Weird Science: King Troll, Slime Bots And Goldmines.

Richard Bradbury: You never know what you’re going to hear when Matt Armitage starts hunting for Weird Science. It could be self-replicating robots, insects that create bio-fuels or planets made of diamonds. On the last edition he skipped the AI, which means he must have a backlog.

Richard Bradbury: You’re sounding a bit miserable today…

Matt Armitage:

• I don’t know if I’m miserable or confused.

• And yes – there’s a bit of a backlog on the AI stuff.

• But before we get there, and the reason for my sad and confused state is…

• The rather strange news this week that Elon Musk has taken a more than 9% share in the social media giant Twitter and has been invited to join its board.

• He is now the company’s largest single investor, although he seems to have pledged not to try and own more than 14.9% of the company.

• Or attempt to take it over.

• He will now have incredible influence over the company’s direction.

• I guess I’m miserable because it’s yet another billionaire buying an information platform.

• And confused because – well – in some ways Musk is Twitter’s largest remaining troll since Donald Trump was removed from the service.

• So now Musk is the Troll King of Twitter – which does sound like a character from a Terry Pratchett story.

• But somehow isn’t as funny.

Richard Bradbury: I think we should probably flesh out some of the details of this story.

Matt Armitage:

• Musk was uncharacteristically quiet about his attempts to amass a stake in the company.

• In a regulatory filing with the Securities and Exchange Commission, it became apparent that Musk had amassed a 9.2% stake in the company.

• Larger than the 8.8% owned by mutual fund Vanguard.

• And more than 3 times the stake owned by founder Jack Dorsey – who, incidentally, will be stepping down from the company’s board at the end of this year.

is for a term that expires in: 2024

• And it’s subsequently come out that he reached out to Twitter CEO Parag Agrawal during the process of buying the shares which led to the invitation to join the board.

• As I mentioned earlier, Musk has agreed not to try to buyout the company and has placed a ceiling on the percentage of shares he will buy.

Richard Bradbury: You mentioned that he was uncharacteristically quiet about his actions?

Matt Armitage:

• Yes. You know, this is the guy who tweets his idea for an underground traffic system because he’s stuck in a jam.

• There are truly two people to whom the phrase ‘I tweet therefore I am’ belongs.

• One has already been kicked off the platform, and the other is Elon Musk.

• He’s gotten himself into so much hot water over his tweets – comments about Tesla leading to sanctions from the SEC.

• Referring to one of the rescuers in the Thai cave incident a few years ago as ‘that pedo guy’.

• Seemed to compare Canada’s Justin Trudeau to Hitler.

• He seems to delight in picking fights and causing outrage and offence.

• Yet he was very quiet about his plans to build up his stake in the company.

• This is understandable – telegraphing it would send prices rising or stir other investors to increase or move to protect their stakes.

• But he seems to have gone to the other extreme, polling his followers at the end of March if they think Twitter rigorously adheres to free speech.

• Around 70% of them stated that they didn’t –

• Which is moot - Twitter, as a private company, isn’t obligated to rigorously adhere to free speech, anyway.

• But in the process, he seemed to hint at starting his own social media company – which is certainly at odds with him building a stake in Twitter, which reports I’ve seen suggest he began at the start of March.

Richard Bradbury: Can you imagine what a Musk controlled Twitter would be like?

Matt Armitage:

• I think in some ways he would continue to go in the direction that Twitter is planning.

• He’s tweeted about his desire to allow people to control or use their own algorithms for the platform.

• So that they see exactly what content they want and when it appears.

• Twitter is already working on ways to retool the platform to make it less about a centralized company and more about the users.

• The company is working on platform redesigns that will allow third party APIs to control the way content is viewed.

• And potentially to allow individuals to design and plug their own algorithms into the service.

• It seems that the common purpose in Musk and Twitter’s vision is one of the reasons

• Agrawal asked him to join the board and help to shape that transformation.

Richard Bradbury: Has he had any service requests?

Matt Armitage:

• That’s one of the funny aspects of the story, I guess.

• That he’s somehow turned into the service’s uber tech support guy.

• He’s had thousands of requests for an edit button.

• Which is genuinely something we all want.

• Especially on DMs. Number of times I’ve messed those up.

• That sparked Twitter to announce on Wednesday that they have indeed been working on an edit feature for some time, and that no,

• They’re not saying that because of a certain somebody’s poll.

• So that’s good news.

• What happens in terms of free speech, I’m not sure.

