We've been playing with chat gpt and gpt-3 in recent days and weeks to refine our sales pitch, generate variations of text for our website, etc. It saves a lot of time when you can just take your rough notes and turn it into a perfectly smooth and readable paragraph instantly. It does not replace what we do but it does speed up the process enormously. We can iterate faster on our drafts, use AI generated mockups to enliven our pitch, etc. We've barely scratched the surface of what we can do with this.
There are lots of people obsessing about feeling threatened in what they do by all this. It's an understandable response but not necessarily a productive attitude. It's really just an opportunity to reconsider what it is you do exactly and become better and more efficient at that. Perhaps if you spend most of your time trying to produce slides, designs, etc., freeing up some of that time for more high value activities is not such a bad thing and something that you should have been striving to do anyway. In our case, using our refined pitch to drive sales is what we should be doing. Spending time to refine the pitch is a necessary evil and not the main thing we should do or even want to do. So, this is an enabler not a threat for us.
I'm actually a CTO; so me spending time on sales pitches is not the most obvious use of my time. But in a small startup, if something needs doing, I probably need to be involved. So, anything that makes that more efficient is more than welcome.
> Perhaps if you spend most of your time trying to produce slides, designs, etc., freeing up some of that time for more high value activities is not such a bad thing
Maybe, but if doing that was my full time job then I would instantly feel less valuable and less secure.
I just don't buy the argument that AI/automation frees up time to do more valuable work at the individual level. Sure it's great for the organization. But if your job became obsolete tonight, simultaneously with thousands of other people, what "high value activities" would you do to earn money? I'm not saying that has happened but the day may be coming for us "knowledge workers".
The trend here is that knowledge work is becoming less valuable. Not everyone will be able to rise above the machines in this category.
Luckily we can probably still clean houses for a living, since it appears that robots will be able to write good code before they can wash dishes or make a bed.
It still astonishes me how much this mindset of "everyone needs to work as much as humanly possible regardless of how much stuff actually needs to be done" is baked into our brains.
Maybe the thing start doing when new technologies come along is ask "how can we rearrange ourselves so that now everyone works less time than they did before?"
Many desired goods are scarce, people will compete for scarce goods. There is a market clearing price.
To some the price is acceptable, to others the price seems like it forces them 'to work as much as humanly possible regardless of how much stuff actually needs to be done'.
People can have different desires. Nothing confusing in there.
Not meaningfully, from both directions at the same time. Most of the of goods we consume are produced with large quantities of mechanical labour and have been since the Industrial Revolution got going; conversely, one of the biggest costs we face in the G7 is the land, and the land is mainly expensive because of the other people all bidding against each other to be in good locations.
Visited Nairobi a few years back, and one of my partner’s local friends was complaining about the rent (one year for her in Nairobi was less than one month for me in Cambridge, because it was locals bidding against other locals for rent in both cases), but that friend was still able to get a smart phone (albeit a cheaper model than most people here would be OK with).
I’m a solopreneuer and I’ve been using it for everything from Upwork job ads to creating Bootstrap templates and React functions. It’s literally like an all-knowing assistant. I’ll happily pay good money for it.
I meant quick scaffolding, not the styling (though it can probably help with that too). Say, asking chatGPT to create a bootstrap template with two columns and three cards in the main container. Or asking it to create a card component with an icon in the center and heading text.
I can find the documentation online but its just much faster to get chatGPT to do it.
I definitely agree with the sentiment about GPT. It's a really powerful tool that can produce some really impressive outputs, but of course, it does have its limitations.
One of the things I really like about it is that it's not always perfect, yet still provides some really useful texts. It usually takes me a few tries to get exactly what I'm looking for, but it's definitely worth it. I've had a lot of fun using the generator to reimagine classic stories with different characters (I loved Cindirella one in but like in Blade setting).
Even with the help of the model's statistics, I still feel like I have a lot of ownership over the text that I create with the generator. It's kind of like ordering from a fast food restaurant instead of cooking from scratch - you still have certain expectations based on the statistics, like food safety or meat content, but you don't have to worry about all the details.
