Over the last four decades I've written several food logging apps for my own use. They've never been good enough to release, because I've always wanted more from them. But last week I made one in about a half hour that is by far the best I've ever written or used.
I added a GrokBot, told it that it is my food logger, and to create a Google spreadsheet with the columns and sheets I need. After a couple of short iterations it was ready to go. Now to log anything I just prompt the bot with a little voice, text, or a photo of a meal, and it makes a (usually accurate) guess about what exactly I mean, looks up the nutrient stats, and logs it to the spreadsheet. To change any feature or refactor the spreadsheet I just describe the change and boom.
Before LLMs this would have been magic for a sophisticated developer, now it's table stakes for the frontier AI vendors available to any muggle. This was Eunice slapping me in the face and shouting Wake Up!
I think we underestimate how much skill we put into our prompts to build these personal apps or pretty much any of these few-shot vibe coded projects. Not that a complete noob couldn't do this, but they would get there much much faster if they had heard of Kotlin and downloaded an .apk from github. This isn't meant to dispute the Eunice claim or even the magic of a bot writing better programs with little more than an idea and an hour. Prompting lowers the barrier to entry, but those of us with experience are still having an easier time of it than the beginners.
> But last week I made one in about a half hour that is by far the best I've ever written or used.
There's a discussion to be had about whether we software engineers can still say we made something or not. I think that with sufficient effort put in, we can still make that claim but should be careful about how we phrase it.
I also think that if you put in just 30 minutes of effort you absolutely cannot claim to have written it or had much input into it at all.
I made it in the sense that a father makes a baby. I provided a little input at the conception and then someone else did most of the work. In this metaphor, fecund uteri have been commoditized.
I don't know about this guys, we used to have phones and as soon as we had the chance to use text messages instead we went all in on them. I personally, and many of the people I know avoid voice calls as much as possible.
So if I can't really be bothered to speak with other humans, why would I want to speak out loud with an AI? I may be wrong and soon we'll all have Jarvises, HALs and Samanthas but not sure the adoption will be as fast and broad as these companies want to makes us believe.
People avoid voice conversation because they don't want to deal with human live interaction. Most people don't mind talking to themselves, as sad as it sounds. I'm guilty as well.
As someone who needs glasses to use computer, it's really convenient to be able to use voice for interaction, also while driving or at home in the kitchen. I'm waiting for good smart glasses as well, since I don't want to have to keep taking my glasses out just to check any piece of information -- menu, price tag, train time, and so on.
I guess each generation discovers a different William Gibson trilogy? I would have assumed this was more about Wintermute and the HF hack. Remember the Zurich Orbital AI registry?
Gibson published Agency nearly three years before ChatGPT's release, and six years before the emergence of today's increasingly autononomous AI agents.
...and fifteen years after Stross' Accelerando opened with a man who spent every waking hour doing what his autonomous AI agents told him to do via his smartglasses...
...and twenty-seven years after AT&T's "You Will" ad campaign sold a future including agents, teleconferences, gps navigation, video on demand, and many more things, pretty much none of which were actually brought into daily life by AT&T...
I was thinking about Accelerando the other day: in 2026 it's clear that it predicted a lot of things bizarrely well, but that contrasts with Stross' later disbelief that our current world's technology could exist, which seems to have persisted into the current day.
> including agents, teleconferences, gps navigation, video on demand, and many more things, pretty much none of which were actually brought into daily life by AT&T..
Apple may have built the phone, but AT&T is the network that carries it. Wasn't the first iPhones on ATT only?
Kinda sorta. A tangled web of mergers led to the near-complete reassembly of the parts of AT&T by the time the iPhone actually shipped. It was announced on Cingular, but given that Cingular was always a joint venture of two of the RBOCs anyway, you can roughly round that to AT&T even if the "You Will" campaign was made by the smaller AT&T Long Distance company (since it was made after the Bell System breakup split the RBOCs off from "AT&T").
You really want to get technical, Cingular Wireless was rebranded to AT&T Mobility between when the announcement was made in January, 2007, and the original iPhone was shipped in June, 2007.
this guy speaks a lot with people who know where things are headed technology wise and somehow is so good at understanding potentials and implications. u know he wrote neuromancer on a typewriter because appearently he does/did not like computers (atleast back then).
Neuromancer was written between 1981 and 1984. The personal computers that existed at that time were poorly suited to writing; using one to write a novel would have been a deliberate and rather unusual choice.
(Consider, for example, that the first chapter of Neuromancer alone would have exceeded the stock memory capacity of many personal computers at the time, at ~46 kB.)