• Both Jack Dorsey and Parag Agrawal are also committed to making speech as free as possible on the platform.

• That’s a hard call – one person’s freedom of speech is another’s hate speech.

• And there are legal requirements around some forms of speech.

• That’s before we get to disinformation.

• Free speech is never as simple as it appears. Or as free as we think it is.

Richard Bradbury: Does he have the time to help run Twitter?

Matt Armitage:

• That’s possibly the bigger point. What will the investors backing his main gig – Tesla – have to say.

• The company has performed exceptionally well over the pandemic.

• But it’s facing a number of problems at the moment – the supply chain crunch which is limiting the availability of many car parts and chips.

• Plus, the Tesla factory in Shanghai is experiencing shutdowns due to the spread of the Omicron variant in the city.

• And Musk already annoyed stakeholders when he unloaded around 10% of his stake in the company last year.

• Am I the only one who pictures a crowd yelling kill the vampire whenever they hear the word stakeholders?

• Shareholders often take a dim view of founders taking on huge outside interests that they fear might be to the detriment of their own investment.

• And Tesla is facing enormous pressure from these external shocks right now.

• So it will be interesting to see how Tesla’s stakeholders react to the news over the next few weeks.

Richard Bradbury: What is it that disturbs you about this?

Matt Armitage:

• Because it’s yet another – media outlet isn’t right – but information platform that’s being bought by some super-rich guy.

• It’s like Jeff Bezos and the Washington Post. True he hasn’t interfered – as far as we’re aware – but is he a guardian of free speech?

• Look at how Amazon responds to efforts to unionize its workforce.

• Tesla has also had its issues with toxic workplace conditions – as has Twitter in its early years, I should mention.

• Whether it’s Bezos, Murdoch, now Musk. There seems to be this idea that when you’re rich you should own some kind of broadcast platform.

• It’s not new – Henry Ford is just one of numerous examples.

• The problem is that none of these men – and they always seem to be men – seems to be a decent human being.

• Musk thinks he’s an edgelord but often he’s just a troll: one of those Twitter power users who can use his account as a weapon.

• It’s like watching Back to the Future but it’s Biff Tannen and not Marty McFly that comes out the winner.

• I guess I don’t like it when the bullies win – it’s more than a little dispiriting.

Richard Bradbury: Cheery. I’m not sure where we go from there…

Matt Armitage:

• I do. This is sort of a similar story. This is something I didn’t have time to fit into our last WS.

• And in a way I’m glad I couldn’t.

• It was the news last month that US cinema chain AMC bought a stake in a company that owns gold and silver mines in Nevada.

• Today seems more weird business than weird science – but I guess that’s ok.

• It’s a business station.

• Incidentally there’s a longer piece on this story on the kulturpop substack.

• AMC became Internet famous for becoming a meme stock alongside companies like GameStop, where a mass of casual investors,

• Galvanized by forums like the sub-reddit wall street bets.

• Piled onto various companies, causing their share prices to rise, and upsetting institutional investors holding short positions on those companies.

• So it was part edgelord – twice in one show – and part, not manipulation, but coordinated movement creating these short term price gains in the stock.

Richard Bradbury: How was AMC doing at the time?

Matt Armitage:

to bankruptcy at the start of: 2021

• On top of the structural changes that the movie industry has experienced since the rise of streaming media.

• Investors pumped in around USD1bn to shore up the company and it has since leaned in to its meme stock status.

• At the end of: 2021

• But also its CEO has announced that as well as restructuring the company he would use some of its cash reserves to invest in other companies that had short term cash issues but looked good for the long term.

• In March AMC announced it was taking a 22% stake in mining company Hycroft Mining Holding Corporation.

• Market watchers announced the move as crazy. That AMC shouldn’t be looking so far from its core business.

• And that’s what the story would have been if we’d talked about it a couple of weeks ago.

Richard Bradbury: That a move like this is odd and possibly crazy?

Matt Armitage:

• Yes. You know – the comments were – and my own thinking was – that the company should be using money to pay down debts, reinvest in the theatres, etc etc.

• But the move seems to have turned Hycroft into its own meme stock.

• At the time we’re recording this, I think AMC’s stake is up something like USD60m.

• Which brings us back to Elon Musk.

• I was wrong about AMC – or at least it appears that way at the moment.

• Maybe I’m wrong to be pessimistic about Musk joining Twitter.

• Maybe it will be the best thing that could happen to the platform.

• And maybe it will stop Twitter being the Great Pacific Garbage Patch of the Interweb.