Having a tool like this to assist me with my writing is really appreciated. I can be a bit chaotic when I write, so having a tool that helps me organize my thoughts is a huge help. I can't wait for the text generator to be commercialized so I don't have to worry about timeouts anymore.
The novel thing with ChatGPT (besides the language model) is the interface and how flexible the informal question answer interface is.
I’m trying to imagine what other interfaces could expose value from these existing AI tools value simply by making it easier and as low effort as possible.
Similar to how GitHub Copilot/Tabnine automatically suggests completions as unobtrusively as possible without requiring a lot of manual intervention. Just as an extension.
Once we have a google glasses type augmented reality this sort of light automated assistance can be fed automatically into our lives. Just by hearing or looking at something it could turn it into an automated prompts and be unobtrusively suggested or visually fed into our lives. Auto-translation of foreign languages is the obvious AI example with AR.
Maybe other interfaces such as direct MacOS integration the way the smart search box/Alfred style works or browser extensions.
We just use the openai freemium layer like everybody else. So, cost for this is so far zero and we are definitely saving time and effort. We'd probably end up paying for this if this starts showing up in mainstream products for creating content. Our CEO actually pays for midjourney, which he uses for mockups and other things. I expect this will start happening pretty soon as there probably are a lot of clued in business people discovering just how useful this is.
You get some free usage, and then pay per token after that. Once you've done that you can play with chatGPT for free, as of now. It does time out fairly frequently since so many people are playing with it.
AI permeated pretty much every business years ago with the advent of machine translation. It absolutely decimated the translation industry, because people who want translation work done rarely have the skills necessary to tell the difference between the bad-but-free machine translations and the good-but-not-free human translations. Everyone loses - there are fewer translators now, the ones left find it harder to get work, and most people have to put up with understandable-but-mostly-terrible translated text.
AI is technically amazing, but the short term impact is going to be to lower the quality of basically everything you
can generate. People will struggle to compete on cost, and if the buyer doesn't know and appreciate the difference in quality they'll never buy from a person when they can generate something instead.
I never really thought AI would be a serious threat to my job, but I'm starting to reconsider that now.
I wouldn't say everyone loses. Being able to use foreign language websites and read documents is incredibly useful, and they don't need the world's best translation - although I would say the quality has been pretty decent recently. I'm sorry for translators but they're vastly outnumbered by people who find machine translation useful.
If the choice is between finding a translator and paying them to be able to read something at all vs immediate free translation with the occasional mistake, I know which one I find more useful.
It's very, very surprising to me that that person has found a way to say that "everyone loses" with machine translation.
If they wanted to look at a foreign website 30 years ago, what did they do? Call up a human translator and get them to translate? No, we just didn't read foreign language websites.
For me, this is on par with arguing that English to foreign language dictionaries are bad, it's just ridiculous.
Sorry if this comes across as too aggressive, I just come across this backlash against technology and labour-saving devices somewhat often, and it really confuses me.
If you look at machine translation in isolation you might actually be right. Perhaps we don't really need accurate translation at all, and translation jobs won't be missed, and having more-but-worse translation is a benefit of AI. But my point was more about using translation as an illustration of the impact of AI on an industry. In the case of translation there are maybe 5 or 10 percent of the translation jobs today than there were 20 years ago. Machine translation has decimated the translation industry. People are happy to have bad translations in order to save money. That is a pretty indisputable fact.
My point is that this is likely to be the case for everything that you can generate with AI.
If people will accept a slow, buggy, shitty app generated with AI because it saves them money then there is every chance that there could be 90% fewer devs in a couple of decades than there are now.
It's hard to accept as a dev. I'd wager that if you'd told a translator back in 2002 that a computer will replace them in two decades they'd have had a hard time accepting it too.
>No, we just didn't read foreign language websites
I don't understand. You're just denying that, before machine translation worked well, some people translated their websites into more than one language? Not many, but some did.
The complaint is that automatic translation hasn't just added translations for those who couldn't afford knowledgeable people, but replaced some of the good translations with bad ones, at the same time driving people out of that kind of work.
I guess it seems at least superficially plausible to me because of the devastation wreaked on journalism by ad-tech.