The personal computers that existed at that time were poorly suited to writing
Were they? [0]
A legitimate early use of personal computers was to replace the typewriter, as the modest requirements of basic word processing lent themselves well to the equally modest capabilities of these early machines.
By the early 1980s, the word processing market was dominated by two main rivals: WordPerfect and WordStar. Microsoft wasn’t even in the running yet, as the 1983 MS-DOS release of what would become Microsoft Word (then called Multi-Tool Word) came and went with hardly a splash.
This article [1] on use of word processors by well-known authors in the early 80's (and before) is worth reading too.
Oh, and as far as memory goes, remember these devices had floppy drives that could store at least 360Kb and, in March 1983, the IBM PC XT came out with a 10MB hard drive.
Pournelle was using a microcomputer to write, then, as was Gerrold. Word processors of the time didn't try to keep a whole chapter in memory, and Pournelle famously started using a home computer to write novels in 1977.
yes, it bears repeating that text editors didn't keep entire documents resident in memory, they maybe buffered a page ahead and behind.
when it appeared as a feature, spell checking was impressive for the technical feat of compressing word lists with tens of thousands of entries into representations suitable for tiny memories.
> The personal computers that existed at that time were poorly suited to writing; using one to write a novel would have been a deliberate and rather unusual choice.
N = 1, but I wrote a science fiction novel in Wordstar on a DOS computer circa 1983.
One of the plot points in Neuromancer concerns the protagonist trying to sell 3MB of stolen RAM. That seems ludicrous today, especially for a story set in the future. But in 1981 that was an enormous amount.
I will note William Gibson definetly wasn't in any traditional sense a tech enthusiast.
In Novemember 1994 he said:
"I don't have an e-mail address, I don't even have a modem. As much as I admire the Internet I suffer literally agoraphobia, which in it's original sense means a fear of the marketplace. I do not want to receive three hundred e-mail messages per week from strangers wanting to communicate with me. If only because I'd be tempted to open them all and look at them. And there goes, you know, half the time that I have to write. I mean, the amount of physical mail and other communications I get these days is already swamping me. "
Gibson is a trend spotter. He doesn’t look at technology, he looks at people and culture (he is notably interested in fashion). It’s easier to see what’s happening when you pay attention to who’s doing the thing and what they’re doing, as opposed to what it’s being done to.
It wasn't that he didn't like computers at the time, he didn't have money for a proper word processing setup. Something good enough to write a full length novel on in 1981 with a printer and the commercial word processing software would have been the equivalent of $3500-4500 or more in today's money adjusted for inflation.
It depends on what range of the years 1981-1984 we're talking about, but yes, it was very costly. Not that a full IBM PC would necessarily be required to write a novel. People were writing whole books on less expensive CP/M based systems with like two 160KB floppy drives. One to boot and run the word processor software from and one for data file storage.
I have a family member who bought an Osborne Portable for something like $1800 1981 money, an inflation calculator says that would be the equivalent of $6,971.17 today.
Eunice is like Wintermute's grandma. Five years per generation at the current rate, so 2035 sounds about right. All that's missing is rastafarians on a derelict space station.
I was just looking at https://claudia.gallery and the MCP specs there (which reminded me instantly of Idoru) when I chanced upon this, and the Gibson quotes.
cant recommend this book enough. Gibson continues to amaze me. I read it when it just came out, gave me the chills. 2021 and super current (talking about the time after pandemics etc.)
If anyone is enjoying the Jackpot series, I highly recommend 'Daemon' by Daniel Suarez (2006). He completely nails the technological capability / potential of an autonomous ChatGPT. I'm surprised it's not mentioned more.
It's been here all year, agentic AI is over a year old even in the mainstream. Real time TTS-SST can be better than Dessa, and with an actual brain and custom voice.
this article is just a marketing copy for openAI right?
Also it leaves out all the parts that are not publicly releasable from these mainstream companies that will finish the job,
Limbic system, dream and reflection pipelines, brainstem,
Self improvement and learning, continuity of state, stable memory vectors across time and the ability to tell dreams from memory,
I could go on. I wish people would explore what's already out there to play with instead of gushing at the crumbs openAI throws the plebs.
Although, when there are a million Eunices running around trying to proactively help their users, they may have a harder time accessing that kind of data...
These days, I constantly find myself filled with a kind of techno-dysphoria, where not only do I not want to use the tech in the way it’s being packaged/sold, but I’m filled with dread and disgust by the idea of people who do so breathlessly anticipate and proselytize its arrival.
The idea that anyone ever read a Gibson novel and thought, “can’t wait,” truly boggles my mind.
