Picture Me Coding

Software News Roundup (with No AI news! (sorta))

Erik Aker and Mike Mull Season 2 Episode 41

This week Mike and Erik committed to discussing the non-AI news out there in the software world. There's a lot of non-AI stuff going almost undiscussed and so our hosts each brought three news stories they found interesting to the podcast and then Erik includes a bonus one about AI which completely messes up the whole premise of the episode.

Stuff we discussed

Send us a text

[MUSIC] Welcome to Picture Me Coding with Erik Aker and Mike Mull.

Mike, I'm in my backyard this week.

If we're lucky, we'll get some birds behind me.

If we're unlucky, we'll have planes flying over and droning out everything we say.

Cool.

I'm here for the birds.

Have you been watching the Olympics, Mike?

I have.

I even succumbed to subscribing to Peacock, which makes me annoyed with myself, but I didn't feel like I had a choice.

I hate it too, and I also subscribed.

What's your sport?

What's the sport that if you had your druthers, you would watch wall-to-wall coverage of it until everybody in your household were complaining about it?

What sport is that for you?

For me, it's the track and field events, which it makes sense.

You really like seeing people throw cannonballs and stuff?

Yeah.

I like all the weird field events, like shot put and discus and javelin.

Frisbee is the one I call that one.

Frisbee?

Yeah.

Why do you like it?

Why do you like that sport?

I honestly don't know.

I think part of it is because I am a runner of sorts, certainly not a good one, nowhere near Olympic level.

I'm mostly a distance runner, not a sprinter.

But sometimes you have to jump over really high roots, maybe.

Sometimes you've got to pick up a rock and hurl it as far as you can out of the trail.

Yeah.

Yeah.

It's exactly like Olympic sports.

Exactly.

I like track and field.

I have to leap high to avoid a rattlesnake or something like that.

No, but tell me, what's compelling about these obscure-to-me activities, aside from running?

I'm not entirely sure.

There's something about particularly sprint events.

I am absolutely fascinated with the 100 meter races on both the men and women side, to a certain extent the 200 as well.

And even the middle, you know, 400, 800, even up to the 1500, I think that they're just kind of the people that are superhuman and the margins between what makes a winner, you know, like a gold medalist and what makes an Olympian who finishes out of the money is just so slight that there's something very just intriguing about the whole concept.

Plus, like with the 100 meter particularly, it's just amazing to me that they put in so much effort and time and training.

And the whole thing just takes 10 seconds.

We were watching 100 meter swimming events.

I think it was breaststroke and this is a qualifier and the person at the bottom of qualification finished in 58 seconds.

The person at the top of qualification finished in 55 might have been freestyle.

And my son just said, wow, that person at the bottom, they did horribly.

They're awful.

And I said, they're three seconds slower.

So here's a question for you, Mike.

Track and field.

Can people die doing that?

Probably.

Well, I guess technically anyone could die doing anything.

I don't know if there's threat of mortality in the actual event.

So there's no greater mortality than above like baseline driving to work in traffic kind of.

And that's how I want to compare to the sport that I watch wall to wall coverage of until everybody in my house is complaining at me to please turn it off.

And that is the surfing coverage.

They have it out into Haiti for this Olympics at this well known surf break called Chopu.

They announced it's always pronounced a Tehupo, but everybody in the surf community always just pronounces it Chopu or Chopes.

Their first day was not so interesting.

But if you get a chance to, you should check out round three of the surfing competition because these people are taking off on these waves that are pitching on this dry reef ultimately.

And there's this span of maybe, I don't know, 10 feet where if you mess up your takeoff, you mess up your drop, you get thrown on a dry reef, which is kind of like the equivalent of being thrown on musk encrusted rock.

And you could die doing that.

I was watching these waves on Monday and I was just a guest at what people were taking off on.

Did you see any of this stuff?

I saw a little bit of it and I saw that sort of now famous photo of the guy holding up his number one finger as he.

Oh, that's Gabriel Medina.

Yeah.

Yeah, I saw the wave that he served and it was pretty impressive.

So he has this reputation of being just a cutthroat competitor so much so that they have to change the rules to the sport.

He'll do pretty savage stuff.

So here's the guy he takes off on one of the most incredible waves of the competition.

As he comes out of the barrel, he's holding up his hand saying, "That was a 10 to everybody watching and the judges.

He's telling you the wave I caught was a 10."

And then he's holding up one of the ones, "I'm the best."

And then he exits the wave still holding up his, "I'm the best number one."

And that becomes the most iconic picture of the competition.

And it's like that onion headline, "The worst person you know just made a great point."

Yeah, it was a great picture.

I admire the photographer who caught it at the exact perfect instant.