Maybe the economic net benefit correlates to a societal net benefit, but that doesn't mean everybody has to be happy about the losses.
> I don't understand. You're just denying that, before machine translation worked well, some people translated their websites into more than one language? Not many, but some did.
No, I'm not. If the website was in a foreign language, I couldn't read it. If it had been translated, it was no longer only a foreign language website.
> Maybe the economic net benefit correlates to a societal net benefit, but that doesn't mean everybody has to be happy about the losses.
The reason this confuses me is that it seems extremely obvious that there is a massive net benefit. I can actually read foreign news from even the smallest website which never had the resources to translate. I can read foreign forums and subreddits discussing local politics - there's no universe in which anyone would've hired anyone to translate forum posts...
It's just patently ridiculous that machine translation might NOT be a net benefit.
It's just patently ridiculous that machine translation might NOT be a net benefit.
Ok. Let's assume that you're right, and apply the same logic to AI writing code:
Everyone gets free access to an AI that can write buggy-but-kind-of-working code for an app they can describe in plain English. If someone wants a basic app, for any reason, they just have to go through a series of prompts and they get something deployable out the other side. This democratizes coding, it enables and empowers millions of people, and it means anyone who wants an app of any sort can have it basically for free. Yay!
The downside is that the development industry shrinks by 90%.
That could be a net benefit for society, and I have no doubt people who want an app right now but can't afford to pay devs would absolutely think it is. I don't think you can assume cheaper access to worse products is necessarily a good thing though. There are downsides, like the consolidation of service provision into the hands of a few companies. For example, if you want translation today and you don't have a ton of money you're pretty much stuck with what Google will give you, with all of Google's biases.
30 years ago was 1992 and the web was already a year old, so there were already a few websites, including some in Hebrew and French (although most were physics-related). Usenet itself is going to be 43 years old next year!
If the choice is between finding a translator and paying them to be able to read something at all vs immediate free translation with the occasional mistake, I know which one I find more useful.
Yes, and that's how people will think about everything that can be generated by AI, from images to writing to music to code. People are very happy to accept a cheap-but-worse alternative. That that thought and apply it to what you do, and you'll see what the problem is.
If the corpus of verified, good translations comes down hard then the ability of machine translation to improve much beyond its current state or to deal with language changes will be limited.
The issue with things like machine translation is how to compensate the humans that fund and create the translations going into the machine - they are not redundant.
Also the translation really depends on the corpus. EU legislation and such is big input. But I'm not sure how well that covers many other cases. Or prose for example.
Chinese/Korean->English is readable, but actually pretty horrid for artistic works. Not something you really want to read for pleasure.
Super interesting point, though it's important to note that this extends beyond text (which everyone is understandably focused on after ChatGPT's release).
AI "art" (image gen) is also becoming more popular and lots of interesting business applications around marketing images and product listing images. One thing to note is if you look at something like Amazon where many listings have terrible images, using AI could actually improve them straight away by using some kind of Stable Diffusion / DreamBooth-type model with an API like https://dreamharbor.ai
For more generic images (e.g. for blog posts) where you just want "a dog sat on a cloud" then places like https://lexica.art let you produce these images for free. The key with both of those is learning prompt engineering for which there are tonnes of resources - sites like https://prompthero.com/ are already getting involved in this space and on that dreamharbor site there's a pretty detailed blog post: https://blog.dreamharbor.ai/blog/how-to-write-prompt-stable-...
And then when it comes to video Synthensia is doing some very exciting work - but still some way to go. They could actually enhance international growth by providing better dubbing (potentially).
> [machine translation] absolutely decimated the translation industry
Do you know of a good source for this? I've heard claims of this, but also claims of the opposite ("free MT leads to more communication across languages which leads to more need for high-quality translation"). But I've never seen any studies or hard numbers (or even anecdata from translators who lost their jobs – though I do know of translators who turned into MT post-editors), and my google scholar searches are turning up nothing.
I know a freelance translator who basically gave up their work because (1) a steady decline in requests from clients and (2) it just doesn't pay enough anymore unless you rush every job by using DeepL and cutting corners on quality.
