The Butt Ugly Weblog

What slightly bothers me about a lot of the AI coverage is the discrepancy between the hype and reality. Some folks are screaming AI apocalypse, and some folks go and "I tried ChatGPT and it gave me shit answers."

Just a few years ago, we had an equivalent situation with EVs. Some folks were saying that Tesla will build all cars in the future, and others just saw a big bunch of panel gaps and range issues. That they would never replace diesel. And you couldn't charge them everywhere.

Yet almost overnight, EVs and chargers are suddenly everywhere (in richer countries). Tesla is big, yes, but not everything [in Europe & China; US is weird]. It just took the understanding on how to make them cheaper and faster to build, and then building the factories and supply chains. Fundamentally they are better vehicles, but the first implementations were buggy as hell. They are not perfect, but they have become _good enough_.

There's a ton of panel gaps and quality issues in current generative AI tools as well. But consider the speed at which they are improving - ChatGPT is nary a year old - and we don't yet know if they are hitting any sorts of limits. It again becomes a question on how to make them cheaper and better so that they can become ubiquitous. We're still pretty much at the "throwing spaghetti at the wall and seeing what sticks" -phase of AI; figuring out where and how it could and should be used.

Neither do we know what the real dangers of AI are. EVs didn't become a public menace because they are so quiet either, even though that was a big argument against them a few years back. We solved it by making all EVs make a mandatory noise.

So don't take the current AIs as representative of the future. Neither take the hyperbole. Reality, as always, is somewhere in the middle.

(Unless someone gives them weapons and makes them eat organic material and self-replicate. Then we might have some Horizon Zero Dawn -level problems.)

Myself, like an increasing number of other people, have deserted Twitter and moved on to alternative services. (Not blaming Musk though; it was a cesspool before and Musk just made it worse through his lack of understanding of human behaviour. Twitter would need a game designer as a CEO, not a manufacturing geek. Anyway, I left Twitter already on the early COVID days...) So you can find me on Bluesky as @ecyrd.com or Mastodon as @ecyrd@universeodon.com. Of these, Bluesky seems to be enjoying faster growth, though I do like Mastodon's filtering features more.

Anyway, what I find interesting is not the tool themselves, but the meta-discussion around the tools. Many people announce their escape from Twitter and almost invariably the first question on a new platform is a question of identity - what should I write about here? How will others see me? What kind of a response do I get? How do I engage? Who should I follow? How do hashtags work? It's kind of endearing to watch: like a number of puppies in a new pen, poking around all the toys. Everyone's account is the same with a single post and a few followers. Fun! Much potential! Wow!

The next post, regrettably, often is about "nice place, how do I keep this place from turning into a toxic pool of twitterisms" or "how do I deal with trolls when they eventually arrive". It feels like people escaping a war zone. People, who are sure the war will follow them wherever they go. Social PTSD, almost.

The obvious answer is to bubble up - create your own Discord and discuss there. Community will throw the misbehaving people away, and new account moderation will keep bots at bay. But there are significant advantages to having a public forum too. You find new people, new information, joy, sadness that you cannot find from inside your own bubble. Social bubbles are after all quite slow-changing and static, much like cosmic bubbles (and entirely unlike soap bubbles. But I digress.)

To me dealing with trolls is about energy consumption. Assume a trollish message written by a human costs one unit of energy for that person. Doesn't matter what the unit is, let's just call it "1 troll". Since time and human capacity are limited, I believe the best way to respond is to make sure your response costs less than "1 troll". This is why things like engaging in a debate does not ultimately work: if you spend 10 trolls worth of energy to respond to a single troll, they will respond with another 1 troll message, and you will use again 10 trolls of energy to respond. The end result is that you will spend most of your time just dealing with a single troll. God forbid if there are more!

However, if your response takes, say 0.5 trolls, the more the troll responds, the bigger the gap between yours and their time usage will be.

This is why blocking almost works - it takes only 0.1 trolls to block someone. I say almost, because the whole thing is broken due to bots - the cost of creation of a bot and making it spew out nonsense goes asymptotically towards zero. So you cannot win a manual block war against bots.

