The HT68F001 stands out: 512 words of program flash and 16 bytes of RAM. ... can only be clocked from an internal 32 kHz oscillator. Since each instruction takes 4 cycles to execute, this results in only 8000 instructions per second!
Too put the facts crudely, the world would be fucked climate change wise without China. The oft heard "why do anything while China is the problem" would be hilarious, if people repeating bald-faced bullshit didn't grate so much.
> And the industry might finally be waking up to the fact that writing code is a small part of producing software.
[Parts] of the industry are very aware of the fact, and have been for decades. In fact there was a book on the subject. You probably already are well aware of it. It's "The mythical mean month" by Fred Brooks.
He didn't have to contend with AI's of course, but the underlying driver was the same. He wanted to speed up writing software. Specifically OS/360, a new operating system for IBM, on which Brooks was a project manager. It was badly late, so they tried the obvious tactic on throwing go hordes of programmers at it. I doubt money was a problem, so the said programmers would have been good at their job. Those programmers weren't AI's of course, but the reasoning behind the move seems to be the same as here: OS/360 is just code, therefore the faster you can produce code the faster it will be delivered.
Brooks Law [0] summarises what he believes happened to OS/360: "Adding manpower to a late software project makes it later." Which is doesn't sound too different to the experiences mentioned here: AI's supplying tens of thousands of lines to a large software project that is well outside of their context window to understand is a hindrance, not a help.
Interestingly, that doesn't contradict the experiences reported by people vibe small projects, who say it is much faster. We had a term for difference between the two types of development back in the day: programming in the small vs programming in the large. It seems to have largely disappeared from the vernacular now. Pity, as I think it sums up where AI coding works and where it doesn't.
And it had same disconnect between the two groups, as we see between the vibe coders and the rest. People who spend their lives coding in the small have no idea what people programming in the large do all day. To them, people working on large projects seem to spend an inordinate amount of time producing very little code.
I was curious, so I asked an AI (Gemini) to compare the wholesale price of electricity in Ontario vs Sydney, in Canadian dollars, including any subsidies in the price. The reasoning is the wholesale price best reflects the cost of production.
The outcome was surprisingly close. Sydney seemed to be a little more expensive, with a spot market average of CAD$73/MWh vs CAD$65/MWh. A wash really.
I don't know what is going on with the retail prices. My rule of thumb is multiply by 3, but your multiple is closer to 4.5. I live in Brisbane for example, where the average price is $100/MWh and we pay around $.30/kWh retail. Have you looked at https://www.energymadeeasy.gov.au/ ?
> There are currently 419 reactors in operation, 76 in construction, 140 in pre-construction and 290 planned/announced. I have a slightly older version of that chart, where those numbers were 69, 92 and 178, respectively.
At about 1 GWatt per reactor, thats about 500 GWatt total new nuclear built over what must be decades, if it is built at all. A fair chunk of the existing 419 reactors will be retired in that time.
Meanwhile, Gemini tells me the planet added well over 100 GW renewable generation in 2024. That 100 GW is dispatchable. It was over 500 GW peak. Almost no renewables were retired in 2024. The rate new renewables are being added is growing at least quadratically.
Maybe Europe sunshine and wind resources mean they have no choice, it's nuclear or nothing. But renewables are being added at the pace they are for a reason. In the places that do have the renewable resources, they are far cheaper. If Europe is forced to go down the nuclear path, they are going to be paying far more than other places on the planet for their energy.
I don't know a lot about this, but for example an would app side loaded or delivered via F-Droid be subject to this policy?
F-Droid notwithstanding using an alternative app is not a very attractive option right now as there no good alternatives. But if Valve can create an Android app store that competes with the Play store, the in principle situation is very different.
Presumably not, as it's a contractual thing when you upload something to Google Play. Not sure if there are rules about only uploading to Google Play if you use it at all.
I've no idea, but I presume it's not even possible to use the Google Play purchasing API if your app isn't on the Play store.
I would not buy a FIDO2 token if it allowed anybody to reprogram it, including me. If you managed to make selling me such a device illegal, then may a pox descend on your house.
If I want to reprogram my own FIDO2 token, I should be allowed to.
If I get your FIDO2 token and reprogram it without somehow also wiping the data on it, your problem is that I got your FIDO2 token, not that I could reprogram it without erasing it (which theoretically could perhaps be true right now)
I'm guessing you don't understand the reason I don't want it to be reprogrammable. Yes, there are some advantages to me being able to reprogram it. But it comes with two big downsides.
