan), when almost everyone else was a business person or engineer. But once Steve died, he was basically stuck deciding what radius to make the iPhone edge for the next three generations. That must be so braindead, and I can see why he tried to push things that nobody really wanted in an attempt to make the designs more interesting (I imagine the butterfly keyboard was about going towards a haptic keyboard and then being able to have a dual touch screen type device or something).

I haven't really followed his other design work, but that in itself suggest he isn't an amazingly notable artist. But whatever, he's done well for himself, but I definitely do not think he is or ever was much more than Jobs' artist friend who could help with the aesthetics on Jobs' big new idea.

Yeah I agree as well. I haven't had to do much MS code for a while, but wow, they just generally have the crappiest architecture design I've ever encountered. I mean it is the place where the lament 'architectural astronaut' was born.

When I was younger, I tried to understand what they were doing with their various driver models, franken-unix style file system, bizarre layer-apon-layer of UI 'object' models, and that nightmare that is the registry. It's just all horrible, and I genuinely think many of the original underlying architectural concepts have been lost to time and incompetence.

When I do MS work, I do find myself just copying and pasting big chunks of code because life is too short to try to understand if there is any sort of coherent design going on (and these days I assume there isn't). If AI code paste can help untangle that, then good for them.

But when I'm dealing with a well designed library, such as Cocoa even, I can't really see the big benefit of such an AI thing. A well designed library has a simple conceptual model that gets out of your way and lets you focus on the project's business logic.

I think the whole college industrial complex is doing a pretty good job of destroying itself anyway. My friend recruits electronics engineering grads. He has had to develop his own basic skills test, because the grades and degrees these people are getting has little correlation with their actual practical abilities. I mean, one question is 'draw the circuit for a bridge rectifier'. Many grads cannot do this. He has had a number say they need a computer so they can check on google. How are they getting EE degrees without understanding basic electronics? There were some useless people in my EE class 20 years ago, but all of them would have been able to explain what a bridge rectifier was.

Also go look at the scandals around places like Juliard where they have milked their reputation while putting increasingly less effort into actually teaching students. This is basically private equity stuff - take over a business with a reputation for quality then cut costs. Here in the UK, when I graduated, about 7% were getting first class honours - now it's at nearly 30%. The measurement is becoming worthless.

I think for many jobs with hard skills (such as software), independent proficiency assessments like LEET testing will be the way forward. Business groups will set up assessments and if someone can get a top result in those they will be in, regardless of whether they got the knowledge through university or by watching youtube videos. College will remain for those who go into career pathways where it's about knowing the right people - but it will become increasingly pointless for anyone who does not already come from a rich and connected family.

If you don't come from a rich family, and don't have hard skills then you will be screwed, and I expect this number will grow massively as automation (including AI) raises the bar for what hard skills you need to have. At some point there will be so many people struggling to find careers that it will become politically untenable to continue running our economy that way, and then we'll just have to hope someone has come up with a better idea by then.

I imagine the delayed payment thing is funded by the merchants (vendors) because it presumably encourages people to buy more and/or allows smaller shops to compete with bigger retailers who might offer their own credit programs. Merchant card fees are quite high (around 2-4%) especially for smaller vendors, so if they can force out other payment providers with this 'deal' then that could be a net win for them. I also imagine they get a certain number of customers switch over to credit at high interest rates because of the low friction.

The real scam in all of this is how merchant payment processors have inserted themselves into the process in a way that is transparent to the user, and then they absolutely fleece business with no ability to negotiation. A business I owned used to take credit card payments over the phone. We were low volume but high value items, and one day Visa or whatever rang us and said they had to up our fee for no reason. We asked why and they said most of our customers were platinum card holders so it cost them more to give them all the rewards, and basically, we were going to have to pay for that. It was easy for us to just bump a bit more onto our prices, but I realised how dumb it was that all the guys flashing around their fancy credit cards and bragging about the 'free' stuff they were getting were actually just paying for it (and then some) at all the high end shops they spent at.

The model isn't to make profit, well, at least not in a sustainable way. The goal is to IPO - which is exactly what Klarna was trying to do before it got hit by the deteriorating economic outlook.

If consumer spending hadn't hit headwinds, then their plan would have been to keep 'growing' sales (loan issuance) while keeping delinquencies at an okay level. Engineer a couple of quarters of profit, and then do an IPO with the promise of some sort of exponential increase in revenue due to leveraging their large user base and partnerships with merchants blah blah blah.

The VCs and early founders cash out and then what happens to the company after that is not really their concern.

These buy now pay later companies have been around for ages. They periodically collapse when it turns out that the 'asset' of loads of unsecured loans from people who have no money (i.e. it's not even worth pursing them though the courts), is basically worthless in any kind of economic downturn. As far as I can tell, the main leverage they have to make people keep paying the loans is the offer of more loans. As soon as this process breaks down (you decide you really don't want to keep giving them more loans) then it all comes to an end quite quickly.