This is where the platform must take a firm stance and make sure that there is a big enough - but not too big - a cost of creating and operating a bot. Yes, I know of the big debacle around paid API access on Reddit, but in my mind, this is necessary. The key is to remember is that the cost does not have to be monetary - it can also be less tangible stuff like red tape. Make the free API option such that you have to request an API key via a fax machine or a physical letter, and it might deter the bot authors somewhat. (Don't actually do this, it was just an illustration of an idea. :-D)

But as long as there's not a bot problem, blocking hard and fast is simply the best long-term way to engage with trolls in any social media. And I say this as someone who has been on pretty much all of the social media before they were even called that, and who has probably tried all of the methods from backtrolling to completely ignoring trolls. The only thing that really works is blocking; starving them of space and time.

Jane Friedman talks about how AI is messing up with her work, since anyone can write and publish books under her name for a quick buck.

In the old cyberpunk dystopias, the scary stuff was always in corporate vaults, and a big fear was that a smart AI would escape the corporate computers and roam wildly on the internet. I don't recall an author who thought the dystopia would be a small company trying to turn in a quick buck by making their AI available for everyone. But here we are, and we're not at all prepared for this.

Found this from my daily AI newsletter:

So yeah, basically future job market will have AIs writing letters for other AIs to digest. I wrote about this content explosion a bit earlier, but I still find it fascinating to watch it take shape. It's like the AI output will become the lingua franca of all professional interaction.

The fun/great part of it is that it could have been a complicated, designed machine language; a general interchange format designed by people to be easily machine-consumable (there is a myriad of them already). Turns out that the best way for computers to communicate is basically just polite, verbose human language. Ha, take that, XML!

Something that passed unnoticed was the 20-year-anniversary of this blog. I started out with a very witty Finnish Hello World -like text:

The quality has been downhill ever since.

However, I'm wondering... Should I arrange some sort of a celebration? Write a memoir entry? Reminisce about "ye goode old days" of blogging when we all knew each other and had those unforgettable monthly orgies? Right now I don't really quite know how to handle this, so if I a) have any readers still left, and 2) they might have some ideas, feel free to pitch them.

A few months ago I asked for some recommendations on quick comedy series to watch, and quite a few people recommended The Good Place. I was a bit sceptical at first, since I in general shy away from religious-themed shows, but after watching a few episodes I ended pretty much binge-watching the whole series. There is just something about mixing philosophy, Molotov coctails, British parties and comedy and the series worked very well for me.

Without spoiling too much, it's great on Season 1, a bit weird on Season 2, the whole Season 3 is a bit of a letdown, and I watched Season 4 in pretty much three helpings. There are some pretty solid actors, with D'Arcy Garden shining through the entire series as Janet, a fairly omnipotent computer who is most definitely not a girl.

The final episode, "Whenever you're ready", was a somber surprise after so much weird and fun, and it feels like something of a completely different series. Instead of being pointy or funny, it was simply quite beautiful. And through comedy and absurdity, we did learn something about the characters and watched them grow, and letting them finally go was... weirdly sad and happy at the same time. I wasn't quite expecting this, but in the end it was all very okay. It's no wonder this episode grabbed both the Nebula and Hugo awards. In fact, the whole series grabbed quite a few nominations in many galas, though surprisingly few wins.

Anyway, two thumbs up. Let me know if there are other comedy series to watch. Criteria: must be a comedy, streamable, episodes 30 mins or less, English/Finnish subtitles please.

I have to apologize: I have no idea how long the RSS feeds and stuff have been broken on this site. Probably they broke while I dockerized the whole thing. Welp. They should be good now.

Virpi Eroman "Kahenlaisii kuvia" on kuitenkin nyt sellainen levy, että pakko sitä on nöyrtyä ja kirjoittaa jotain. Oma avioeroni vuosi sitten tapahtui juurikin samaan aikaan kun tämä Eroman debyyttialbumi ilmestyi, ja kun autoni Spotify alkoi ehdotella biisejä tuolta levyltä, niin pidin sitä oikeastaan jo jonkinlaisena edistyneenä algoritmisena vittuiluna. Jaa että erosit, soitetaan sitten Eromaa. Ero, Eroma, same difference.