Only a few hours ago I found myself in front of a bad translation, accompanied to the product of a billionaire (who published a video in multiple versions, with translated embedded infographic text). A locution was employed which is used for something very different, and unironically (e.g., "With the departure of the CEO the company has lost his head" - to try and give the idea).
John Deere saw this coming a long time ago; they've locked down the data their tractors generated with iron bars. Lots of previous HN discussions.
Now they make bank based on that data, they use their customers' millions of hours of driving tractors to train self-driving tractors. They can do it not because of fancy AI, but because they spent years raising abundant task-specific data. It's genius: the customers pay John Deere to create data for John Deere, which John Deere uses to get more money from other customers.
I think you're mixing two very separable behaviors. They're using data from their tractors to train AIs (1), and they're locking the data away from their customers, including from their customers' own tractors (2).
The tractor is still driving on a pre-programmed GPS route.
To avoid a semantics debate, I'll agree for a brief moment the AI is "driving" the tractor to avoid an object and get back on the pre-programmed route.
Self driving encompasses a bunch of things. Obstacle detection is one. Vehicle control, including micro adjustments, is one. Path planning is another one.
GPS is one way of doing one part of self driving. But it's ALL self driving
A pre-planned route != self-driving (in my opinion).
I say that because it's not driving itself, it's driving a route you plotted or recorded.
You can't take that same tractor and have it work on your neighbors pasture as-is, it has to be programmed.
But again, this is getting into a semantics debate. Personally I'd prefer to differentiate using autonomous and automated, self-driving is too ambiguous.
By that logic, am I driving the car while following the route plotted and dictated to me by Google Maps? What if it's all over two-lane freeways, so all I have to do is stay on track, not crash into the vehicle in front of me, and follow Google's instructions? And what if I then turn on adaptive cruise control? Am I still driving the car?
If I’m driving a car, and hit a brick wall that someone’s put in the middle of the road, it’s inarguable that this is a failure of my driving. There are no hairs to split here.
Agreed. It will always be pre-programmed GPS because it’s the only way for a tractor to get an absolute position and heading without a star tracker and a sophisticated dead reckoning system (expensive gyros and accelerometers) or an electronic perimeter.
Tractors don’t sit idle all of the time, the owners swap out tractor attachments so their “user experience” depends on handling the transition between manual driving and automatic operation, even if that UX is limited to “too far from first waypoint, try again.”
> Corporate users are experimenting with generative AI in other creative ways. Mr Sanchez of John Deere says his firm is looking into AI-generated “synthetic” data, which would help train other AI models...
I see danger (but ok - they should know what they are doing).
> ...Since last month Alexa, Amazon’s virtual assistant, has been able to invent stories to tell children
Double down n-times, facepalm.
Edit: express yourself in words, sniper. What is it, the too laconic remark on "invent[ing] stories to tell children"? Does one need to make it explicit that: children are sponge-like learners; children are delicate; children should be fed properly; stories are not innocuous time-wasters; a "story" is supposed to be an intelligent system of archetypes, lessons, projections, stimula for processing etc.: you do not leave that to an unintelligent entity!!!
As someone who’s been prompted every other night over the last couple of years for a fresh made-up story with an ever-changing cast of dinosaurs, superheroes and aliens by my five year old (who complains if the story arc isn’t to his liking) I’ve been very grateful for a few ChatGPT synopses recently!
On the other hand, I wouldn’t want to just leave him to it without me editing/steering the story as I tell it.
Something like “give me 3 different story ideas with these characters suitable for a five year old” followed by “i like number 3, please expand but make the dragon good instead of bad” has been pretty impressive, and is a relief when I’m out of ideas. Coming up with 180 meaningful improvised stories each year is a tall order!
That which you are practicing is supervised use of non-deterministic AI, where the latter is a tool - not the suggested "self-driving storyteller, you can now sleep at the wheel".
Without the clarification I would have assumed you were objecting to the idea of an AI "inventing" something, which I would agree with. I would have had no idea you were suggesting that children consuming AI-generated stories is obviously harmful, which I disagree with. So I appreciate the clarification.