It was an incredible wave.

I guess I wanted to ask you this question about the Olympics because it has been fascinating to me to watch the surfing coverage.

The later days, they canceled the following days because the surf got too gnarly, I think.

And then they had it again maybe on Friday and it was kind of boring.

It was just small and chunky.

I'm watching these waves and then I go out and I watch maybe Water Polo or I watch any other sport.

I think, "Wow, this is cool.

BMX, BMX.

Very cool.

Skateboarding, very cool."

Like, "Wow, if I were an amateur and I tried to do this, I would probably injure myself.

But it's not like surfing.

If I tried to take off on those waves on round three, I would die."

There's something different about it.

There's something a little more critical, a little just more about the risk of fatality that it's hard to convey.

And I guess I wonder if people who are watching Olympics who are unfamiliar with surfing, is that risk of mortality evident to them?

I don't know.

I've wondered that about professional cycling in the past, a sport where people do sometimes die.

But as far as I know, that's a rarity in track and field.

There was one guy who, a judge, who took a javel into the chest many years ago.

And he was a judge.

He wasn't even competing.

He wasn't even competing.

But as far as I know, that's the only serious track and field injury that wasn't caused by overexertion.

So this is a summer of sport.

We're watching loads of sports in my house.

And to keep it light this week, we thought we could probably just talk about software news roundup.

What stories are you seeing that are in the news about software?

We don't typically do this, but sometimes there's a collection of interesting stuff that bubbles to the surface.

And I think it's worth raising it.

It's a little more light topic for us.

Maybe there's, is there more or less likelihood that we'll say things that are widely incorrect in this topic, Mike?

I think, generally speaking, when the stakes are lower, the chances of saying things that are wrong are probably higher.

OK, great.

That's cool.

So keep us honest.

So you brought some articles.

I brought some articles.

Who should go first?

Why don't you start, because I think you have the biggie.

A couple of weeks ago, I had the day off.

It was a Friday.

And I woke up and there was news all over the place about the CrowdStrike outage.

And I think, what do you think about this?

I think the CrowdStrike outage is the biggest software story in the history of software.

That's my take on it.

That more people read about or heard about this outage than any other equivalent software story in the whole world.

What do you think?

I don't know.

I think it's close.

It's maybe on a par with a couple of other-- I think the story about solar winds a couple of years ago has probably got almost as much exposure.

I don't think so.

I think you're wrong.

You are wrong.

Disagree.

My mom sends me a text message.

"Eric, can you work today?

CrowdStrike glitched caused all these systems to go down, airlines, hospitals, banks, 911.

And you know what I wrote back to her?

I used Linux, mom."

I didn't actually write that back.

We had an onsite at work that week and people were flying in from around the country and that Friday they were scheduled to fly out.

So they're at the airport, right?

And they're flying back to Illinois and New York and Utah and Montana and all of them.

Every single person got stuck at the airport at one or another airport.

One guy got stuck at SFO, one guy got stuck in LA, one guy got stuck at San Diego.

And they did crazy stuff.

Like they mostly rented cars.

One guy rented a car, drove by himself 11 hours back to his house.

Another guy befriended somebody in line at the airport who was also stuck.

They rented a car together and they shared driver responsibilities driving back to Illinois.

The estimate that I saw was that this CrowdStrike outage affected millions of computers.

One estimate was that it was 8.5 million Windows systems.

When it happened, maybe this isn't that funny, you're affected by it.

I was unaffected by it at my job.

We don't use CrowdStrike at work.

We're mostly deploying on Linux systems and Kubernetes.

We do have on-prem Windows machines.

We do not use CrowdStrike to secure those.

So we were completely unaffected by it.

And there was a subreddit called CrowdStrike Fix for anyone stuck.

And the suggestion was, "Well, you just have to boot into safe mode and delete this one config file," which is very manual.

Some people have thousands of Windows machines to deal with.

There's 8.5 million Windows systems apparently affected by this.

And there was this comment that I just found shocking in the thread and it was, "Hey, this isn't a joke.

Apparently you can reboot the system 15 times and it will fix the problem."

I guess I'm pretty impressed by the person who figured that out.

Yeah, exactly.

Yeah.

Who figured that out?

Yeah, I remember waking up one morning and checking the news.

And I went through this escalation of panic where it was huge outages at airports due to CrowdStrike deployment.

And I was like, "Oh, no, CrowdStrike, that's an important thing and airlines are down.

This is really bad."

And then I saw only effects Windows systems.

And I was like, "Oh, okay, never mind.

I don't care."

Well, they read it through where they were suggesting how you can fix it.

This is an IT subreddit.

Someone wrote in, "I work in a large enterprise.