Richard Bradbury: When we come back – Artificial intelligence with an artistic flourish, magnetic slime and the ghost in the machine.

BREAK

Richard Bradbury: It’s weird science on Mattsplained today. So far it’s been weird but there hasn’t been much science. No more misery. It’s time to get our freak on.

Matt Armitage:

• OK. So let me introduce you to a friend of mine.

• INSERT CLIP: https://www.youtube.com/watch?v=XaZJG7jiRak&t=12s (0m38s to 1m08s)

• As she said, that was Ai-Da the world’s first AI-powered robot artist.

he delivered in Oxford, UK in: 2020

• She was created by a gallery owner - Aidan Miller, in association with robotics company Engineered Arts, AI boffins at Oxford University and engineers from the University of Leeds.

• And rather than just making her – and they are calling the machine her – a series of cameras and robot arms,

• They decided to make her as human-looking as possible.

• Which conversely makes her even more inhuman.

Richard Bradbury: Does her work sell?

Matt Armitage:

• Her initial show raised over UKP1m.

• One of the things that distinguish Ai-Da is that she’s programmed for imperfection.

• With the cameras and the processing and the arm she could effectively be a printer, creating photo realistic drawings.

• But these flaws allow her work to be unpredictable, and surprisingly much more pleasing.

• Otherwise, she would essentially be a very expensive printer.

• She also creates her own poetry – this is more than just a one-off robot.

• This is a proof of concept for machine-generated art and a critique on the ethics of the advances of AI.

• I’m mentioning the story today because AI-Da will be staging a solo exhibition of her work at the Vienna Biennale titled Leaping into the Metaverse from April 22nd.

• And those upgrades include the fact, that incredibly she now paints.

Richard Bradbury: That bigger question though: can a machine create art?

Matt Armitage:

• That’s the point of Ai-Da. To have those discussions.

• Ai-Da is constantly being refined. In her original form, she could hold a pencil and sketch and draw.

• Now she can paint. And with her cameras, she can create portraits on the spot, which will always be different, even if the subject is the same.

• The purpose of Ai-Da is not to prove that machines have a soul, or an essence or whatever it is you think is the source of artistic talent

• Ai-Da is an ethical project.

• I’ll quote the blurb for the exhibition:

• The Leaping into the Metaverse exhibition explores the interface between human experience and AI technology

• to explore the future of humanity in a world where AI technology continues to encroach on everyday human life.

• As Aidan Meller put it in an interview with the UK’s Guardian interview:

• How comfortable are we entering an age where he can’t distinguish between the work of humans and machines?

• For the next one: AI or robot?

Richard Bradbury: Robot?

Matt Armitage:

• Good choice. A few months ago we had a story about a magnetic corkscrew robot.

• Using magnets, it can be piloted around the body. This is a similar thing but way weirder.

• This is a robot made of magnetic slime. In the test video, it looks like a slug.

• It’s described as having a custard-like consistency.

• It was developed by researchers at the Chinese University of HK.

• They created a mixture of magnetic neodymium particles combined with borax and a resin made of polyvinyl alcohol.

• The magnetic particles are toxic, so they’re coated with a silicon compound that makes them safe to use in the body.

• And the result is a slime that can be directed by an external magnetic field.

Richard Bradbury: Why does such an awful sounding thing exist?

Matt Armitage:

• I don’t know how many of the people out there have watched 80s horror flick The Stuff but, yeah, this is The Stuff.

• Using that external magnetic field to stimulate it,

• The slime-bot can form around objects and cling to them, and even repair broken circuits inside machinery.

• If tests with human tissue go well, it’s hoped that it could potentially be used to remove foreign objects from inside the body.

• Things that might have been swallowed by accident, for example.

• They’ve used a prototype to remove a lithium battery from a model of a stomach – the kind of thing a child or a pet might swallow.

• They’ve also used it to grasp very fine wires as well as to stretch itself, almost like a liquid and manoeuvre through incredibly tight spaces.

• The slime can change its form – it can elongate itself, or as its makers show, roll itself into something resembling an octopus tentacle.

• It can even reform itself if the slime is broken or cut into separate pieces.

• Like I said – it is The Stuff.

• The next step is to make sure that it can be tracked inside the body so that doctors can be sure it’s retrieving the foreign object

• …and not turning a lung inside out.

• That was a joke. I think.

Richard Bradbury: We’ve had the robot. Let’s hope the AI story is less disturbing.

Matt Armitage:

• Much better. This is a story about protecting wildlife.