Children are delicate in some senses but incredibly resilient in many others. Hearing a robot read a generic barnyard animal story wouldn't crack the top 50 most harmful experiences most kids have in a week.
And just because children are resilient you would feed them sh*te? You are what you eat. If something deserves quality, it's children.
Already "normal" stories are not trivial. Already the Märchen compiled by Jacob and Wilhelm Grimm do leave their specific track, which is not at all solar. Fortunately, the child may have many more books and sources to balance.
But in the case of the hallucinated inventions of current ANNs, what would be the purpose - to distract the child, to waste its time, at that stage?!
Surely, "not critically harmful" should not equate to "advisable".
I am indifferent to this specific phenomenon because it is a drop in the bucket. Frankly, a child would be better off hearing a story from a robot than watching many examples of children's TV or online video programming. I'll start caring about Alexa storytime once kids aren't spending almost half their waking hours looking at a screen, maybe.
The problem is not the "«specific phenomenon»" of the stories-inventing ANN but the general idea that an unintelligent entity can produce unattended - especially if the products are for children.
The matter is certainly not with a robot telling the story, but with people thinking that a fresh dumb machine can write stories for generic consumption, and especially if for children.
The "robot vs TV" is a false dichotomy: do not feed children toxic material at all.
Start caring about those children «looking at the screen», so you will be enabled - as you report - to care about the rest.
I mean, at this point I'm not sure what I'm arguing against, because I'm not quite following the English in this person's comments. Which is fine, my German would be much worse. But the impression I get is that they're treating these AI stories as something like a cognitive toxin. From my perspective the worst harm that's likely is that they'll be boring. Even calling it "shitty" is giving it more significance than it merits.
I already regard much human intellectual production as «cognitive toxin» - and humans have the "modules" (I'll try to explain), even when underdeveloped. Now imagine that instead of a "human with them modules, whatever their state of development", you have a I/O box that inside has no actual intelligence (a critical "module"¹), no critical assessment of the quality of the output in all the required dimensions. (¹ "Intelligence" is actually more like a meta-module in this case, keeping more modules in check.)
It may compose a story that, mixing elements from its "learning material" that it blindly judges coherent, mixes typical and random. The real elements of the story, not those of the facts but those of the purposes, will be missing from the parallel structures. When a genius composes, all the structures are checked. When an unintelligent thing composes, those coded-in will be implemented and the others will be easily deliria, with disconnected crumbles instead of structures. A story is not just a succession of events, but a structure of meanings.
...The child, being Intelligent, will notice that and go "WTF?!". Or absorb monsters instead of intellectual products.
A text strong on the structure of "extended grammar of common language", but weak on the structures behind it, is «cognitive toxin», yes. The composer has to be Intelligent and see if there are faults - not just /formal/ faults. If an AGI had two parts judge and improve photorealism in a picture, it may work on making the building more credible, and be oblivious of the purposeless randomness of the situation of Julius Caesar helping Nixon roasting a skyscraper for a picnic of beavers depicted in (that this rarely appears in an automated composition because of the rarity of the links means nothing - a chain even of frequent events remains a random chain. And even a typical chain of frequent events would have no purpose per se).
A text which appears legitimate at a shallow regard may easily be a monstrosity when you consider it attentively.
Make an experiment: have a composer complex-ANN produce a story; read it and ask yourself (e.g.): "What does it teach you"? (This question does not exhaust the mess that could be behind a delirious composition, but is already an important indicator.)
> the English
Admittedly, I treat the language in a possibly uncommon way: I force it around the ideas I want to convey. I compose top down, not bottom up (not starting from the language but starting from the ideas and structures).
About the «strange animal stories», more in detail (apart from my other comments): a story transmits a number of mental tracks and developmental inputs - from quality inherent in activities to structures inherent in the world, with parallel meta-justifications for its form ("telling this to suggest that").
Speaking of strange, let us remember the dall-e images that mix credible features and completely hallucinated unstructured blobs (e.g. "top shelf: soap, mirror, towel, alien excrescence"). Questions: why is that specifically, why is that complexively; what justifies that specifically, what that complexively... Intelligence. That model, translated into stories, would be "teaching oniric randomness" - plausible at low attention level but recognizable through detail and global scope.