The first is if I can reprogram it, then so can anyone else. I don't know what the situation is where you live, but government has passed laws allowing them to compel all manufacturers of reprogrammable devices to all them to reprogram is with their spyware.
The second is places I interact with, like banks, insist on having guarantees on the devices I use to authenticate myself. Devices like a credit card. "I promise to never reprogram this card so it debits someone else's account" simply won't fly with them.
The easy way out of that is to ensure the entity who can reprogram it has a lot of skin in the game and deep pockets. This is why they trust a locked pixel running Google signed android to store your cards. But take the same phone running a near identical OS, but on unlocked hardware so you reprogram it, and they won't let you store cards.
But that's the easy way out. It still let's a government force Google to install spyware, so it's not the most secure way. One way to make it secure is to insist no one can reprogram it. That's what a credit card does.
In any case, if someone successfully got the law changed in the way the OP suggested, so people could not use their devices as a digital passport, it won't only be me wishing a pox on their house.
1. if your government decides google has to put spyware on your phone, you wont be able to remove it, unless your device is reprogramnable.
It's actually the other way around, the only way to garantue that your device is free of spyware is you reprogramming it. You shouldn't have to trust the potentially compromised manufacturer.
True, but it's turtles all the way down. There is lots of non-reprogramable firmware in what you call "hardware". The recent article here pointed out the 8087 (an old floating point co-processor) had so much firmware (for the time) Intel had to use a special type of transistor to make it fit. Modern CPU's have many such tiny CPU's doing little jobs here and there. I'm being you didn't even know they exist. They not only exist, they also have a firmware programmed into ROM's you can never change. The bottom line is you have to trust the manufacturer of the silicon, and that isn't much different to trusting someone else who loaded firmware into the device.
The fact that there is always something you must trust in a device, as opposed to being able to prove it's trustworthy to yourself by just looking at it is so well known it has a name: is called the root of trust.
The interesting thing is it can ensure the root of trust the only thing you need to trust. The ability to do that makes your statement factually wrong. In fact it's drop dead simple. The root of trust only need let you read all firmware you loaded back, so you can verify it is what you would have loaded yourself. TPM's and secure boot are built around doing just that. Secure boot is how the banks and whoever else know you are running a copy of Android produced by Google.
Hey pabs, think about it. You know this doesn't work.
It doesn't work for the same reason the electricity company doesn't let you reprogram your electricity meter. Unlike the raucous response here as far as I far as I can tell, no one complains about that arrangement, despite the fact the meter is on your property, on land you own, and you effectively pay for it. They put up with it because of want the electricity, they know the electricity can't trust all their customers with metering it, and when it's all said and done putting a small box on their property the electricity has absolute control over is hardly a big deal.
It's exactly the same deal with your computer, or should be. There is a little area on a device you own that you have no control over. Ideally visible and running open source software with reproducible builds, so you can verify it does what it says on the box, and yes neither you nor anyone else can change it, so it meets your condition.
But it's purpose doesn't. It's purpose is to load the equivalent of electricity meters, which are software other people can change and you can't. Thus this area on the your device carves out others areas it can give ironclad guarantees to a third party they solely control, you can not reprogram, and you can't even see the secrets they store there (like encryption keys). These areas don't meet your definition. The third party can reprogram them, but you can't, you can't even see into them.
These areas can do things like behave like a credit cards, be a phones eSim, house a FIDO2 key that some their party attests is only ever stored securely.
Currently we depend on the likes of Google and Apple to provide us with this. I'm not sure Apple can be said to provide it, as they insist on vetting everything you can run that doesn't live in a browser. Google does better because you can side load, if you are willing to jump through hoops must people can't. Wouldn't it be great if debian could do it too? But to pull that off, debian developers would have to be believe allowing users to hand over control of a space on their computer they can't see or alter, to a third party debian didn't trust somehow works open source. It's not a big jump from the current firmware policy.
I can see that some verification is necessary. However i still think stuff that I can't be reprogramm should be heavily regulated. I want it to be kept at minimum.
Samsung already installs very suspicious auto updating, can't be removed without root, apps and ads. This is the natural consequence of locking out the users capabilities. If you want to get rid of them completely, youd have to root it, breaking compatibility with banking apps. Thats the world you are rooting for.
> but government has passed laws allowing them to compel all manufacturers of reprogrammable devices to all them to reprogram is with their spyware.