The big difference is that even if you own a car, you still have to pay to use it. You have to pay taxes, insurance, fuel, parking, maintenance. Even if your car sits in the garage all the time, you're still paying a lot to own one.

With software and media this is not really the case, though a lot of software does have ongoing costs these days so some kind of recurring charge is valid.

I also think most people are just annoyed about the exploitative nature of software as a service, not the actual concept. For example, Apple has the dumbest tiers ever for their iCloud storage. You can either get not enough or way more than you need. Other firms (ahem accounting software) make it hard to change over, and keep jacking up the price each year. Adobe makes it hard to stop the subscription.

If we could have faith that the service would be fair, then I think many people's problems with SAAS would disappear.

For example, if Apple just charged me a fair amount for every GB I used, I would absolutely go nuts on iCloud, but their dumb MBA designed system pisses me off so much that I put the effort in to using alternatives.

I really don't think money is the motivating factor for him. If it was, it would be way easier to understand the guy. If he cared about money he wouldn't have completely tanked Tesla sales with his stupid behaviour. No, the trouble is that he wants power and to smash up a system he doesn't like - he is ideologically driven by a weird ideology. Wanting to change the system is not necessarily a bad thing (there are many problems with our current system) but the guy is ignorant. When he learnt that there is not a giant vault full of the dollar bills everyone had saved up over the last 50 years to pay out social security, and than pranced around telling everyone that he had made the incredible discovery that social security payment come from current tax receipts, it was like 'how can this guy not understand basic economics?'.

That level of ignorance is incredibly dangerous when combined with hubris and access to power. He probably genuinely thinks he's saving everyone because Theil gave him a copy of Ayn Rand for dummies a couple years ago.

The whole lot of them seem incredibly ignorant, but I also blame the establishment for this. The establishment is not ignorant of the structural issues we face, but basically refused to do anything about it for the last 20 years. So the people who have been screwed over by neoliberalism are now supporting anyone who will actually try to change the system. As someone who got hammered by not getting on the housing bandwagon earlier in my career, I could scarily seem myself falling into that camp if I didn't have decades of critical thinking skills developed through an engineering career. But I vote for establishment players with gritted teeth, because I know things can get a lot worse than simply having no real financial future.

Yep. And for his next trick, watch Tesla roll out a fleet of robo taxis in June (as promised) that seems to perform better than Waymo and don't need lidar so are much cheaper.

We will then hear how these will be rolled out en-mass 'next year'.

The vehicles will just be teleoperated, but if anyone pulls them up on this they will point out that Waymo also has remote safety drivers. Nobody will be able to tell how much the vehicles are being driven by AI vs the teleoperator, so this is where the 'magic' will happen.

I have absolutely no doubt that this will work and $TSLA bulls will eat it up.

Thanks to starlink he might even live demo a car being remote summoned from the other side of the USA (tele operation now works anywhere!) as originally promised a decade ago.

Elon is not the only one who has destroyed the tech world with BS (ahem, anything AI...), but he is becoming one of the worst. It's just good to know there are still businesses out there who are busy doing the hard and boring work of actually solving real problems (thanks Waymo).

I agree. I saw their cars in SF recently, and while I didn't ride in one, it was incredible to see how well they integrated with other traffic. I was also surprised at the variety of conditions they could drive under - people trying to do weird u-turn manoeuvres, pretty poor road conditions/markings etc. Basically driverless cars are here, and I think the media has expended so much effort on the Tesla driverless hypefest over the last 10 years that they are totally missing the massive revolution that is now happening.

The cars are going to be everywhere within a few years. It's going to absolutely decimate driving jobs. If their tech can do what it's doing on busy city streets, truck drivers on the highway are essentially unemployed at this point - subject only to Waymo getting around to addressing that market.

The entire automotive industry is going to be disrupted - young people don't care about cars anymore, and will happily give up having to learn to drive to use taxis. The die-hard ageing boomers who wouldn't have given up their steering wheel without a fight will realise that these things will allow them to live independently as they rapidly age, so will embrace them as well. The whole concept of cars as a status symbol was already dying and these things will kill it.

Kudo to the the team at Waymo. Rather than pumping the hypefest and all cashing out as millionaires, they have keep grinding away for over a decade now to actually make something that works.

While my memory is fuzzy, I'd probably agree that homework at school level is pretty pointless. The reality is you are not learning very much at school. But when you get into a decent university program (e.g engineering), there is simply not enough time to learn everything in lectures. You can get a brief overview, but if you really want to understand something you need to explore the space in your own time. At least for myself, I found this was the case. There were certainly people who saw homework as just a box to tick, and frankly, those people became terrible engineers.

For example, I can listen to hours of signal processing lectures, and quickly get the general gist, but if I had to apply a concept to a project, I'll have to work my way through a text book and grind equation out on my own. The detail is just a whole different thing from having an overall understanding, and I don't see how you could learn that in a lecture or lab context.