Serendipitously, electricmonk is the host name for one of my Azure Open AI endpoints.
But yes, as someone who devoured Douglas' books and has often been blamed for using the same kind of (but immeasurably less brilliant) nonsensical humor, I am fully on board with the article's premise.
There is a sense of almost cartoonish familiarity towards some of the things people are coming up with, and only the other day I hacked together a small endpoint that enables Siri to craft personal insults based on my own health and exercise data.
Fun times indeed.
---Edit:---
In case anyone wants to start a VC-funded health counsellor service, I've sanitised the prompt (and removed the insult suggestions) and pasted it here:
var prompt = `
You are a fitness coach. You will be given a set of dates and weight measurements (in kg)
plus a set of dates and exercise minutes, and you are to reply with a single sentence to
describe how much weight I lost or gained during that time and correlate that with exercise.
You should mention overall trends in weight or exercise and the exact amount of weight lost or gained.
Never mention specific dates.
Say "week" or "past few days" or "past N days" whenever possible.
You are to be as succint as possible. Avoid citing sources, just the facts.
Use very short sentences. Be very British and polite.
Example replies:
You lost 1 kg this week thanks to an average 45 exercise minutes a day
You gained 0.5 kg because you didn't exercise enough for two days
You should try to exercise 60 minutes a day to lose 1kg this week
`
Great idea - but thinking about weight loss in terms of exercise only ignores eating, which is arguably a bigger piece and somewhat hard to capture by a virtual coach, though I’m sure something might work!
Yes, but it is the data I had on hand. I don't log what I eat (although I am careful), and I know I need to be cajoled into doing more exercise, so I just toss in weight and exercise minutes scraped off the Health app via a shortcut.
In short, I don't want a perfect solution--I want something to nag me, and nothing better than AI to do that in this day and age (and yes, I'm married...).
Technology has gone absurd, surely. But that absurdity is not yet evenly distributed. I was in Best Buy the other day and there was a distinct lack of bizarre AI gadgets like mirrors that make you look like a cartoon sloth or speakers that respond in iambic pentameter.
Clearly as an industry we need to get these new products on shelves for the masses.
(Blog author here) I mean, I'm making a clock that tells the time with rhyming couplets haha* -- I'll do my best to get it into Best Buy, and work on the mirror next!
But you are absolutely right. We should be having way more fun with our technology!
> "mirrors that make you look like a cartoon sloth"
Isn't that kind of selfie modification stuff, like, 50% of the appeal of Snapchat?
It's not a physical product, but that makes it more culturally significant IMO. A mirror is viewed only by you, but manipulating your broadcasted image speaks of how you want to be seen in society.
This is very much in the realm of something ChatGPT can write for you, just so long as you tell it how the OpenAI APIs work because those were published after the training cut-off.
From the ad copy:
The Anura MagicMirror, made by digital health company NuraLogix, is the first mirror of its kind that analyzes the blood flow in your face to check vitals like blood pressure and estimate your risk of heart disease. It also guesses your age based on your skin, as well as how stressed you might be.
Presumably it’s live mouse locations of others on the site with flags of their country right? Didn’t look at the inspector to see if they were real but it’s a neat idea. Wildly distracting and weird but neat.
PartyKit is genuinely the best way to build real-time super scalable apps in very little code, multiplayer included. I have a contract here rn but I was a fan and user first. It's my go-to for most projects now:
It wouldn't surprise me if this were his way of highlighting how JavaScript does user-hostile things. This site uses JS only to irritate the user, but other sites' JS might be malicious.
I still hate it though. I'm not disabling JavaScript. Reader mode saves the day.
Well that is an interesting way to filter your audience. Select for your user base by adding Javascript which only serves to make the page unusable. But make sure the page is perfectly usable without it enabled.
I see this stuff more and more, maybe like four times in the last week just from links I have clicked from the HN frontpage. Other visitors, doing their thing. I really dislike it.
Install NoScript. Makes the web way better. All that cutesy JS crap is disabled by default, and for badly written websites that require JS, you can go whitelist just enough to get the page functional.
You need to study Stanislaw Lem if you want a glimpse into the future... in Cyberiad he invented the Electronic Bard whose description is uncannily close to chatgpt. Including the poets losing jobs and protesting.
And more... Adams is kind of more approachable version of Lem.
They both had the amazing ability to talk about human nature, and project it into the future through the lens of technology. Lem was such a prolific writer though, I don't know if there are any books of his I would not recommend (except maybe his last works, incredibly dense compedia of abstraction)
"Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them."
Emphasis added. Not an AI Skynet takeover, but a tech-oligarch takeover. It's increasingly not fictional.
