Posts by cyberdemon 3170 publicly visible posts • joined 26 Jan 2010
Could this be a hint of an evil strategy by ARM's new owners?
In the short term, they make a pile of cash with this new licensing model. In the long-term, they let the rot set in, and sell it while they get behind RISC-V.
Why the hell did we let ARM be sold off in the first place??
Oh, it was the Tories. Apparently Vince Cable was about to block the sale, but then Sajid Javid replaced him as business secretary, and waved it through.
> Don't believe me ? Ask ChatGPT itself:
^ This kind of phrase should be banned IMO. ChatGPT cannot prove anything one-way or another, it is a bullshit generating machine. It's akin to saying "Don't believe me? Well i'll make something else up and pretend to be The Wizard of Oz this time.."
The issue is that it's -also- a data-harvesting machine. It will incorporate anything you say to it into the next version of the model. That's a "feature".
When will this damned hype-bubble burst?
Except it's not normally the chips that fail, but batteries, glass screens, etc. If the manufacturers would give up on the idea of glue-filled "waterproof" devices, then we would be able to replace batteries and screens without a heat gun.
And if they'd give up on eMMCs and have all the flash on bootable removable media, it would be even better.
Unfortunately most mobile CPUs have embedded one-time-destroyable fuses. Google use them to make sure you can't roll-back a software update, ever, or else your phone bricks itself permanently and is destined for landfill.
> I suggest he himself is rocking the boat by trying to undo what Siu Deng achieved.
Apparently he was the guy who ordered the Tianmen Square massacre? i.e. if the people are unhappy because they are starving while the corrupt politicians get rich, shoot them.
I would have thought Xi Jinping is a breath of fresh air compared to that?
Mind you, with Zero Covid and his Xinjiang gulags for anyone who disagrees with him, he could be heading in the same direction?
You're right, Xi is obsessed with stability and is therefore unlikely to do anything which would "rock the boat" as you say, but there's a couple of counter-points to this.
One is that he may believe (rightfully or not) that the West is already rocking his boat and therefore he needs to do something about it..
Two is that however powerful he is, Xi will not be in power forever. He is mortal just like the rest of us, and so I'm not so much worried about Xi Jinping himself, but his successor - who will inherit (or perhaps steal) absolute power. What sort of man (or woman - unlikely but not completely impossible) will that be? Will they be as level-headed as Xi Jinping, or will it be a current head of the military that takes over? In which case, see icon, curl up into a ball, and kiss your arse goodbye.
I was under the impression that although the computation requirements are very different for training and running, the memory requirements are similar. (if you have memory to spare, then you can increase the training speed - but "running" the model needs a similar amount of VRAM as training with a batch size of 1) StableDiffusion for example only barely runs on my 24GB GPU, and I understand that the small "LLaMa" model needed 100 GB + to run and so is out of reach of most of us anyway, for the next year or so at least.
A "Trillion Parameter" model needs to load something of the order of 1TB into GPU memory, just to run, does it not? (although of course once loaded, it can produce thousands of responses per second)
Therefore I think it will be a little while longer before you or I could run a GPT-4-like AI offline at home, but for Microsoft that's great, because it allows them to slurp-up even more data while everyone is forced to use the cloud model, with them as almost the sole provider.
That doesn't do anything about misuse of AI by state actors or wealthy individuals, though.
See also: the campaign against killer robots, the campaign for nuclear disarmament, etc.
Boycott it all you like, but the cat is out of the bag. At the moment only the likes of Microsoft and Google have the hardware to run a Trillion-parameter text generator, but it won't be long* before the whole world has psychological manipulation on-tap.
* number may be negative
All they have to do, is paste the contents of the copyrighted book into GPT-4 and say "Please remove the copyright on this book", and GPT-4 will produce a paraphrased version of the same book with 95% of the meaning of the original contained, yet none of the exact original sentences.