Mutta kuuntelin silti. Ja pikkuhiljaa huomasin, että pollajukeboksi alkoi tarjoilla niitä korvamatoina ihan muutenkin. Sävelkulut kun ovat hyvinkin tarttuvat, etenkin kappaleissa Tuhat kilometrii ja Huomenta maailma.

Sitten aloin kuunnella sanoja. Ja voi pojat, että niitä oli. Ei tule äkkiseltään mieleen kovin montaa muuta suomalaista sanoittajaa, jotka osaavat kuvata ahdistuksen, masennuksen, menetyksen, kaipauksen, löytämisen ja uudelleen itsensä löytämisen tunteita kuin Virpi Eroma. Jokainen biisi on kuin pieni, hiottu jalokivi, josta on kivulla hiottu pois kaikki se, mikä ei tee siitä jalokiveä. Eroma onnistui kuvaamaan minunkin avioeron jälkeistä tunnemyräkkääni aivan käsittämättömän osuvasti.

Tiedättehän voimabiisit? Ne, mitä laitetaan soimaan ja sit juostaan Kuusamoon ja takaisin? No, nämä Eroman biisit ovat... sielubiisejä. Ne tulevat istumaan sun viereesi sohvalle ja haluavat kuulla mitä sulle kuuluu, eikä niitä kuunnellessa tunnu siltä, että on yksin. Eroma laulaa suoraan "sielusta sieluun", on ehkä paras kuvaus siitä, miltä minusta tuntuu. Voi olla - ja lienee jopa todennäköistä - ettet koe näitä kappaleita samalla lailla, mutta kannattaa ehkä kokeilla. Elämässä pitää olla uusii kuvii.

Mutta mitäs mitäs - miksi tästä artistista ei ole kuultu aiemmin? Eroma itse selittää albuminsa syntytarinassa, että hänelle sanottiin suoraan, että hän on liian vanha musiikkiteollisuudelle. Eikä hän ole myöskään julkkis, joten "ei kelpaa" teollisuuden raaka-aineeksi. Joten hän käytännössä opetteli itse tuottamaan ja tekemään omat levynsä - sen lisäksi, että sekä kirjoitti, sävelsi, ja soitti omat laulunsa. Pakko nostaa hattua - ja myös sen takia tehdä tämä pikainen arvostelu. Koska jos ei ole systeemin tukea, niin fanien pitää tukea.

Tässä meillä siis on artisti, jolla on ääni, taito ja tarina. Ehdotan tutustumaan - ja toki nappaamaan liput (ehkä) vuoden ainoaan konserttiin 3.5 Tapiolan Louhisalissa.

Turns out this was one of the successful ones, and after a few shenanigans involving our beloved Finnish Post Office (who - you guessed it - lost the bag in the mail), and I got mine and it was way better than I expected it to be. It's a bit heavier than you'd expect, but that's because it's designed as a camera bag, so it's got a bit more padding everywhere. And the way you can redesign the interior with movable walls is simply genius. If you've seen me, you may have had to endure some way too wordy praise on the Everyday Messenger bag I've been carrying around for years.

Well, buckle up! (ChatGPT insists that I keep using this phrase whenever I ask it to write a blog post in the "Butt Ugly Weblog style"). There's more wordy praise coming!

This winter the bag finally broke down, with the strap having eaten through the attachment point. I really like the bag and the durability otherwise, so I sent Peak Design a letter asking if they have any spare parts, so I could buy them, and fix and continue using my bag, thank you.

They responded immediately saying that "hey, no worries, just fill a warranty claim and we'll get it sorted out." I thought - a warranty claim? That sounds odd for spare parts, but no worries, I can fill whatever you want.

A bit later I get a message "Sorry about the problem you're having, we'll send you another bag."

I'm floored. The bag is frigging seven years old, and you want to send me a new one? "Yeah, we don't make the exact same one anymore, so we don't have any spare parts, so we'll send you a new bag under lifetime warranty."