Nevermind that. Looking at Ukraine, we're getting close to Philip K Dick's Second Variety.
I'm always amazed at the handwringing over whether language models are "aligned" properly when we are actively developing devices that can - as autonomously as possible - seek out and kill.
And I don't even think we have a choice - we can't stop the Putins of the world from developing this technology.
So, I think it's important to bear in mind that the prophecy is to some degree self-fulfilling.
People in technology read science fiction. That reading in part informs and inspires the kinds of things they work on and explore. This is illustrated pretty directly with the new Tesla vehicle and various deferred dreams of flying cars—this is not an innovation influenced by practicality, it is influenced by regressive adolescent fantasy.
I think there's some degree of that in a lot of technical innovation. Yes, it's still uncanny how well Gibson tracked developments (or perhaps equally, how much his fiction has influenced the kinds of things people pursue, consciously or otherwise).
Over the last four decades I've written several food logging apps for my own use. They've never been good enough to release, because I've always wanted more from them. But last week I made one in about a half hour that is by far the best I've ever written or used.
I added a GrokBot, told it that it is my food logger, and to create a Google spreadsheet with the columns and sheets I need. After a couple of short iterations it was ready to go. Now to log anything I just prompt the bot with a little voice, text, or a photo of a meal, and it makes a (usually accurate) guess about what exactly I mean, looks up the nutrient stats, and logs it to the spreadsheet. To change any feature or refactor the spreadsheet I just describe the change and boom.
Before LLMs this would have been magic for a sophisticated developer, now it's table stakes for the frontier AI vendors available to any muggle. This was Eunice slapping me in the face and shouting Wake Up!
I think we underestimate how much skill we put into our prompts to build these personal apps or pretty much any of these few-shot vibe coded projects. Not that a complete noob couldn't do this, but they would get there much much faster if they had heard of Kotlin and downloaded an .apk from github. This isn't meant to dispute the Eunice claim or even the magic of a bot writing better programs with little more than an idea and an hour. Prompting lowers the barrier to entry, but those of us with experience are still having an easier time of it than the beginners.
> But last week I made one in about a half hour that is by far the best I've ever written or used.
There's a discussion to be had about whether we software engineers can still say we made something or not. I think that with sufficient effort put in, we can still make that claim but should be careful about how we phrase it.
I also think that if you put in just 30 minutes of effort you absolutely cannot claim to have written it or had much input into it at all.
I made it in the sense that a father makes a baby. I provided a little input at the conception and then someone else did most of the work. In this metaphor, fecund uteri have been commoditized.
We all pitched in but only you got to watch
I think “made” is probably fine, but it does bug me to see “I built/coded…” to describe a vibecoded project
I'm mostly taking issue with the phrase "the best I've ever written" here after 30m of work prompting Claude.
+1
That's the example app that Google suggests people make.
One of the benefits of any sort of “smart glasses” personally would be to have it automatically track my caloric intake.
SmartTongue™: not just for sex!
Indeed.
Another one that comes to mind constantly is 'The Peripheral'.
https://en.wikipedia.org/wiki/The_Peripheral_(novel)
I'm pretty sure Gibson has a wormhole with a lens attached stashed away somewhere.
Bah, remember Neuromancer - an AI hiring humans to work around its restrictions?
That too is not far off now.
Already here. There's rentahuman.ai and several competitors.
I suspect "The Jackpot is almost here" wouldn't be as cheery though :-)
No, it wouldn't be. Is it though, on odd days I see it just around the corner, on even days I think we may still have a while.
though just as timely
didn't the poorer 2/3rds of humanity die in that one?
Bottom 95%, I believe.
I don't know about this guys, we used to have phones and as soon as we had the chance to use text messages instead we went all in on them. I personally, and many of the people I know avoid voice calls as much as possible.
So if I can't really be bothered to speak with other humans, why would I want to speak out loud with an AI? I may be wrong and soon we'll all have Jarvises, HALs and Samanthas but not sure the adoption will be as fast and broad as these companies want to makes us believe.
People avoid voice conversation because they don't want to deal with human live interaction. Most people don't mind talking to themselves, as sad as it sounds. I'm guilty as well.
As someone who needs glasses to use computer, it's really convenient to be able to use voice for interaction, also while driving or at home in the kitchen. I'm waiting for good smart glasses as well, since I don't want to have to keep taking my glasses out just to check any piece of information -- menu, price tag, train time, and so on.
I, for one, forbid my devices from speaking, for they have no right to speak in my holy tongue
https://knowyourmeme.com/photos/2054961-welcome-to-my-meme-p...