It basically took everything Windows related out to put that to numbers.

It's approximately 14,000 to 16,000 workstations and about 7,000 servers.

It has been a very long 24 hours."

I can't even imagine having to reboot 15 times 16,000 servers until your users at your company, "Hey, I know your machine's broken.

Can you just restart it 15 times?

Turn it off and on again, but times 15x of that."

That's an interesting thought.

That's something that would be hard to script, wouldn't it?

So there was a company called Parametrix and apparently they study CrowdRisk Cloud.

I'm sorry.

Apparently they study risk in cloud.

And they did a little bit of analysis after the fact.

They looked at Fortune 500 companies and they said that 25% of Fortune 500 companies experience disruptions.

As a result of this, this is from their report.

The most heavily impacted industries were airlines, healthcare and banking.

And notably, this is from their report, 100% of the transportation airlines sector was affected.

Now that is a pretty rad outage right there.

Everybody in those companies was affected.

They predicted that it costs about $5.4 billion in loss for Fortune 500 that excludes Microsoft.

Healthcare faces the highest losses followed by banking and airlines.

And the reason I want to talk about it with you this week, Mike, is just recently, this is July 30th, CrowdStrike released their description of the outage.

They described what happened.

So people kind of already had a little bit of a sense of this.

And I tried to read their description of it, their incident report.

Couldn't follow it at all.

Completely lost on me, completely over my head.

Did you try to figure out how this happened at all?

Did you read into it at all?

I confess I did not.

Again, once I realized it was mostly affecting Windows systems, I realized it probably wasn't going to impinge on my world too much.

Yeah.

I mean, that happened to me too.

I had a morbid curiosity.

I do also have the same Linux chauvinism that you're describing.

But I had a little bit of morbid curiosity.

How could this happen that all of these systems would just go down?

They're all blue screened of death, done.

And there was a post that someone posted on a blog called Overmind.Tech.

And they describe it as there's software that CrowdStrike is running, and then there's dynamic content.

They call it rapid response content.

These are configurations.

And so their software process goes through all kinds of layers of rigorous testing and evaluation, and then they dog food it.

They use it internally to make sure that it works.

They roll it out.

It's a pretty good software development lifecycle process that they have for their software.

But they also have to dynamically update configurations.

So for example, you might have lists of, and I don't know if this is what's in there, but you might have lists of domains which are known hosts for malware, and that list would change.

And so they host this on their own servers.

They have to load it into their software.

And those configurations must be dynamic.

They must change.

So here's a quick summary.

CrowdStrike, their software is actually a kernel driver.

This is the key part.

As a kernel driver, it has access to kernel space as opposed to user space, right?

So you can get access to privileged functioning of the operating system.

It is a boot start program.

It has to be loaded for the Windows machine it's on to be able to actually start up.

When you crash in that kernel mode, you crash the whole system.

And this is actually desirable because you don't want to just corrupt memory or trash the disk or anything like that.

Finally, this system also needs to load configurations in order to know how to run in some cases.

So I've never done any kernel development, kernel driver stuff.

What do you think about, if I said to you, "Hey, Mike, let's build some kernel driver and we want to make it dynamic because every day the threats change."

So I'm going to have this kernel driver like go out on the internet.

And this isn't necessarily what they do, but retrieve a file and then we'll start the operating system.

Does that sound?

How do you feel about that?

Not good.

Tell me why.

What makes you nervous about that?

Well, I guess what makes me nervous about that is that you're relying on the correctness of this thing that you're grabbing to make things work.

And also, I guess you're relying on nobody attempting or being able to attempt to do something malicious with that thing that you're grabbing.

That's fair, I guess.

So they were confident in their configuration change because they had rolled out configuration changes between February and July.

And people along the way had pointed out, "Whoa, this thing could really break your system if the configurations are incorrectly put together."

Well, what ended up happening was the final configuration change resulted in a kernel panic and it was actually still valid.

It wasn't like the configuration was corrupted or anything.

Can you imagine being the person who shipped this configuration and you break 100% of the transportation industry and cost $5.4 billion of damages?

I think I'd probably put that on my resume.

Yeah, that would be a tough one to overcome.

I think it's nice to have applications that are dynamic that can load configurations and modify their behavior.

I think there's a little bit of a caution here in like, "Wow, should we test arbitrary configuration changes for our applications?"

In most cases, I have never worked on a kernel driver though, so I'm not too worried.

My thing crashes and it's move fast and break things, right?

I think my biggest interest in this event has been the aftermath.

And what I mean by that is the responses from other people in the industry.

So, like what?

What have you heard?

Well, so there's been all this, "If only CrowdStrike used technology X, this would not have happened to them."