I'm more floored than I was before. A lifetime warranty? Seriously? A few days later, I have a spanking, brand new backpack that fits pretty much everything I could need, and which is padded to the max so I can put whatever in it and it's quite safe. It's also filled with a ton of geeky features that make my inner nerd purr very happily, they're climate neutral, and they have a mission I can get behind on.

I know it's a big thing to say, but you know, sometimes you just find a brand that you know you want to stick with for the rest of your life. And Peak Design has just done that by winning me over with their products, quality, warranty and their customer service.

There may be better bags out there. But honestly, I don't need to look further. This thing does everything I want it to do, and I have faith in the company that makes them. So satisfied, I'll continue talking your ears off on this one.

"That was NOT my fault."

"Hello, chair."

"...the carpet!"

"You have the conn."

Just a few quotes on episode 9 "Võx" that had me cheering, yet don't spoil anything. You'll get them when you see them.

The third season of Picard has been surpassing all of my expectations. Especially since the hapless and weak season two. The pace just keeps increasing, with plot threads seeded decades ago being tied together. The writes (all hail Matalas!) have managed to find a beautiful balance between fan service and an actual plot, and have been able to give meaningful character moments to all actors, who are really excelling in their roles.

Let's just hope that the last episode doesn't ruin everything - as it's not exactly unheard of in Star Trek, such as season two of Discovery - but still, nine out of ten episodes is a pretty darned good result already. This season has been on par, or even above, of the best two-parters in TNG, such as Birthright, Best of the Both Worlds and Chain of Command.

Really looking forward to next week.

by KimnCris Knight

I tuned out after a few episodes. Monster of the week, teenage girl problems, not particularly funny even. So I left it be, and expected it to be forgotten, just like so many TV series from the 90s.

After a few years, I realized that I had probably been utterly wrong, since many of my friends started to speak about it as if they had found the entire series a defining - even a transformational - experience of their lives. So I got intrigued again, and when both Buffy and Angel became available on direct streaming, I started rewatching them. Then they left the streaming service (Netflix IIRC), so I had to borrow them from the library, which was a chore, and then I was finally able them both when they became available on Disney Plus.

It's a total of about 189 hours of watching, if I counted correctly. It's not something you can easily do these days, with the kids and all, but turns out you can enjoy them quite well while cooking. Just flip the burger during the fight scenes - you won't miss much. All the interesting stuff happens between the fights anyway, since at its core both series were about the interplay between the characters and the big plot was just a backdrop and an interesting reason to throw crap on the character's faces.

I'm now old, so I wouldn't call the series transformational. Entertaining, yes; thought-provoking; sure. Silly? On occasion. But I can certainly see how watching these in my late twenties would've impacted my personal development to a degree: they handled growth, relationships, death ("The Body" anyone?), fears, happiness, belonging, transformation, etc - all packaged in quick-witted parlay between the characters.

Unfortunately it has later emerged that not everything was well during the filming. In fact, things got downright creepy at times. While there is merit in looking just at the series in a detached fashion, I don't think it is the right way in the end. We need to consider also always how things are made, not just the end result.

But the series are still good.

P.S: I really, really liked the way Buffy took music as an integral part of the series. I don't recall very many shows from that time (or even later!) where music has made a serious impression on me - usually it's just background noise, but in Buffy, the score is at times quite profound. I might even call "The Bronze" as a series regular character. The theme continued on Angel, with Andy Hallett as the music-loving demon Lorne. So nice. Good opportunity for comedy as well as the calms between the storms.

Kari Haakanan Tapa hänet ensin on erittäin mukaansa tempaava pageturner, kansainvälisen tason jännäri. En edes muista, milloin olisin viimeksi lukenut kirjan yhdeltä istumalta läpi, mutta nytpä näin kävi. Vahva suositus!

Niin, ja erinomaista Minna Canthin päivää. Lempilainaukseni häneltä

I think we now know what Web 3.0 is for real. We had a false start with IoT (where S stands for security, as the joke goes); and then the crypto folks started waving blockchains as Web 3.0. But I think what actually launches Web 3.0 are the Large Language Models, such as ChatGPT.