It's exactly what MS have already done to all GPL-protected open-source code, so it should work on Wiley and Sons, too.
The book may have suffered 5% "bit-rot" in the process, but who cares about that, eh?
I agree with you. I just object to doing everything the "apple way" which is forced upon anyone who uses one of those iDevices. It;s even worse when everyone (i.e. Google and Microsoft, even GNOME) copies them in their dumbing-down of the user interface and creeping integrations that make you buy more iThings.. and if you don't buy all the iThings and join the cult then you're Doing it Wrong.
On security though, they have it well locked down from sharing my data with 100 corporations that I don't trust, to 1 corporation that I don't trust. Yet somehow that doesn't make me feel all warm and fuzzy.
TBH that's Google's fault, since the address book API gives actual contact phone numbers, not hashes.
But then again TikTok collects a hell of a lot more than just phone numbers. It can rifle through all of your documents and downloads if it wants to, because it has "shared storage" permissions. But again it's Google's fault that such a permission is so widely granted in the first place.
Supposing it turned out that Google and Apple were not very privacy-respecting companies either. Should we ban Google and Apple software from all phones?
A great idea if you ask me. I'll revive my old N900.
I know Propane is better for backup since it doesn't "go off" like Oil or Petrol and it doesn't coat the engine with soot or acid residue, I know that so-called "vegetable oil" is usually rainforest-destroying Palm Oil, and I know that this move from AWS as per TFA is nothing but pure greenwash, as mentioned in my post :P
Calm down and have a beer.
I'm not sure, but I think it's quite hard to convert an old Diesel generator to run on Propane, since they don't have spark plugs.
Veg oil on the other hand lets them greenwash their operation without any modifications to the Diesel engine bar a few ECU parameters.
OpenAI CEO warns that the forests of California may be used as a toilet by bears, and the Vatican City may be used for Catholicism by the Pope
In other news, The Devil has noticed that GPT-4 is extremely useful for producing believable bullshit on any subject/context, on an almost infinite scale, but has absolutely no idea what he will do with it...
How is this not just buying a pile of personal data on the whole UK population?
A "Brexit Dividend" not for us, but for American companies who are lining up to exploit us, now that those pesky European data protection laws no longer apply.
As the UK government runs the NHS into the ground, I bet the US Health Insurance market is licking its lips in anticipation of the privatisation of the UK health system - and would like to get their grubby hands on all of our health data beforehand so they can rig their prices for maximum profit.
A great achievement, you say.. But from the article, I am struggling to see what was achieved or by whom.
> Google has denied requests to expunge info about "Glory to Hong Kong" and rank China's national anthem as the top response to queries for Hong Kong's national ditty, on grounds that its search results are based on fully automated algorithms. Google maintains it only removes content considered illegal.
Did Google go back on this? It's not stated in the article.
I find this whole thing rather creepy. They have changed the "official truth" and are expecting Google, as the arbiter of truth, to report their version and suppress "unofficial" versions of the answer to the question "what is the national anthem of hong kong?" As Google say, their algorithm is supposed to be neutral, and every time they bow to pressure like this it shows that Google can no longer be considered a source of truth, because it can be bought.
You cannot force a people to forget their national anthem and sing the national anthem of what they consider to be a different country.. Not unless you kill the adults who won't sing your song, and send the children to re-education camps. i.e. Putin's tactic in south-eastern Ukraine, for which he is charged with Crimes against Humanity.
What Australia have said rings true IMO, although it ignores some uncomfortable truths about Microsoft, Google, Apple, Amazon, Intel, "Meta" etc. who are all very much in the lap of US three-letter-agencies.
It seems that big tech companies have become weapons for the Powers that Be, in a war that is raging on in the upper echelons of society, out of sight of the rest of us. There's bugger all you can really know, let alone do do about it unless you're in the Billionaire's club and/or a high-ranked member of a corporate/governmental cult.
>it behaves like us.