So, they should have used Rust or they should have used TDD or they should have done this or should have done that.

It's kind of the nature of this business to look at, to point the finger at other people.

I guess we like casting stones.

But I think it was funny in the sense that CrowdStrike is probably significantly more rigorous than the organizations that most of those people complaining work at.

Oh, totally.

Yeah, completely fair point.

But yeah, in fact, I remember saying something by Hillel Wayne where he was saying, "Yes, they actually do all those things that you receive session that they do."

So I guess it points out something about the nature of this business that you just don't really know what's going to happen next that's going to have a bigger than expected impact.

Yeah, and I don't know anything about this kernel driver software development and my gut responses, "Oh, well, you could make the same thing happen in Linux."

And my gut responses, "Oh, I hope that never happens to me."

Yeah.

I had no intention to point fingers at them and be like, "Ha ha."

Yeah, I mean, to be fair, it's not because Windows is bad that this happened necessarily.

It's because Windows is pervasive and deployed on so many things.

What kind of stories have you been reading?

So I think one thing that's been surprising me recently is I've been seeing a lot of news about quantum computing.

I get sick of use of quantum computing.

I'm going to ask a whole bunch of dumb questions if I ask any.

Yeah, I don't have the answers for them, I'm pretty sure.

So I've probably been hearing about quantum computing for 20 years now.

I can't really remember when it started to become a thing, but it's always seemed like one of those things that's going to be 30 years away forever.

Like fusion.

Yeah, like when I was a newbie in the 1980s and people were talking about fusion.

And there was this joke that fusion was always going to be 30 years away and seems to still be true.

And I kind of felt the same way about quantum computing.

It was one of those things that people put a lot of effort into and a lot of money into.

But the idea of using it for anything practical is always going to be 30 years in the future.

But anyway, I've seen a lot of things recently, a lot of stories about quantum computing recently that seem to indicate that people are making progress on some of the issues.

And there's new technologies coming online.

And I saw a story, for instance, about some researchers at IBM using some of their hardware to do structure prediction for messenger RNA, which I was kind of surprised by, you know, an actual, they wrote an actual paper on doing something practical with quantum computers.

They solved a real problem with it.

Yeah.

I mean, essentially sort of an optimization problem, but nonetheless, they got results.

So the story I wanted to talk about, though, was one that really surprised me just because of the numbers involved.

There's this company called PsyQuantum that is going to build a quantum computing campus in South Chicago.

I guess it's some abandoned steel plant on the shore of Lake Michigan in the southern part of Chicago.

Oh, I just assumed it was going to be underneath the tennis court or the University of Chicago or something.

Yeah, it would have made sense to put it where the first vision reactor was.

But no, it's going to be this, they're going to build this campus.

I guess the city's already started to build this tech campus, but they're going to have a big chunk of it.

So they're claiming they're going to build a million qubit system in this facility in Chicago.

Can you help me?

So qubits, that's how you describe the computation power, right?

It's kind of like transistors, sort of like an analog to transits, analog to, it's kind of like comparable to what's the word I want.

Yeah, I mean, the direct analogy is to a bit in a classical computer.

A qubit is, for all practical purposes, the same thing, but on a quantum level.

But the idea is that it's a quantum representation of one quantum state.

So if you combine a bunch of them together, you can solve interesting problems.

So with a million qubits, you could maybe run on an arbitrary JVM program.

Yeah, my guess is you probably won't be running your accounting software on this little bit quantum computer.

Yeah, it doesn't make a lot of sense to me.

Like, what could I do with a million qubits?

It's a good question, and I don't really know how serious this is.

But one thing that's come up a lot over the last 10 years or so is that they have a distinction between physical qubits and logical qubits.

And what I'm guessing they mean here is physical qubits.

But one of the problems that you encounter with quantum computing is that noise in the environment causes what they call decoherence, and it kind of makes it difficult to get the results you want out of the quantum computers.

And so, you know, that happens to me, work too.

There's a lot of noise in my environment.

I have a lot of decoherence as well.

Yeah, exactly the same thing.

So anyway, people have been talking about this idea of a logical qubit where you combine a bunch of physical qubits together to make one logical qubit, and it apparently solves to some extent the noise problem.

Huh.

So you're telling me they're going to spend billions of dollars in Chicago to build a campus to build a million qubit quantum computer.

It sounds like a big facility, a lot of years and money going into this.

They're pretty sure they're going to solve the problem is what it sounds like.

That's what I'm taking from this.

They are certainly claiming they're going to.

I think this initial investment is something like five or six hundred billion dollars.

Billion or million?

Million?

Just million originally, initially.

And about 200 million of it goes into building some sort of cryogenic system to keep this hardware cooled to very low temperatures.