In this case the government may mandate to have spyware pre-installed in the factory - which is already the case for phones and laptops in some countries.
> I promise to never reprogram this card so it debits someone else's account
When reprogramming, the card should wipe private keys so it becomes just a "blank" without any useful information.
That doesn't work for two reasons. Firstly the law in my country specifically forbids introducing what they call a "systemic weakness". Among other things, that bans them from demanding every device is bugged. Instead they must get an judge to authorise targeting an individual, then get the manufacturers to replace the firmware in that device.
Secondly, they have no control over companies not based where I live. So I could just import it myself, provided you are successful get ever country to pass a law the denies me the right to do this the way I want to do it.
> The second is places I interact with, like banks, insist on having guarantees on the devices I use to authenticate myself. Devices like a credit card. "I promise to never reprogram this card so it debits someone else's account" simply won't fly with them.
If that's the only option they have, it will fly. Just like you used to be able to use banking apps with any Android before they had the option to restrict that to only Google-controlled ones.
Fair enough. Sort of. You can get the same assurances OTP gives you using secure boot + open source + reproducible builds.
Regardless the rest us who don't want to go through the extra work OTP creates still of use want to put our credit cards, fido2 keys, government licences, concert tickets and whatever else in one general purpose computing device so we don't have to carry lots of little auth devices. To do pull that off securely this device must have firmware I can not change.
The OP wants to make it illegal to sell a device with firmware I can not change.
In asking for that, they've demonstrated they don't have a clue how secure and opening computing works. If they somehow got it implemented it would be a security disaster for them and everybody else.
> I feel like everyone in this thread is assuming this is a good faith move by Australia to help kids in school and with socialization.
Most Australian schools banned phones a while ago. Attempts were made to measure the outcome. For example, South Australia saw a 72% drop in phone-related issues and 80.5% fall in social media problems in early 2025 compared to 2023 [0]. Other states reported similar results. These early figures are a little rubbery, but overall look very good. The social media ban is in part a response to that success.
The only major concern I have is de-anonymization of the web. It's worse than just de-anonymization. They've opened the gate for organisations like Facebook to demand government ID, like say a photo of a drivers licence. It contains a whole pile of info these data vultures would like to get their hands on, like your actual date of birth and residential address.
The sad bit is I doubt de-anonymization was goal, in fact I doubt they put much thought into that aspect of all. If it was the goal there far more effective ways of going about given the corporations permission to "collect whatever data you need to make it work". They could have implemented a zero knowledge proof of age service. But given the track record of their other computer projects, a realistic assessment is it had near zero chance of being implemented at all, let alone on time and on budget.
But if they had of insisted the providers implemented some sort of ZKP themselves, I would have found it hard to argue against given the past experience in schools.
> School behaviour improving after mobile phone ban and vaping reforms
Vaping !?
If we're discussing effect of phone bans at school, I think looking at a period where nicotine addiction was also strongly reduced makes the numbers pretty hard to interpret.
Is that the same report that failed to mention they changed the testing methodology for the year after the phone ban, and that an improvement was expected in SA test scores regardless?
I've owned a few XPS/X1 high end laptops in my time. Every single one of them had serious design flaws, as in "I'm not buying that again" type design flaw. That's true of every vendor.
I was sucked in by the advertising I guess. They looked very good on paper - good battery life for the time, thin, light, powerful, sleek, latest everything. I've built computer systems for most of my professional career. Looking back on it, how I could have possibly thought some fresh shiny new design first off the production line was going to be rock solid work horse is beyond me. Lack of critical thinking skills I guess.
Now, I buy something like a Dell Latitude. It's an enterprise machine. Translation: a plain, boring design with parts that have been trialed by the XPS/X1 suckers, so most of the bugs are ironed out. Enterprise tends to mean expensive. But they lose 75% of their value in 2 years, so second hand prices are very reasonable, and since Dell offers 5 years warranty on them they can effectively come with the same guarantees as a new one.
Enterprise also means well supported. It's almost night and day. Ring Dell about a Inspiron or even an XPS issue, and you are met with a wall of excuses. Contact them about an Latitude issue, you get a fast response. The one time I wasn't happy with the outcome, I said so in their "how did we go" questionnaire, and they rang me back begging me to let them have another go.
The HT68F001 stands out: 512 words of program flash and 16 bytes of RAM. ... can only be clocked from an internal 32 kHz oscillator. Since each instruction takes 4 cycles to execute, this results in only 8000 instructions per second!
reply