• Camera traps are frequently used in conservation.

• My wife helps to set them up in Malaysia for an NGO called MYCAT that protects tigers and other big cats.

• One of the drawbacks of the traps is that the data they give you is sometimes quite old.

• Many cameras are in remote and inaccessible places. So the cards can only be picked up every few months.

• And they then have to be analysed and sorted through.

• Researchers at the University of Stirling in Scotland have been working with wildlife rangers in Gabon to create AI-powered camera traps.

Richard Bradbury: Is there a benefit of having such powerful technology in remote locations?

Matt Armitage:

• It depends on what the traps are there to achieve.

• If you’re just logging the animals that come into the field of view and it’s not time-sensitive, then no. Probably not.

• But where it is time-critical it could rebalance the scales.

• The AI can be trained to detect certain species, so that when they step in front of the lens,

• …the camera can send a message and the photo data to the cloud.

• For endangered species like the African forest elephant, which is native to Gabon,

• …knowing the movement of the animals may help to protect them against poachers.

• The system can even be trained to recognize a human carrying a gun and send an alert.

• This would give rangers the kind of speed of response that could stop the poachers from killing any of the animals.

• All of which could help to protect biodiversity in the region.

Richard Bradbury: When we talk about systems like this – providing power for them in remote places is often the biggest issue.

Matt Armitage:

• For sure. This isn’t related to the last story, but it does address what you were saying.

• This is about solar panels that can generate electricity at night.

• Typically, solar cells don’t work at night.

• Solar panels essentially work on the principle of heat transfer to create energy.

• The panels are cool and absorb the heat of the hot sun.

• A layer of semi-conducting material in the layers of the panel takes this energy and converts it from heat energy into electricity.

• At night, that process reverses. The panel is generally warmer than the sky above it, so the energy transfer flows in the opposite direction.

Richard Bradbury: So this is like a night vision solar panel or some trickery to get reflected light from the moon?

Matt Armitage:

• I’m trying to imagine what that would look like.

• Green electricity, I suppose. Or electricity made of cheese.

• Researchers at Stanford University decided to use the property of the panel at night – that it’s cooler than the air around it.

• As another means to generate power.

• They modified a normal solar cell and added a thermoelectric generator, which produces current from the difference in temperature of two objects.

• Admittedly, the power produced is very small – just 0.04% of what the cell creates in the daylight.

• But that’s enough to power an LED light or recharge a smartphone.

• Further development could increase that power generation, and the product itself could easily become a commercial project.

• There’s probably no way to scale it into large scale power delivery.

• But for off the grid locations, it could reduce the reliance on battery stored energy for use at night.

• And ensures that those cells are doing at least a little bit of something for the hours that the sun is absent.

Richard Bradbury: Normally I’d ask you to do a really weird story to end with, but there doesn’t seem to be any lack of really weird this week.

Matt Armitage:

• How about another mind-boggling one?

• I’m still planning to do that story on the world as a computer simulation, once some, or even one, of the experts I’ve contacted, responds to me.

• But this is a story about an AI that scientists programmed to run a simulation of computer code within itself that an AI could run on.

• So, it’s an AI running a simulated AI.

• The normal process of teaching an AI how to do something is to train it.

• It can recognize someone with blue eyes because you feed it thousands of images of people with different.

• Researchers at the University of Pennsylvania wanted to find a simpler approach.

• They wanted to build a neural network that could run its own computer.

• They created a really simple neural net and linked multiple of them together so that they could carry out more complex tasks.

• This mimics the logic gates you find in computer chips.

• They then used it to – among other things – run the classic arcade game Pong.

• As a simulation within an AI.

Richard Bradbury: And if you scale it up, then you have the possibility that we’re just a simulation inside the mind of an AI?

Matt Armitage:

• I could just leave it there on that chilling thought.

• The purpose behind this, apart from having a very odd gaming machine,

• This kind of network could be teamed up with s new breed of machines called neuromorphic computers.

• Running the software on virtual neural nets could squeeze even greater performance out of the machines.

• If you’re a bit scared, NS quotes neural net expert Abdelrahman Zayed at Montreal Polytechnic

• Who states that replicating these logic gates two or three times is not the same as emulating a high-performance chip that may have billions of logic gates.

• So there’s a chance for you sceptics that this technology won’t scale much beyond Pong.

• Now, doesn’t that take your mind off Elon taking over Twitter?

Share Episode

Shownotes

Transcripts

Follow

Links

Chapters

Video

More from YouTube