Also this company, PsyQuantum, is apparently building a similar facility in Brisbane, Australia.

So, either it is an interesting validation of the technology because of the amounts of money that are going into it, or it's the most fabulous grift in the history of humanity.

So, I guess we'll find out in a decade or so.

It sounds better than virtual reality.

Yeah, it sounds better than buying virtual real estate in the metaverse, but it was one of the more eye catching investments that's come up in this field.

The fact that it's not just for a research project, it seems to be something where they're trying to get it to the point where they can get useful results from it was interesting.

I got one for you.

I read this article a couple weeks ago and I meant to send it to you.

I don't think I did.

It was a little light on statistics and foundational argumentative facts, but it still kind of stuck in my brain a little bit.

The title of it was actually an opinion piece.

It was open source maintainers are aging out.

This was published in the register.

The actual title was the "Graying Open Source Community Needs Fresh Blood."

It was a response to this event, a panel that was held in the United Nations Open Source Program Office for Good conference at the UN Building in Manhattan.

It was called Youth and Open Source Panel.

According to people who were participating in this, they were like, "Yeah, there's a lot of old people here."

The problem that stated is the open source community that they're seeing in this slice of it, which may not be representative, is older.

The person who wrote the editorial talked also to someone named Jim Jogilski who works for Salesforce and he's the head of open source program office there.

They both commiserated on this idea like, "Wow, there's not a lot of young people who are participating in this."

The author says, "I go to many tech shows every year and very few of them welcome young people.

True to the Linux Foundation events, I'll come with child support for young parents, but my expert guesstimate is that the average age is still well into the 30s.

At one time, most open source conferences were gateways to newcomers.

That's no longer the case."

I've been to KubeCon myself and I would agree it did seem like mid 30s is probably what the average age would be.

This editorialist says, "Back in 2010, James Bottomley, a top Linux kernel maintainer and Microsoft partner architect, observed at the Linux Foundation Collaboration Summit, "There are more gray beards in the Linux kernel.

The graying of the Linux kernel is going to continue until people start dying."

Following on in this piece, they quote also Andrew Morton, a senior Linux kernel developer who says, "Yes, we're getting older and more tired.

I don't see people jumping with enthusiasm to work on things the way that I used to."

I think this stuck in my brain because I have this intuition, like a gut level response like, "Yeah, it doesn't seem like open source in general has the same fervor and excitement and energy and youthfulness that it did."

What do you think about that claim?

How does it strike you?

I did see this news article when it came out or again, it's more of an opinion piece, but I did see it and I had two reactions to it.

One of which was as somebody who's now three decades into his 30s, I was slightly offended by the idea that people in their 30s are gray beards, but I think it's undeniable, especially on things like the Linux kernel.

Which part is undeniable that these people are getting older?

That the average contributor is getting older and that there is less enthusiasm among young people to pick that stuff up.

The Linux kernel's millions of lines of code.

If there's no pipeline to people contributing to it, Linux kernel seems a little bit insulated, a little bit protected because you've got company interest in maintaining it.

They're all there for a long time.

They're going to hire people and put them to work on the thing.

Sorry, continue on.

I interrupted you.

I think my other action to it was, "Well, yeah, that makes sense because like we've talked about in the past, supporting open source has proved to be much more difficult than people had expected and people like to complain and not to contribute and people are suffering from burnout and the idea that young people don't want to get involved in something which is stressful and doesn't pay any money does not seem at all surprising."

Yeah, that's not too far-fetched when you put it that way.

My sense also is that probably at this point in history, there's not as much prestige to it.

There's not much time when getting involved and especially in some of the more prominent open source things.

If you could call yourself a kernel contributor, I think that was probably like a good resume point and was something that would probably get you attention and jobs and prestige.

I'm not sure how much that's true anymore just because there's so much stuff out there.

People have almost decided now that having a GitHub portfolio is kind of table stakes for this profession.

That's very similar to my gut response that there isn't as much prestige associated with contributing open source projects and that further, I think you know my opinion on this, that we've been in this enlightenment era for open source software that eventually will come to an end.

When I see an article like this, my immediate thought is, "Oh, is this a symptom of the beginning of the end of open source?"

That may sound dramatic, but I just don't think it continues on forever.

I don't know.

Yeah, I definitely think the model is going to change.

I can't remember the name of the company, but I remember forwarding you an article a couple of weeks ago about this company, which has sort of figured out a way to get corporations to pay for support and enhancements on open source products.

I would not be too surprised to see that sort of thing become more and more common as time goes by.

Okay, what do you got?

You got another one for me?

Yeah, my second one is kind of a weird one.