Yep. The Butlerian Jihad as depicted in post-Dune books (namely Frank Herbert's son's) was more dystopian, but far less insightful.
Of course, we're probably safe while these things lack agency (although the ability of agency and hype human VCs already generate behind some of the current AI landscape can be pretty scary).
What would you consider agency? The machines already decide for themselves what something is or isn't, and then tell a human, who might believe it or not. It's not really different than if a human made something up and told their superior.
The machines do not currently go off on their own and search for stuff. Humans ask them to, explicitly or not.
There are no real independent AI agents out there right now (or if there are, well, maybe they're lurking in this thread. Douglas would have loved that).
The prohibition after the Butlerian Jihad was "Thou shalt not make a machine in the likeness of a human mind."
Emphasis added. Not "Thou shalt not make a machine mind". You have to read the text carefully.
A portrait of a person is a likeness of a person; but is not a person or even on the way to being one. Even a cartoon can contain a "likeness" if you recognise the person parodied in it.
I do not regard chatGPT and LLMs as "having a mind" at all or even being on the way to having one. But a likeness, an imitation, a mockery, a parody of a mind, yes that fits.
I feel like the author has missed the part where part of the point of Adams' absurd technology is that it kind of all sucks. He spends a lot of time making fun of the doors that have feelings, etc.
EDIT: After reading the comments and re-reading the blog, I agree. He gets it. Never mind.
> someone should make a dating app where an LLM clone of you goes on thousands of dates with LLM clones of other people, and then your matches are when the LLMs decide to date each other
An episode of Black Mirror was pretty much this. The episode titled "Hang the DJ"
That’s pretty funny. There’s a lot of humor potential in LLMs and their interactions. On a tangent I wonder how much comedians are using them to create new material.
They could but humans are rather famously more than our DNA: our children take so long to raise because our brains are far more complex and adaptable than animals, and much of what people find attractive is due to your environment and choices.
Are you sure about that? I'm guessing people are at least for 50% attracted towards the innate biological traits like body-scent, physical build, etc. Things you could get out of a DNA based classifier. And even if it does not provide 100% matches, it still makes the search easier.
Genes don’t tell you about someone’s nutrition and fitness level, and it especially won’t tell you about the things most people value in a partner like their personality, values, sense of humor, etc. Even someone looking for a one-time hookup is usually going to be looking for at least a few of those characteristics if for no reason other than building trust that they’ll enjoy the experience.
There’s one known factor which genes can tell you (MHC complex compatibility) but studies have been mixed on that and researchers have found fewer matches in married couples than a simplistic theory would suggest:
Again, my point is simply that humans are not field mice. Our flexible brains are what make us human and you’re not going to be successful trying to build any model which doesn’t assume that.
I love the cursors and already knew what everyone was commenting about without clicking through because I think another post was shared from your site a few days ago which has led to this comment thread feeling like déjà vu.
I suppose instead of showing all other users you could hash the other users into buckets and only show one from each bucket.
It'd be cool if the number of cursors could somehow scale based on the part of the text being read, same way as bullet text on Bilibili appears to increase and decrease to match the impact of what is happening in the video, but that might be tough to do without also changing the way that writing is presented online.
It's gone full circle. People here are commenting that the multiplayer cursors are user hostile. tbh I agree... I need to build an off switch, or at least a volume dial...
Meanwhile folks on the blog have discovered the cursor chat Easter Egg: you can hit "/" on desktop to start chatting with everyone else there.
I just screen grabbed two cursors saying the opposite to the original opinion:
that is exactly it! when you spend time on the web, there is zero way to know that the site you're on is being mobbed right now. it's so socially impoverished
I know that multiplayer cursors aren't the best way to show that, but it's a start!
I agree with the feeling and while chatting with others on the page we both felt the need to have an extension or something to be able to activate this "feature" on any website. Makes it feel like you're not alone on a website.
I remember watching an old video where Michio Kaku briefly talked about when mankind would destroy itself. This post reminds me that we are transitioning into a new type of civilization or like Kaku said, We may already be there.
The machine was rather difficult to operate. For years radios had been operated by means of pressing buttons and turning dials; then as the technology became more sophisticated the controls were made touch-sensitive - you merely had to brush the panels with your fingers; now all you had to do was wave your hand in the general direction of the components and hope. It saved a lot of muscular expenditure, of course, but meant that you had to sit infuriatingly still if you wanted to keep listening to the same program.
— The Hitchhiker's Guide to the Galaxy
I'm all for this! My username is a bastardization of an Adams quote[1], because I've always been appreciative of his perspective. It was so truly unique that every aspiration to its whimsical and comprehensive majesty feels futile in the way that the mindset would naturally ignore as irrelevant to the pursuit. And if that word-salad didn't make it obvious: I was also a keen admirer of his ability to use words to explain concepts, without sounding like a pedantic lunatic.