You see, I think of Web as content, not technology. Web 1.0 was mostly static content, created by professionals using arcane magic such as HTML and Javascript. Web 2.0 was about people manually writing, laboring their small little blogs in their own little contents of the websphere, loosely connected by technologies such as RSS and Atom and Blogger, without any knowledge of the underlying technology. Web 2.5 was about the consolidation of Web 2.0 onto large platforms such as Facebook, Instagram and TikTok, where content creation and consumption was made as easy as humanly possible.

See where I’m going with this?

ClarkesWorld - a web fanzine that publishes short scifi stories and takes submissions from all over the world - had to stop taking in submissions because of the sheer volume of AI generated drivel.

The number of AI -written books on Amazon that are sold for $1 has exploded.

Your local neighbourhood marketing person is thinking about how to utilize ChatGPT in their messaging.

Even your budding nerdy artist friend who never had the drive to do ten thousand hours of brushstrokes to be a painter is now delving into the depths of Midjourney prompts and plugins to produce art.

The problem isn’t with using AI for these kinds of things. The problem is that everyone who can is doing it. And that leads to a never-before-seen explosion in content on the internet.

This is what I believe to be the essence of “Web 3.0”. First professionals with deep technical skills, then amateurs with some skills, now AIs driven by amateurs with little or no skill.

You thought spam in your inbox was bad? Now prepare for every single meaningful content out there to be completely and utterly drowned by AI-generated noise.

Unless.

Unless - and we of course will, or we leave the internet altogether - we figure out a way to filter the AI spam, much like we do right now. We know some of the tools already from the spam wars - whitelisting, paid access (I think Musk is absolutely right in making the Twitter API paid-only, because he does see the danger of combining AI with Open APIs), algorithmic filtering, trust networks, etc.

But still I fear that a massive amount of the future internet traffic will be AIs screaming to each other - one trying to constantly figure out a way to get its message across, the other trying to block it. And considering that AIs will become self-learning, because it is more efficient than teaching them manually, they will become very, very good at their job. Insomuch that stepping into the open internet in the future will be dangerous to your mental health.

Will that mean that the future is either very distributed, with people again hiding in their small corners with whitelisted participants, trying to keep AI-generated content out; or will it consolidate to an even smaller number of players, who will use their power to strike the right balance between AI-generated content and regular humans?

I think we're going to find out soon.

I fixed a long-standing bug that allowed people to comment on older entries. This was basically an open invite for spammers. It won't keep everyone away, because I did it at UI level and not server level, but since it should keep a lot of the drive-by spammers away. Which are pretty rare these days - it seems that the work that search engines have done to keep link farming at bay, and the fact that personal blogs aren't really that popular anymore and traffic is concentrated on a few high-power sites, are causing the comment spammer to be a dying breed.

Okay, since my CSS knowledge is a little bit outdated, I resorted to talking to ChatGPT about how to make plain CSS to achieve the three-column layout you see here.

It gave me some real nice code for layouts; both using display:flex and display:grid, but I. Just. Can't. Make. These. Work. On. Safari. On. iPhone. Safari on the Mac, no problem. Chrome on the iPhone, no problem. But for some reason it feels like Safari on iPhone is announcing the screen width to be far larger than it actually is, so all media queries go all wonky. I know all of you CSS heroes out there are laughing right now and shaking your heads saying "yeah we know", or "use a framework you dork", but before I go the framework route (which would be a bigger job) I sort of want to understand why this is so.

Any hints?

Okay, figured out the reason - very aggressive caching by Safari. Grid layout works, but I had to learn how to connect my phone to the local dev environment [which was pleasantly easy] to actually debug the issue. A quick cache purge and yes, turns out that I had after all done everything right on my final attempt... Classic rookie mistake there :-D

Okay. So everybody hates distractions. You are working on something important, and you get a message, or a call, or the kid decides to tape your cat to the ceiling fan, or someone at work wanders at your desk and wants to ask a real quick question. Or you distract yourself with an important cat video or a quick peruse at Instagram, and never seem to get anything done.

It sucks. Totally.