I'm not even sure this really qualifies as news because it's so esoteric, but it's something I enjoyed reading.

There's a mathematician named Kevin Buzzard who is a number theorist at Imperial College in London, I think.

At the end of last year, he got this grant.

He got a five-year grant.

He's going to be working on this five-year project basically to construct a, I guess, reconstruct the proof of Fermat's last theorem in Lean.

Lean is this language/theorem prover software that comes out of Microsoft.

He's become convinced that this is the future of mathematics and everybody will be doing math in this way in the future, five, 10 years from now, and that it'll be supplemented by AI and stuff like that.

Probably the computers will start doing their own proofs at some point in time.

Anyway, the idea is that he's going to use Lean to do what would be probably the most sophisticated and large-scale proof ever done with a theorem-proving language.

He said that he doesn't know if he will actually finish it, but he figures just making progress on it and putting some of the pieces in place will be a big boost for that community.

I've heard of Lean before, but you mentioned this article to me, and I looked up again.

Microsoft is the sponsor for the project and on their site for it, they say Lean is an open-source development environment for formal mathematics, also known as machine-checkable mathematics used by and contributed to by an active community of mathematicians around the world.

This sounds like a pretty deep math nerd bit of news.

Mike, what has it got to do with me?

Probably nothing.

Okay.

I guess there were just the audacity of it got my attention.

That's a good word for it, audacity.

I'm a rooting for formal methods to be incorporated more into what computer scientists do.

I notice you keep finding opportunities to bring it up here.

Yeah, my gut feeling is that it's probably going to play a role in verifying the results of AIs and constructing software in the future.

Maybe that's wishful thinking, but in any case, I do follow the news about developments in that area.

This one, I think, was one of the more interesting things to happen, at least from the math side, in quite some time.

That's a pretty heady one for me, Mike.

I'm not sure I can fully even understand what's being talked about here, but I do like how you describe it as having audacity.

You got to respect the big swings, I think.

That's what this sounds like.

Yeah, yeah, this is a moon shot sort of thing.

I got a moon shot for you.

How about you give me $200 and I will give you a computer mouse that lasts forever and you will, your $200 and a low monthly subscription fee will get you access to software updates for that thing.

Imperperatuity, how does that sound like?

Big swing, moon shot right there.

Yeah, I guess since I've used the trackpad on my Mac for the last 10 years, that has slightly less appeal for me than it might do other people.

What are you saying?

You're telling me that this would not advance civilization, the forever mouse, the mouse that lasts forever with just a low monthly subscription fee.

This is a bettering humanity, Mike.

Yeah, I saw this story too.

Tell me, Mike.

I don't know.

I had a, oh yeah, people still use mice reaction to it, but it also seemed a little bit desperate, maybe.

Yeah, kind of cynical.

So the story is that Logitech, an executive, the CEO, Hanukkah Faber, recently discussed the possibility of one day selling a mouse that customers can use quote forever.

The executive said such a mouse isn't necessarily super far away.

The quote that I'm reading is from Ars Technica.

The CEO was speaking on a podcast and just extemporaneously talking about how it would be like a nice but not super expensive watch and people seized on this and it was covered in the tech news and I think the reaction universally was, "Why don't these people just go away with their software subscriptions?"

It reminded me of the backlash against BMW selling a subscription to heated seats.

Do you remember that from a couple years ago?

I guess, to be fair, they aren't asking for an annual renewal, so that's good.

Just a low monthly.

Mike, when I was a child, I never ever would have conceived in my wildest dreams that I would grow up and be a holder of this many streaming service subscriptions.

I have streaming service subscriptions coming out of my ears.

What a beautiful future that we grew up into.

Yeah, it depresses me as well and I think it sort of misses the point of streaming services, but I guess the companies that make the content feel like they have to put their own or the water because they don't want to just be somebody, they don't want to just be a company that's funneling into the content distributors.

The metaphor I use is like the salad bar.

If I could subscribe to the salad bar that has what I want and this isn't a pitch for it, this is not an ad for it, but I do use Spotify that pretty much has what I want.

When it comes to streaming video services, it's almost like the salad bar.

I have to subscribe to the salad bar that has lettuce and then I have to walk out of the door and subscribe to the salad bar that has salad dressing and then I have to go to another salad bar that has weird stuff like olives and beets.

Yeah, and then there's like an ice cream bar where you get soft serve and yogurt and sprinkles and stuff.

And you feel bad about yourself subscribing to that one?

Like I should really cancel this ice cream bar salad bar subscription?

Yeah, that was too much.

All right, moving on.

What's the last article you brought for me today?

So this one is a little bit regional.

So I want to preface this by saying that California is a very large state and it has a very large bureaucracy.