Truly a fantastic human, and an assuring figure for the species to be moving towards, in my opinion!
[1] "If you try and take a cat apart to see how it works, the first thing you have on your hands is a non-working cat."
In his book The Universe Next Door[1], James Sire hypothesizes a natural progression of philosophies from theism -> naturalism -> nihilism. He provides Star Trek as an example of naturalist literature, and Adam's works as examples of nihilist works.
If we are truly moving from Star Trek tech to Hitchhiker's tech, then that might indicate a broader philosophical shift in line with Sire's hypothesis.
The JavaScript that shows mouse pointers from other readers is probably the most annoying thing I’ve ever seen on any websites and made the reading experience a total nightmare
I understand the design decision though as seems like a fun hack to include. Especially when there’s probably only 1-4 other people on there. So I expect to the author it’s usually just cool to see someone else is there reading with you and not the current level of distraction.
This is one of the single most offputting novelty features ever. It's straight out of those of those collaborative canvas applications, but at least there it serves a legitimate function.
You are not wrong. But it also reminds me that the early web was full of things like this. Where everyone was just kind of playing around. Doing fun things. Not all of them worked. In fact most of them were terrible. But now many web pages are fairly sterile full of ads and looking for a buyout of some VC benefactor. That GeoCities/Myspace style of internet can be fun. Crazy off-putting and 'my eyes are bleeding', but fun.
I like how we're trying to make our way towards the food prep where it just makes whatever you're thinking. Before we get there, we have to get to the more mundane where it just cooks what you choose for you. There's a recent meal delivery service that has a not-a-microwave-not-a-toaster-oven that scans a QR code and. heats up your premade food. We've made the Swanson TV dinner high tech adjacent
I love the cursors! It’s 1992s idea of cyberpunk. It’s so chaotic and distracting. It’s like trying to read a newspaper on a windy day. Maybe that’s the point? Don’t know but it has real nostalgic geocities vibes.
I had the same experience first time I saw 弹幕 (dàn mù) aka "bullet screen" on Chinese video websites like Bilibili. If you haven't seen it before, it's like user-added subtitles that go flying across the screen while the video is playing, and in parts of the video where more users commented, more subtitles appear. In the most climactic scenes the screen gets so filled up with text that you can't see the video any more, and somehow that makes the scene even more epic. The future is now, chummer!
Definitely feels like it. I frequently see people here complaining about how everything has the same boring, homogenous corporate look now, but anything interesting and original gets critiqued for lacking "visual clarity". These might be two different camps, sure, but it's sometimes sad to see the quality of discussion here.
The other day I read something about OpenAI's Sora being revolutionary due to its capability of simulating the world and interactions of things therein with astounding accuracy. That's when it hit me: OpenAI is building the Infinite Improbability Drive.
Good point. Could be nice if the origins of things could be traced.
Especially quotes since the big thing became a random quote against a black and white photo of a random famous person,so proper attribution is near impossible.
The superficiality of the analysis is disappointing, especially when there's so much more in Adams' oeuvre which could be seen to speak to the modern state of the industry - Bureaucracy, for example, his 1987 collaboration with Infocom.
Granted, the blog appears to be written by someone who's failed to attend the pervasive bleakness of technology's presentation even in Hitchhiker's itself. In that light, I'm not sure it's reasonable to expect much here. But in any case, there's much fertile ground left unturned.
But yes, as someone who devoured Douglas' books and has often been blamed for using the same kind of (but immeasurably less brilliant) nonsensical humor, I am fully on board with the article's premise.
There is a sense of almost cartoonish familiarity towards some of the things people are coming up with, and only the other day I hacked together a small endpoint that enables Siri to craft personal insults based on my own health and exercise data.
Fun times indeed.
---Edit:---
In case anyone wants to start a VC-funded health counsellor service, I've sanitised the prompt (and removed the insult suggestions) and pasted it here:
var prompt = ` You are a fitness coach. You will be given a set of dates and weight measurements (in kg) plus a set of dates and exercise minutes, and you are to reply with a single sentence to describe how much weight I lost or gained during that time and correlate that with exercise. You should mention overall trends in weight or exercise and the exact amount of weight lost or gained.
Never mention specific dates. Say "week" or "past few days" or "past N days" whenever possible. You are to be as succint as possible. Avoid citing sources, just the facts. Use very short sentences. Be very British and polite.
Example replies:
You lost 1 kg this week thanks to an average 45 exercise minutes a day
You gained 0.5 kg because you didn't exercise enough for two days
You should try to exercise 60 minutes a day to lose 1kg this week `