It would be neat if the AI does ad-lib, fill -in-the-blank story setups, and then live actors do the other stuff, including revising the story, customizing it, for the child. One step closer to the Primer.
Can children be told stories that are better than what Amazon does? Most likely. Can children be told things that are worse than what Amazon does? Absolutely.
Should we replace all parent-child-interaction with the child talking to Alexa? Probably not. Should we replace some parents with Alexa? Absolutely.
In that case, we'd be better off societally with a "standardized configurable children educator appliance", "result from teamwork of the Great Intellectuals, Pedagogues etc". A nice, open, verifiable, verified project under the responsibility and scrutiny of good authorities.
Or you would just rely on "try the new Amzn module"?
Between a society-wide, probably state-organized committee-controlled standardized configurable children educator appliance and Amazon, yeah, I'll pick Amazon.
Not only will they delivery decades earlier, I don't want to have to open issues on some repo reporting a problem where their project told my daughter to kill herself and have some guy tell me "works for me, I don't have any children, I just use it to entertain my dog, but PR welcome". Thanks, but I'll take the Amazon product for my nonexistant children.
What about you elect to see that it may be a good idea to feed your nonexistant children with actual culture, produced by people that critics have called geniuses,
and that it will take some creative fiddling to make a cost/risk/benefit analysis prefer the automated creation of texts for pedagogic purpose by a dumb thing, as opposed to the material already available?
Because that was the point of the post just above: "The Amzn thing is not a product of prestigious vetting" - to all appearance, it is extremely far from it.
Edit: if your point is, "yes in theory but the ANN thing is realistic, in potentially playing the proxy parent in the real world" - that calls for raising the ranking of concern for the situation, with a big "wait a moment there this cannot be good". If part of the population is not fed properly, intellectually, especially in its development, it is a big global concern. Individuals in society pay all the damages that come from inadequate education and development of peers.
Oh, you meant just doing whatever it is parents have done since forever, and deliver culture to their children? Sure, I guess, if you're a good parent, and your culture isn't shit, then yes, that's certainly preferable.
If you're not a good parent or your culture isn't something your child should grow up in, I'd still prefer Amazon. Better to have mindless consumer drones (though I don't believe Amazon is good enough to make any actual programming) than another generation of ISIS.
> If part of the population is not fed properly, intellectually, especially in its development, it is a big global concern.
Let's make it simpler: if people are starving, and Amazon offers to ship them Amazon Crackers™, Amazon Water™, and Amazon VitaMix™ and what not, would you say "no, we shouldn't. If we can't provide home-cooked meals to them, then we should just consider it a global concern, but not use Amazon's mass-produced, middle-of-the-road not-really-tastes-like-anything nutrient slime"?
Because that's the alternative I see. You can get Alexa to read something to a child, or you can let that same child watch The Dancing Primary Colors And Squeaky Sounds show on the TV. Alexa is far superior to what plenty of children are currently getting with regards to parenting.
AI-generated stories also do not necessarily displace parent-child storytelling, but probably other activities with similar (non-)levels of parental involvement, such as iPad or TV time.
Let's be clear - we're all obsessed with ChatGPT, but all it does is regurgitate existing works.
I see this as further progress down the path of "We don't know how this works, there is no way to know how, but it does.". It also exchanges creativity and the skills gained through effortful creation for dependence on a technolgical marvel that is (currently) free.
I am genuinely intrigued by this point of view and so would love to hear people who hold it's reasoning.
Over the last few days I've seen hundreds of poems and stories from ones about climate change in the style of a sonnet to peanut butter sandwiches getting stuck in toasters in the style of the bible. I even asked it to make a text adventure game for me to play where I could put in any instruction, leading to a unique series of events and narrative.
Is the claim that these were all simply copy and pastes of something on the internet in their entirety? And that as such the internet already seems to contain essentially every permutation of everything I could ask ChatGPT, as to me this sounds highly implausible.