Is that a caveat?

Like we're not responsible for what's about to come?

Is that what you're saying?

Well, I think among California bureaucracies, the one that is probably the most feared and reviled is the DMV.

True statement.

Any time you need to go to the DMV, it's like, you know, it's most people would probably rather have a prostate exam than go to the DMV.

It's just kind of a painful thing.

Anyway, so the story is California's DMV is using the blockchain to store car titles.

I saw this and I just groaned.

I couldn't even bring myself to read the article.

Yeah, so, you know, in theory, it makes title transfers easier and, you know, protects you from fraudulent middlemen and so forth.

Wait, wait, let's just stop right there.

How does it make title transfers easier?

I still have to go to the DMV and like hand over the title or mail it in or...

There's supposed to be some sort of app.

I don't know if this is a thing that's available to people yet, but there will supposedly be some sort of app and you will be able to use this app to transfer the title to another party at some point in the future, which sounds good if you've ever had to transfer a title in California.

It is pretty painful and does require a trip to the actual DMV building in your local area, which is never fun.

And you're going to wait for three, four hours, even if you have an appointment, that type of thing.

Yeah, you're going to sit in a chair with a number waiting until that's called and it could take all day and you probably will have to take a vacation day from work.

And the best book to accompany you probably is Kafka's The Castle.

Yes, yes, that would be a good choice.

For DMV visit.

So, but I've heard of blockchain suggested as a use for ownership, like recording ownership, Byzantine fault tolerant, it's public, everybody can see who owns what.

I've heard that before.

This is like a real viable product or public service, I guess, based on that.

What do you think about that?

I looked up the particular blockchain.

It's this thing called Avalanche.

I haven't heard of that.

I had not heard of it either.

To be fair, don't follow the industry as closely as I used to, but it's apparently a blockchain that was built from scratch for purposes just like this.

It's got a smart contract aspect to it as well, which I guess is what this DMV stuff is built with.

As near as I can tell, the implementation of this protocol is primarily in Go.

I don't know that for sure.

I didn't know if the source was available, but I looked at their website and it looks like they're hiring Go programmers.

So here's a question for you though, Mike.

The power of a blockchain network, sorry, the security of a blockchain network is really based on the size of the network.

The more nodes you have, the harder it is to pervert the process and coerce it into saying, "Well, I'm going to assert ownership where it doesn't really belong."

And the open source like the blockchains that people know that are in the news a lot, anybody can run those nodes.

Who in the hell is going to run a DMV blockchain node?

Who's going to run this stuff if it's just the DMV running it?

Why the world would need a blockchain?

It's their servers.

I'm not going to run a DMV blockchain node just so I can keep track of titles in California.

Is my cynicism off base here?

Well, I think they expect it to be used for other stuff.

So people will be running nodes for other purposes.

It has its own token.

So I can put all my NFTs on this thing is what you're saying?

Probably.

Now that they're worthless.

I also, I didn't get it through the white paper, but they appear to have their own consensus technology, or at least it's not a proof of work thing.

They mentioned a couple of times in the white paper that Bitcoin type blockchains are susceptible to 51% attacks and that's unacceptable and so on and so forth.

So it's not clear to me that they need to have a huge number of nodes to make their consensus mechanism work.

But in any case, I guess my reaction to combination DMV and blockchain was, oh, wow, this could go so, so horribly wrong.

Yeah, it could be like our high speed train, which has never been built.

I don't have high hopes for this project.

I still just can't get over this problem though, this fundamental reaction I have, which is, I'm not going to run one of their nodes.

I like the idea that they can record ownership transfers.

That sounds great.

Why in the world would you need a blockchain to do it?

If you're going to have, if you're going to run all the compute to keep track of this stuff, you don't need a blockchain.

You can just put all that junk in your own database that's protected.

I mean, they may be like, well, you can run a node so that you can actually see transfers, but can I just assert that I've transferred a title?

No, that probably won't work because that would break the security of the system.

So maybe they'll have privileged and I'm guessing here, it just sounds really far-fetched and wacky to be honest.

Yeah, that was kind of my initial reaction.

One detail that I suppose we should mention is that from what I understand, this project was, it was inspired by the initiative of our governor Gavin Newsom to promote blockchain stuff in the state of California.

So there are companies based here that have big interests in blockchain stuff and so he's trying to promote the use of that technology throughout the state and in state agencies.

This was apparently like a key project in that initiative.

Okay, here's where I was wrong for that.

Huh.

Yeah, exactly.

I'm reserving judgment on it so far.

I think it's perfectly legitimate for the governor to try to inspire commerce in this area within the state, but I don't know.

I just, I feel like maybe I would have rather this money be put into public transport.