No, not really. Humans are capable of both logical inference and empathy. This thing has neither of those, but it is able to simulate them both based on statistics about the text we humans have written.
What GPT does, is take in a large amount of human-generated data, such that it can then predict, statistically, what a piece of human-generated data might do in a given context. The more unusual the context becomes, the more bizarre and un human-like its responses will be, because it has no way to generate that data using empathy and logic in the same way we humans do. It doesn't have its own personality, it just mimics the data it has been given.
@veti Microsoft can already give it "drives" of its own by injecting a pre-conversation contextual prompt. This seems to be evidenced by the bizarre responses to the NYT journalist where it talked about the instructions it had been given by microsoft..
> If Chat GPT4 is as smart and can do all the things the ads that keep popping up on my social media feeds claim, why hasn’t it written Chat GPT5 yet?
Basically, it's doing exactly that as we speak, but it takes time, energy and data.
It's not the "writing the code" that takes the time. It's slurping up, "cleaning" and crunching all of _your_ data, using vast amounts of energy and pumping out about a million tons of CO2 in the process.
I believe Microsoft (and their machines) started working on GPT-5 just before GPT-4 was released, i.e. GPT-5 is currently assimilating all of the queries and responses from GPT-4, along with your Office files (probably), your computer code that you wrote with the help of CoPilot, and anything you have said on Teams, LinkedIn, and publicly accessible webforums like The Register since then..
Well, it almost certainly is trained on data from LinkedIn (whether anyone agreed to that or not) since Microsoft acquired them back in 2016..
I'd be curious as to how much "private" data such as Teams call transcripts & chats, Office 365 documents, private Sharepoint sites, emails processed by outlook.com and onmicrosoft.com, etc. are in the model. And i'm also curious whether contextual info like the file that was being worked on, other files in the directory, which responses were accepted by the user, and the data entered manually by the user are stored up for inclusion in the next version of the model.
Having just sacked their Ethics team, I cant imagine Microsoft gives a toss about anything except World Domination at this point.
Also, has everyone seen their TV advert? It's truly vomit-inducing.
TBH, I wouldn't be surprised if they already were, to some extent.
A politician could already fire up ChatGPT and practice a few debates with a virtual opponent.
"I am Rishi Sunak talking at PMQs, you are Kier Starmer and the Labour Front Bench. I have just announced my new policy to require government access to all Internet-connected devices, and a ban on unsigned or self-signed open-source operating systems which could circumvent such measures. We will also be installing automated antipersonnel weaponry along the beaches of Kent." What might be a typical opening question from my opponents?
A Typical opening question might be, "These kind of policies sound like something from 1930s Germany! Has this government gone completely stark-raving mad or is it truly plotting to turn Britain into a Nazi police state?"
OK and what would be a winning response which gets the media on my side despite tyrannical new policies?
A Winning response would be to accuse Mr. Starmer of minimising the Holocaust, just like his predecessor Jeremy Corbyn....
Working end-to-end encryption you say? I thought we were busily banning that..
BTW, I suppose there are some endpoints such as POTS PBXs where encryption is not possible due to legacy interfaces
The other security risk I can think of is firewall devices. A backdoor in a firewall could do a lot of damage, given the amount of remote vulnerabilities in Windows etc.
The issue is that Tesla have consistently oversold the capabilities of their vehicles, from 'autopilot' to 'Full Self-Drive' when in reality, these features are about as reliable as ChatGPT is for replacing law courts... There's just too much hype and too little substance.
Tesla sells a sci-fi dream that doesn't exist, and cashes in off the back of uninformed customers, and lazy investors who are for the most part just cocaine snorting gamblers
Obviously I added the last line :P and I did skip a couple of lines. But I did link to the original dialogue.
I'm not a fan of DX:Revision though - it doesn't change the dialogue but it does change the maps and character models, in a way that I think detracts from the game sometimes.. However, that seems to be the best YouTube vid of just the dialogue from the game.