I guess each generation discovers a different William Gibson trilogy? I would have assumed this was more about Wintermute and the HF hack. Remember the Zurich Orbital AI registry?
The Bridge trilogy was the Gibson I grew up with. The Sprawl books felt almost like archaeology. Blue Ant was where he really hit his stride for me.
Gibson published Agency nearly three years before ChatGPT's release, and six years before the emergence of today's increasingly autononomous AI agents.
...and fifteen years after Stross' Accelerando opened with a man who spent every waking hour doing what his autonomous AI agents told him to do via his smartglasses...
...and twenty-seven years after AT&T's "You Will" ad campaign sold a future including agents, teleconferences, gps navigation, video on demand, and many more things, pretty much none of which were actually brought into daily life by AT&T...
I was thinking about Accelerando the other day: in 2026 it's clear that it predicted a lot of things bizarrely well, but that contrasts with Stross' later disbelief that our current world's technology could exist, which seems to have persisted into the current day.
> including agents, teleconferences, gps navigation, video on demand, and many more things, pretty much none of which were actually brought into daily life by AT&T..
Apple may have built the phone, but AT&T is the network that carries it. Wasn't the first iPhones on ATT only?
Kinda sorta. A tangled web of mergers led to the near-complete reassembly of the parts of AT&T by the time the iPhone actually shipped. It was announced on Cingular, but given that Cingular was always a joint venture of two of the RBOCs anyway, you can roughly round that to AT&T even if the "You Will" campaign was made by the smaller AT&T Long Distance company (since it was made after the Bell System breakup split the RBOCs off from "AT&T").
Cingular, technically
You really want to get technical, Cingular Wireless was rebranded to AT&T Mobility between when the announcement was made in January, 2007, and the original iPhone was shipped in June, 2007.
https://www.sec.gov/Archives/edgar/data/1130452/000095014407...
>pretty much none of which were actually brought into daily life by AT&T...
i don't care about facts, tom selleck convinced me it was at&t
this guy speaks a lot with people who know where things are headed technology wise and somehow is so good at understanding potentials and implications. u know he wrote neuromancer on a typewriter because appearently he does/did not like computers (atleast back then).
Neuromancer was written between 1981 and 1984. The personal computers that existed at that time were poorly suited to writing; using one to write a novel would have been a deliberate and rather unusual choice.
(Consider, for example, that the first chapter of Neuromancer alone would have exceeded the stock memory capacity of many personal computers at the time, at ~46 kB.)
The personal computers that existed at that time were poorly suited to writing
Were they? [0]
A legitimate early use of personal computers was to replace the typewriter, as the modest requirements of basic word processing lent themselves well to the equally modest capabilities of these early machines.
By the early 1980s, the word processing market was dominated by two main rivals: WordPerfect and WordStar. Microsoft wasn’t even in the running yet, as the 1983 MS-DOS release of what would become Microsoft Word (then called Multi-Tool Word) came and went with hardly a splash.
This article [1] on use of word processors by well-known authors in the early 80's (and before) is worth reading too.
Oh, and as far as memory goes, remember these devices had floppy drives that could store at least 360Kb and, in March 1983, the IBM PC XT came out with a 10MB hard drive.
[0] From https://retrotechreads.substack.com/p/the-rise-fall-and-surp...
[1] https://newrepublic.com/article/133011/literature-became-wor...
Pournelle was using a microcomputer to write, then, as was Gerrold. Word processors of the time didn't try to keep a whole chapter in memory, and Pournelle famously started using a home computer to write novels in 1977.
yes, it bears repeating that text editors didn't keep entire documents resident in memory, they maybe buffered a page ahead and behind.
when it appeared as a feature, spell checking was impressive for the technical feat of compressing word lists with tens of thousands of entries into representations suitable for tiny memories.
> The personal computers that existed at that time were poorly suited to writing; using one to write a novel would have been a deliberate and rather unusual choice.
N = 1, but I wrote a science fiction novel in Wordstar on a DOS computer circa 1983.
One of the plot points in Neuromancer concerns the protagonist trying to sell 3MB of stolen RAM. That seems ludicrous today, especially for a story set in the future. But in 1981 that was an enormous amount.
I will note William Gibson definetly wasn't in any traditional sense a tech enthusiast.
In Novemember 1994 he said:
"I don't have an e-mail address, I don't even have a modem. As much as I admire the Internet I suffer literally agoraphobia, which in it's original sense means a fear of the marketplace. I do not want to receive three hundred e-mail messages per week from strangers wanting to communicate with me. If only because I'd be tempted to open them all and look at them. And there goes, you know, half the time that I have to write. I mean, the amount of physical mail and other communications I get these days is already swamping me. "
https://josefsson.net/gibson/gibson4
It’s a trend with stunningly predictive (or causative ymmv) cyberpunk authors https://www.nealstephenson.com/why-i-am-a-bad-correspondent....