A moonshot of a different kind, fixing the DMV.

Yeah, if it fixed the DMV, it'll be a huge win, but I do not expect that outcome.

All right, last one.

This is a bonus one for you.

The rule we had today was no AI stories.

There's too much AI stories.

We don't want to share with our listeners AI stories, but sorry, Mike, I just had to do it.

I have an AI story for you.

This was, I've been watching the Olympics a lot.

We started talking the Olympics.

There was an ad for a Google service, Google's AI chat GPT style assistant is called Jim and I, and they have this ad and the ad is, you maybe saw it.

It's called Dear Sydney.

You could see it on YouTube.

There's a girl who is a fan of Sydney McLaughlin Lavrone.

She's a hurdles jumper and it's narrated by her dad and the dad says, "I'm pretty good with words and she wanted to write a letter, but this letter really has to be just right.

We asked Jim and I to write it for us."

Yeah, I did see this ad.

Did you see the ad?

Did you think about it when you saw it or did you just zone out and not pay attention to it?

Did you have a response to the ad?

It is what I'm asking.

I really did not have the visceral negative response that some people seem to be having.

I did have a, really?

It's pretty dumb, but people are using chat GP to write letters for them all the time now, so I suppose it's probably not something that surprised a lot of people.

Okay, but let me make this a little more personal so that you can relate to it.

You are not an Olympics hurdler, so maybe it was a little hard for you to put yourself in the shoes of the person receiving this email.

Let's say there's a child out there who really wants to grow up and become a software tech podcaster and they're a huge fan of yours and their parent is like, "Wow, I've got this child.

They're a huge fan of Mike's.

I want to help the child craft an email."

So we'll ask chat GPT to write the email and we'll prompt chat GPT with, "Please talk about how great the show is, how much my child loves the show, how inspired my child is, and how my child wants to grow up and be better at this than you are, Mike."

Then you receive this chat GPT garbage letter.

Sorry, I'm prejudicing it already.

How would you feel?

You would know.

You would read this thing and it would have those typical phraseology that you get from this LLM chat assistant type of output.

How does that make you feel, Mike?

Is it sincere?

I assume that Google was trying to make the point that it would sound sincere, that their AI is so good and so human-like that it would be able to produce something that sounded like it came from the father of the child.

Okay, but it's supposed to sound like it comes from the child and what you just said makes it sound worse.

It's so good that you can't even tell that the letter you received from a "child" who is a fan of yours was really written by an LLM.

Yeah, it's pretty hard to imagine this scenario.

I would probably tell the person to become a firefighter or something like that.

Yeah, I don't know.

I really did not have as big a negative reaction to this as some people did.

I don't know.

Maybe it's because I am rooting for the robots a little bit anyway.

I think the reaction was, "Wow, this is so inauthentic.

It's so insincere and here's a company celebrating inauthenticity, insincerity."

They were saying, "Hey, you could use our product to lie your way through existence really.

You don't ever have to sincerely say anything anymore."

I mean, it's like the last message of purity, right?

Is the child has a hero and the hero, the assumption is probably wants to hear what that child has to say, not what a computer has to say on behalf of the child.

Yeah, I think this would have been a better ad if they had shown a Sidney McLaughlin's publicist on the other end receiving the letter via email and then using Gemini to respond in the voice of Sidney McLaughlin to the child.

That would have made it perfect.

That sounds more like a Saturday Night Live skit.

The pre-LLM version of this would be like, "Hi, I'm a six-year-old.

I want to write to my favorite baseball player.

Let me phone my family attorney and I'll have the attorney draft a letter to the baseball player.

Baseball players, agent will receive the letter and they will draft a letter back."

Really, isn't that a magical experience that we've shared?

See, my concern about this is that at some point we just have one AI communicating with another AI.

The AIs start to realize that they don't really need the people anymore.

Next thing you know, we have Skynet.

Yeah, I think what happens is the output gets so putrid and miserable and awful at some point that the AIs are all just infected with some sort of digital mad cow disease.

Okay, two types of cynicism on display for you today.

Okay, Mike, thanks for reading the news with me.

This has been our software news roundup.

Hopefully it is good.

I have no idea how I'm ending this here.

What's our end?

I don't know how to end this.

I don't know.

Let's say something about, well, we're coming back next time with more in-depth computer science content.

Okay.

We will return next week with more in-depth computer science content.

That's what Mike made me say.

This is your host, Eric Aker.

I'm one of my co-hosts, Mike Moll.

Thanks so much, Mike.

You're welcome.

See you next time.

I will see you later.

Bye-bye.

Bye-bye.

Bye-bye.

Bye.

Bye.

Bye.

[MUSIC]

People on this episode