AGI is no longer a big buzzword in Silicon Valley. The unfulfilled expectations will affect the future of artificial intelligence.

Like the idea of a true driverless car or nuclear fusion, artificial general intelligence is not going to happen soon, despite hype in recent years suggesting it is not far away.

Most modern artificial intelligence is a form of deep learning or natural language processing. AI is also used for computer vision. Things have moved fast in the last decade, but it is more about sifting through vast piles of data than building machines that think.

General Artificial Intelligence is about building machines, or more accurately software, that can think in ways that resemble human thinking.

That’s a loose definition, experts argue over exactly what GAI means and doesn’t mean. They don’t even agree on its name: there are many variations in use each with a different, nuanced meaning.

GAI doesn’t necessarily mean sentience. At best it is an early step on the path, but it is possible machine sentience can never be achieved.

As The Next Web reports the financiers behind GAI have gone quiet lately. The story suggests they have lost their appetite for the technology.

Source: The AGI hype train is running out of steam

Lisa Fong (National Cyber Security Centre)
Lisa Fong (National Cyber Security Centre)

Criminal cyber attacks targeting Aotearoa New Zealand skyrocketed in numbers over the last year and continued to grow in sophistication.

At Reseller News Rob O’Neill writes: National Cyber Security Centre reports a surge in criminal cyber attacks.

While the proportion of state-linked malicious cyber activity was down slightly from last year’s 30 per cent, this was because of the greater proportion of criminal incidents recorded.

The report showed there were 404 incidents affecting nationally significant organisations in the 2020/21 year, a 15 per cent increase on last year.

These numbers reflect the NCSC’s focus is on incidents affecting New Zealand’s nationally significant organisations, and on incidents likely to have a national impact, which means the numbers represent just a small proportion of the total incidents affecting New Zealand

This squares with anecdotal evidence from New Zealand businesses that they are now constantly under attack.

When it comes to government and large scale business systems, it is likely the attackers are already inside the systems waiting for an opportunity.

Everyday crime rates are dropping in most rich countries like New Zealand. There’s a clear switch from activities such risking your life with weapons to rob a a physical bank and getting online to steal money. Computer fraud is on the rise everywhere.

Technology doesn’t help. Bitcoin, a cryptocurrency, may not have been invented to smooth the way for criminals, but it is used by the underworld to move money around. Drug gangs are carrying fewer suitcases full of banknotes and dealing with more crypto transactions. Encrypted messaging services are used to communicate.

While these tools have legitimate uses, criminals have embraced them and depend more on them.

Ransomware remains the biggest threat. Criminals lock up data or disrupt systems until victims pay them, almost always the transaction is in Bitcoin.

At first ransomware gangs targeted small business. It turns out that was all about learning their trade. Today they target government departments, a DHB in New Zealand, police departments overseas.

They operate on an industrial scale and there are well established digital underworld supply chains.

We know most of the gangs are based in a small number of countries. Officials don’t like to talk about this because of diplomatic niceties. As a journalist I can tell you that Russia, other parts of eastern Europe and China are the main sources. We also know some states turn a blind eye to the activity so long as the gangs focus on foreigners. There’s evidence criminal gangs and state hackers co-operate.

Governments have been slow to focus on fighting cybercrime. We can expect that change, but don’t expect a let up from the gangs.

Last year the world created or replicated 64.2 zettabytes of data. The number comes from IDC, a market research firm.

The figure is remarkable considering three years earlier IDC was forecasting the 2020 number would be 44 zettabytes.

A zettabyte is a trillion gigabytes.

In part IDC puts the faster growth down to the Covid-19 pandemic: a “…dramatic increase in the number of people working, learning, and entertaining themselves from home.”

IDC says: “…less than 2 percent of this new data was saved and retained into 2021 – the rest was either ephemeral (created or replicated primarily for the purpose of consumption) or temporarily cached and subsequently overwritten with newer data.”

Between now and 2025 the amount of data is set to grow at a compound annual rate of 23 percent.

The fastest growing source of data is the Internet of Things, not including surveillance video cameras. Social media is the second fastest growing source.

IDC says the amount of data generated is growing faster than our capacity to store data. The world had around 6.7 ZB of storage and that is growing at 19.2 percent year on year.

Which means we save less and less of the generated data.

This is less of a problem than it might appear because a large fraction of data is useless. A decade ago experts found as much as 90 percent of stored data was rubbish. It can include empty files, duplicates… or many multiple copies of identical files and temporary files that were never deleted.

Science fiction writers have worried about the idea of artificial intelligence getting out of control for years.

Now scientists at the Max Plank Institute in Germany are concerned.

A group of artificial intelligence researchers have published a paper that says once AI starts working at levels beyond the scope of today’s programmers, it won’t be possible to set limits on what it does.

This may seem like an idea that is far off, but there are already AI systems in use today that are making decisions and running things that the programmers who built the systems don’t fully understand.

Part of the problem is that while we could limit the capabilities of AI systems to make sure their capability doesn’t snowball out of control, the whole point of AI is to solve problems that are beyond us using conventional techniques.

I should point out these researchers don’t have any comforting suggestions.

David Leacraft found his Canon Pixma MG6320 printer would refuse to scan or fax documents if there was no ink in the device’s cartridges.

So unhappy he filed a class action lawsuit against Canon claiming “deceptive marketing and unjust enrichment“.

The point here is that you don’t need ink to scan or fax documents. So there is no reason the device should not perform those tasks if it is out of ink.

Canon advertises the device as an All-in-One printer that does printing, scanning and faxing. At no point does it warn potential customers the features don’t work if there is no ink.

There are 20 Canon printers that behave similarly.

It’s yet another example of customer hostile practices from the companies that make consumer printers.