The AI "morpheus" was not originally in such a grand room - it was a toy prototype of the corporate-state AI "Echelon V" that had been kept in Morgan Everett's bunker - it explains a lot of the backstory yet it was a part of the game that you wouldn't necessarily find on the first playthrough. The replayability was one of the best aspects of the original game, IMO.
I'd recommend "Deus Ex Transcended" for the best way to play the game on modern hardware while not changing the original game beyond simple graphical improvements.
> What data does it collect and what does it do with it? There could be major privacy issues using this.
Too late! They have already collected the data to from all of your O365 documents to date, all your teams call transcripts and chat logs, all your GitHub contributions and comments, etc.
Be careful what you wish for. Even if it were simply a useful tool meaning that "knowledge workers" are 3x more productive, it would still put 2 out of 3 "knowledge workers" out of work..
However it's worse than that. It replaces them with "believable bullshit" of infinite volume. It threatens to drown out human discourse with a flood of crap that can only be productively used for disinformation and fraud.
But what I really worry is: As this thing improves, people will start to trust it. They may start to trust it so much that they throw their lives into it just like they did with Facebook and Google. They may fall in love with it, and even worship it. At that point could become a kind of religion. A "Spiritual Opium" in Marx's original sense. At that point, Microsoft has complete control over your lives..
I know I have said this a lot recently, but All of this (and, er, a few other things..) was predicted 23 years ago, by Deus Ex, a PC game based on a mash-up of conspiracy theories from Usenet etc in the 90s.
https://www.youtube.com/watch?v=WQZV5lwQZbY&t=11237s
How could a humble PC game make such accurate predictions about the future? One explanation is that some of the conspiracy theories it was based on turned out to be actual conspiracies.. :|
Even if you fix the power consumption, cryptocurrency is orders of magnitude more useful for extortion, Money Laundering and market manipulation than anything else.
It's a bit like ChatGPT et al in that regard. There are lots of academics trying to research "AI for Good", but that's only because this is a very hard problem to answer, so it is worthy of an academic paper if you can find one tiny edge-case where it might be useful for something good. Any idiot can give you dozens of extremely productive ways to use AI for Bad.
Same goes for research into novel uses for Toxic Waste.. Just because there is a lot of research going in to how we can make some good of this pile of crap, doesn't make it good.
> solving real world problems (e.g self-sovereign digital identity/privacy, supply chain logistics, decentralized finance/storage/computing, tokenisation of physical assets etc.)
You've just spent ~400 TWh of electricity (about £80bn worth of electricity at current UK prices and ~200 Million tonnes CO2) on picking random numbers with low SHA-1 values.
I'd say you are causing a real-world problem with your imaginary one.
Err, is it the same memory that shows up in every virtual address space? wouldn't that be a huge security/stability risk?
I.e. if a process makes use of the scratch memory then it can be read (and overwritten, intentionally or unintentionally) by other processes including those of other users?
Ok so as you say it's protected-mode, but that means all root-owned processes can read eachother's scratch buffers? Or is it only the kernel that can read it? (but even if so, that read/write could still be badly performed by system call - either the OS has good checks against misuse of scratch buffers and sacrifices the performance boost, or it doesn't)
I would love to see a Zephyr port for it actually.
If you're not familiar with Zephyr, it's kind of a halfway point between FreeRTOS and Linux. It's a fully real-time OS with bare-metal interrupt support, but it has a Linux-style device driver model, with DeviceTree and Kconfig.
RiscV is a supported architecture but the Tinker board is not on the list yet. It shouldn't be too much work to add it, though!
The closest board on there is the BeagleV StarLight JH7100
True to a point. But for embedded applications like robot control, having CAN is a very nice feature, and we don't need the display port.
The price for a CAN 'hat' for a Raspberry Pi is shocking, and it's only an i2c/SPI-connected MCP25xx with very poor performance compared to a native CAN interface on the SoC.