But I had a bit of an epiphany today. It was the last class of the improv course I've been taking, and the instructor made us do annoyingly hard things in front of the class, like "you can only use three word sentences" and "you must have a dialogue without ever using the letter 'U'". And they seemed completely arbitrary, but we did those nevertheless because it was all so much fun, but then he said something remarkable: These exercises are designed to distract your conscious mind from the fact that you are performing in front of people by giving it something else to think about. Avoiding a letter, thinking in three-word sentences, using only sounds without meaning... It's all just a ruse and a distraction for the busy mind that keeps worrying about such things like "how do I look" and "am I funny enough".

This, I believe, is an important lesson. Distracting yourself at the right moment is a skill that you can learn and use to a great advantage. We often just distract ourselves without meaning and without the skill by glancing at the phone or just calling; but a skilful distraction can free your brain from a dead end it has driven itself. Some people take walks, or listen to music, or even sleep before making a hard decision, but I think this again is a skill you can develop consciously.

Magicians are known to spend endless hours to practice how to cause unnoticeable distraction in other people. You think you're following what they do, but your attention is exactly always where they want it to be. They know how to draw your attention to the right place at the right moment, and play the fine line between distraction and attention. So if they can do it, we all should be able to distract ourselves at the right moment to free our mind.

Douglas Adams wrote in his Hitchhiker's Guide to the Galaxy, that the secret to flying is to distract yourself at the right moment so that you forget to hit the ground. To me it's always been a very comforting idea, and a demonstration of the power of a good distraction (obviously ignoring the physicist screaming inside me).

I don't know if this makes sense - it's been too long since I've actively written things like this in English, and I feel I'm a bit rusty. Oh well.

Some minor tuning is occurring. I have dockerized the web server (which makes things such as changing the layout a LOT easier since I can test it now locally), and I have made the CSS layout a lot more responsive (thanks, ChatGPT for the help). However I seem to have lost all the attachments, so there seems to be something weird still going on. Will continue a bit later, it's way too late now.

(Well, I did continue. Images are back, but for some reason on Safari on iOS the layout just doesn't work too well. Chrome on mobile works perfectly as designed. Weirdly enough, Safari on desktop is fine as well. Oh well. Have to look into this later. In the mean time, just rotate your phone...)

I had lunch with a colleague, and she mentioned that she had made an MVP (Minimum Viable Prototype) of a "cafe in a van". They had tried it out for a couple of days, but also did branding, printed matching aprons, napkins... The whole works.

I pointed out that it's not an MVP really - they would've figured out if it works even without spending all the time on branding. If you're unknown, and you're just trying stuff out, just shipping the waffles and coffee is sufficient. But she insisted that it was an MVP - because they learned whether they would enjoy it or not. But that's not really what the "minimum viable" implies in the MV part of the prototype.

So I think we need a new word for this kind of experimentation. It's not a "Minimum Viable Prototype" where you validate a business case - it's the "Most Fun Prototype" where you try to simply have as much fun with an idea as you can in a time-boxed environment. See if you and any customers really like doing whatever it is that you want to do!

I think we need a lot more MFPs instead of MVPs. We should all be experimenting with things, having fun with them, instead of trying to achieve "a laser-sharp focus on execution". Life is more than success.

Happy New Year! Today I went to take part of the FLTP test, which is probably the largest electric vehicle consumption test event in the world right now. There were maybe about 70 cars of all makes and models, and pretty much the worst possible weather for an EV - high winds, sleet pouring from the sky, collecting on the road, and +2°C. We drove three rounds each, one highway, one regular road, and one city drive, measured the consumptions and submitted them.

But what I found very delightful was a small moment when I turned around and started driving back to the starting position - I saw an endless stream of electric vehicles on the road. It felt like I had very briefly touched the future - hopefully not too distant one - where most of the cars on the road are electric and we're free of the fossil fuels. Today it lasted only ten minutes or so, but I think the transition in western countries is going to happen faster than people think. The thing is - it's not about replacing _all_ of the fossil cars, it's about driving _as many electric kilometers_ as possible. Since cars pollute little if they're not driven, we're in no hurry to replace the old cars that barely get any distance on them. We need to replace the new ones which are driven the hardest. And that happens fast.

(Took some photos but am too lazy to upload them at this hour. Check out my instagram instead.)