To refresh your memory printer makers charge ridiculously high prices for ink refills while scaring customers with dubious stories about what happens when you buy third party cartridges.

And they are so keen to watch what you do with your printer that some devices only work if you punch a hole in your cyber security.

Printers are a problem in other ways.

Leacraft and the 100 other Canon customers who have joined the class action are looking for at least $5 million in compensation. The action has yet to be approved by the New York Court.

Overpriced home printer ink is annoying. On its own it doesn’t pose a security risk.

We can’t say the same about the technology printer makers use to keep paying too much for their overpriced ink.

There are home office printer models that stop working if they are not connected to the printer maker.

Connection risk

That’s risky on two fronts.

First, it means you can’t print if you lose the connection.

There are many ways the link can fail. It could be your local wi-fi network, your internet connection or the submarine cable connecting your country to the printer maker’s servers.

All these need to work for you to print a page on your home printer.

To be fair, connections don’t fail often. But the failure rate is not zero.

Dependence

Another way you might lose the connection is if the printer maker’s servers stop working.

Given that the cloud giants all experience downtime, it’s possible your printer maker might be offline when you need to print something in a hurry.

The outages may not be long, but it is ridiculous that your ability to print at home depends on the conditions in a remote server on another continent.

To big to fail

In a similar vein, your printing days could be over if the printer manufacturer goes out of business.

Admittedly that’s not a huge risk, but, again, it is not a zero risk.

Add all the risks together and you realise you have to put a lot of faith in things you have little control over just to get a page out of your printer.

Security alert

A bigger, more worrying, risk is your security.

An internet connection going to your printer potentially punches another hole in your cyber defences.

Connecting printers to the internet isn’t new. It’s been possible to remote print on your home inkjet from anywhere in the world for years.

Modern devices can have embedded servers. They are, in effect, computers in their own right. Again, this is not new.

They perform tasks like installing new drivers and telling printer makers you are using third party ink.

Vulnerable all the time

The difference now is today’s printer servers have to be on all the time. If you block the connections you can’t print.

Chances are the server on your home printer is one of the weakest links in your security chain.

These servers are rarely protected with more than a password. Sometimes not even that.

Patching

You may be careful when it comes to updating your computer, phone and apps. Keeping a printer patched is harder work.

Apart from anything else, it can require manual intervention. Automatic software update options are rare.

The controls can have minimal, hard to understand interfaces. There are plenty of opportunities for things to go wrong.

In the past there have been attacks where printers are used to remotely print messages. That’s not serious, but it illustrates the vulnerability.

Compromised

The main problem is that a compromised printer can open the door to everything on your home network. It can be taken over and used to snoop for data or mount external attacks elsewhere. Your printer could become part of a botnet.

Until now printer-to-internet connections have been, up to a point, optional. You could almost always print out pages without needing a live internet link.

Looking at the bigger picture, adding an extra connection back to an account with the printer maker is a small additional security risk. But we live at a time when the idea is to eliminate security risks, not add fresh ones where there is no benefit.

It’s yet another reason to keep the printer turned off and to work on weaning ourselves off printing.

Footnote: Matt East points out that turning the printer off isn’t a great idea. He is right, but that’s where the printer makers have taken us: to a point where you have to make tradeoffs that should not be necessary. 

After a couple of weeks using the beta and a week with the final version of Windows 11, I’ve yet to find a real reason to use it.

Steven Vaughan-Nichols nails the problem with Windows 11 in at Computerworld. 

For many people it is, he notes, “a pointless upgrade”.

That’s the conclusion I reached.

The main justification for moving to Windows 11 is that it will be more secure than Windows 10. To get those benefits you need to have the right hardware.

Windows 11 is picky about hardware. Most versions of Windows have been able to run on computers that are more than a couple of months old.

That’s not the case with Windows 11.

Are you ready to buy a new computer?

For many people reading this, that means buying a new computer.

And anyway, you can get the security updates if you stick with Windows 10.

Which, as the man says, makes the move to Windows 11 pointless.

At least for now. If you want to stay with Windows, you’ll get it with your next hardware upgrade.

You have to ask yourself why Microsoft is moving to Windows 11.

Last version of Windows

When Windows 10 came along the message was this could be the last ever version of Windows. From that point on the idea was that there would be regular incremental upgrades rather than big leaps.

“Last ever version” lasted six years.

In comparison, Windows XP lasted eight years.  Well, five years if you don’t count Windows Vista. Even Microsoft would prefer to see Vista written out of the history books.

Aside from the security benefits, Windows 11’s other selling point is a fresh new look. This is little more than cosmetics. A lick of paint and a brush-up. If anything it now looks more like MacOS.

Some of the changes appear to be change for change’s sake rather than researched improvements. There are background performance changes that users might experience without noticing them.

There is a promise that Windows 11 will run Android apps. That’s unlikely to happen for another year and, unless you have something important you do only on Android, is less interesting than it sounds.

Options

None of this is to say Windows 11 follows the tradition that says every second version of the operating system is embarrassing. It’s usable, popular and up to a point familiar to the majority of users.

On a personal note I was so disappointed with Windows 8 that I investigated, then moved from Windows to Mac. In hindsight it was a smart move, my productivity soared.

This time around Windows users have other options to tempt them away from the mainstream. Desktop Linux is mature and well worth investigating.

If that’s not for you, there are Chromebooks. An iPad Pro can do most things you buy a PC for. You may fancy a change without moving too far from Microsoft’s orbit. A Windows 365 Cloud PC is an option.

Yet I suspect most Windows users will choose to stick with 10 for now and see which way things go. There is no pressing reason to make a decision today. Most enterprise IT departments will wait at least 18 months before changing, you don’t need to take that long, nor do you need to hurry.