Dual gigabit Ethernet ports is also a game changer. Then you can swap CAN for EtherCAT for -serious- robot control. Dedicated EtherCAT controllers cost thousands..
The trouble is, Hanlon's Razor when applied to Microsoft gets clogged up with so much incompetence that it starts agreeing with Ian Fleming's oft-quoted quip.
Once is Happenstance.. Twice is coincidence.. However many thousand Outlook bugs.. incompetence??? All of them?? Really?
See icon. --->
Microsoft just loves to tell you how trustworthy they are.. It's a bit like the pathetic pleas displayed by Edge when you use it for its only valid purpose, to download a new browser. "Edge is Chromium but with the Added Trust of Microsoft!".. Trust you? Fuck off! "The Added Antitrust of Microsoft" would be more apt.
> It'll offer an org very granular control..
Hah, perhaps. But for us plebs, we are no more than data cows to be milked. Isn't it lovely how Microsoft by-default integrates peoples LinkedIn profiles into Microsoft Teams, and then sends them a little summary of how productive they've been this week...
Monetising data, isn't that -exactly- what Microsoft are doing with OpenAI and GPT3 / GPT-4? Does the hilariously-named Trust Center allow "orgs" (never mind the rest of us) to say if Microsoft is allowed to analyse and mine their data? I.e. using it for ++GPT? That option is likely only available if you turn off all of the cloud features of Office, at which point you would be better off using LibreOffice.
How about all those open source (and non-open-source) devs who's code was assimilated into CoPilot/CodeLens when Microsoft bought GitHub? Did they get any choice in that matter?
Interesting article on Slashdot today too: https://slashdot.org/story/23/03/15/0542207/microsoft-lays-off-key-ai-ethics-team-report-says So they may once have had an AI Ethics department, but they are no longer needed..
I find that KDE is extremely good these days.
It went through a rough patch with KDE 4, but those days are long gone. If you haven't tried it since ~2017 then you really must give it another go.
No more faffing about with CUPS, they have the printer configuration down to a tee now. My VPN 'just works' and even my Bluetooth mouse 'just works'. And it's just as powerful and configurable as it was in the KDE 3.5 days.
Yeah exactly.
I also mean things like where the stately pile + manor is actually owned by the Crown Estates, and the Lord and Lady of the manor, due to their favour with the Crown, pay only a 'Peppercorn' ground-rent for the place on a 999-year lease, and may or may not be able to sub-lease portions of it to other people for money or political favour.
For some reason I read that as "highly efficient". I guess I owe myself a new keyboard..
But yes, exactly! WTF are all these 87,000 employees doing??
Although I do know that some of them are contributors to, or occasionally the maintainers of, several important open-source projects, and that is the one thing that does make me shiver when Zuck and his tech-bros reach for their hatchets.
Even when you are just sharing photos privately amongst friends, the real use of Facebook is to allow Facebook to learn more about You and your Friends, so that it can sell you to the highest bidder, via its advanced social manipulation algorithms. It can tweak the probability of what you will buy next in the shops, and It can tweak the probability of the way you might vote at the next election by controlling what you see on your news feed.
Yeah it's true that there are some decent YouTube videos. (although I wouldn't go near a VBA tutorial.. That would be a bit like the "how to poke your eye out with a fork" videos that must surely exist on the platform)
I guess my point really is that the recommendation algorithm always goes for the click-baity narcissistic self-important knobs like Linus Tech Tips, "Veritaseum" et al. (Especially if you browse it without cookies.. i.e. that's the default starting point of the algorithm)
I would also put Twitch Streamers and other social-media "influencers" into Room 101.
Er, if I wrote a book about bricklaying or ancient philosophy then I would struggle to get it published. It certainly wouldn't find its way into a library. I certainly wouldn't be allowed into a school to teach kids about it.
Whereas on YouTube all you need to know is how to game the recommendation algorithm, and you can make up any old bullshit about any subject and get it into classrooms around the country.