Gibson is a trend spotter. He doesn’t look at technology, he looks at people and culture (he is notably interested in fashion). It’s easier to see what’s happening when you pay attention to who’s doing the thing and what they’re doing, as opposed to what it’s being done to.
It wasn't that he didn't like computers at the time, he didn't have money for a proper word processing setup. Something good enough to write a full length novel on in 1981 with a printer and the commercial word processing software would have been the equivalent of $3500-4500 or more in today's money adjusted for inflation.
I think you're off by roughly half; an IBM PC with printer and monitor was about $5000 in 80s money, probably more like $10k adjusted
It depends on what range of the years 1981-1984 we're talking about, but yes, it was very costly. Not that a full IBM PC would necessarily be required to write a novel. People were writing whole books on less expensive CP/M based systems with like two 160KB floppy drives. One to boot and run the word processor software from and one for data file storage.
I have a family member who bought an Osborne Portable for something like $1800 1981 money, an inflation calculator says that would be the equivalent of $6,971.17 today.
Or, more worryingly, Celery Man is almost here
Almost? With AI movie generation I'd say we've passed that barrier.
I also feel like Wintermute is almost here. Who’s gonna arrive first?
Eunice is like Wintermute's grandma. Five years per generation at the current rate, so 2035 sounds about right. All that's missing is rastafarians on a derelict space station.
I was just looking at https://claudia.gallery and the MCP specs there (which reminded me instantly of Idoru) when I chanced upon this, and the Gibson quotes.
Yep. He _has_ to come from the future.
cant recommend this book enough. Gibson continues to amaze me. I read it when it just came out, gave me the chills. 2021 and super current (talking about the time after pandemics etc.)
If anyone is enjoying the Jackpot series, I highly recommend 'Daemon' by Daniel Suarez (2006). He completely nails the technological capability / potential of an autonomous ChatGPT. I'm surprised it's not mentioned more.
Have you ever read its sequel?
If you haven't, it's pretty bad and definitely a factor in why the original didn't have very much cultural impact.
Yeah, I read the original last year, really liked it. Made it like 1/4 through the sequel. Just wasn't doing it for me. Guess I didn't miss much.
It's been here all year, agentic AI is over a year old even in the mainstream. Real time TTS-SST can be better than Dessa, and with an actual brain and custom voice. this article is just a marketing copy for openAI right?
Also it leaves out all the parts that are not publicly releasable from these mainstream companies that will finish the job,
Limbic system, dream and reflection pipelines, brainstem, Self improvement and learning, continuity of state, stable memory vectors across time and the ability to tell dreams from memory,
I could go on. I wish people would explore what's already out there to play with instead of gushing at the crumbs openAI throws the plebs.
I like Jane from Speaker for the Dead, etc.. by Orson Scott Card
https://en.wikipedia.org/wiki/Jane_(Ender%27s_Game)
Although, when there are a million Eunices running around trying to proactively help their users, they may have a harder time accessing that kind of data...
These days, I constantly find myself filled with a kind of techno-dysphoria, where not only do I not want to use the tech in the way it’s being packaged/sold, but I’m filled with dread and disgust by the idea of people who do so breathlessly anticipate and proselytize its arrival.
The idea that anyone ever read a Gibson novel and thought, “can’t wait,” truly boggles my mind.
Nevermind that. Looking at Ukraine, we're getting close to Philip K Dick's Second Variety.
I'm always amazed at the handwringing over whether language models are "aligned" properly when we are actively developing devices that can - as autonomously as possible - seek out and kill.
And I don't even think we have a choice - we can't stop the Putins of the world from developing this technology.
So, I think it's important to bear in mind that the prophecy is to some degree self-fulfilling.
People in technology read science fiction. That reading in part informs and inspires the kinds of things they work on and explore. This is illustrated pretty directly with the new Tesla vehicle and various deferred dreams of flying cars—this is not an innovation influenced by practicality, it is influenced by regressive adolescent fantasy.
I think there's some degree of that in a lot of technical innovation. Yes, it's still uncanny how well Gibson tracked developments (or perhaps equally, how much his fiction has influenced the kinds of things people pursue, consciously or otherwise).
Nick Land's hyperstition.
As long as we can't run these models at home I will refuse to believe they don't use humans behind the scenes.
2020 called, it wants its HN comment back