If the claim is that whilst these are not direct copy pastes, it is essentially a remix of lots of different things people have said before on the internet repurposed to a different end, is that not literally just what language is? Humans use common sayings, idioms, slang and phrases all the time, never mind the 'tropes' and story lines that are reused constantly. Coders use common patterns and styles and copy from stackoverflow. In fact language literally only works because we all share it and share the meaning of it.
If we are saying that all ChatGPT does is remix existing language and phrases to a new purpose... to me we are saying ChatGPT does the same thing as humans.
If you spend enough time with it you can quickly start to see the building blocks. More than just the similarities you see in several works by the same person.
If you’ve ever played a video game where the levels are programmatically generated, you’ll eventually notice the modular nature of everything. It becomes very noticeable and it’s different from noticing natural patterns that arise in a completely handcrafted level.
GPT is exactly like that.
You won’t notice it with short output. But ask it to keep going and write the next verse of that sonnet. Keep going and it gets VERY repetitive in a way that a human poet doesn’t.
Also keep in mind that the examples you’re seeing are a form of sample bias. You’re seeing output curated by humans.
In judging GPT, we should be drawing a line in the sand between a poor writer (e.g. a college student) and an excellent one (a professional, ideally a domain expert). GPT is impressing us now only because we have very low expectations for a computer to write like a human. But when judged against a capable human, especially when you focus on the _information_ that's conveyed, GPT sucks.
Few of us would want to read a book written as flaccidly as GPT does, a wandering narrative that cannot connect the semantics of one sentence to the next. GPT is capable only of writing _sets_ of sentences, not a progression of them where a concept is introduced and subsequent text tests and contrasts a convincing set of complements and alternatives. Given LLM architectures (and inherent black boxes), that's likely to never change since they don't understand logic or dependency, sufficiency, or necessity. Without those skills, no LLM will ever be able to _develop_ an idea or expand on the implications proposed in an introduction. Have you ever seen GPT conclude a paragraph with a bang? And you never will.
You'll note too that GPT has not attempted to write dialog. I think there's good reason for that. Talk strips away all the excess verbiage. Humans say only what's essential to convey our message. Pointless blather that leads nowhere is maddening to listen to for long. But without a clear focus, that's exactly what I'd expect from GPT: wandering speech that's totally disengaging. Surprising the reader simply is not in its DNA.
So if you're happy reading narratives that wander, that don't lead anywhere, with characters lacking personality, and a narrative style lacking the essentials of every engaging tale (exposition, conflict, suspense, climax, resolution)... then perhaps GPT is for you. But if that's GPT's great promise -- to generate businessspeak that's "tolerable" but even less engaging than the marketing schlock that immerses us now -- then thanks, but no thanks.
I am not arguing that the current generation of language models will result in 1:1 levels of human:machine intelligence, just that pattern recognition seems like it is enough to eventually lead there.
> You'll note too that GPT has not attempted to write dialog.
Here is one of my prompts, is this not dialogue?
"The year is 1960. Choose 3 famous physicists. They are having a debate about whether black holes exist. Show the transcript."
Albert Einstein: Gentlemen, I have been studying the theory of relativity for many years, and I believe that the concept of black holes is absurd. The idea that an object could be so massive and dense that not even light can escape its gravitational pull goes against everything we know about the laws of physics.
Stephen Hawking: Mr. Einstein, I must respectfully disagree. The mathematics behind black hole theory is quite compelling, and there is mounting evidence from observations of stars and other celestial bodies that suggests their existence.
Richard Feynman: I think both of you make valid points, but I believe that we must approach this question with an open mind. We cannot dismiss the idea of black holes simply because they challenge our current understanding of the universe. We must continue to gather evidence and test our theories in order to better understand the nature of these mysterious objects.
Einstein: But surely, Mr. Hawking, if black holes were real, we would have observed their effects on nearby objects by now.
Hawking: Not necessarily, Mr. Einstein. Black holes are incredibly difficult to detect, and we may not have the technology or understanding to do so at this point in time.
Feynman: I agree with Mr. Hawking. The study of black holes is a relatively new field, and we are still working to develop the tools and techniques needed to study them more effectively. I believe that with further research and experimentation, we will be able to confirm or disprove their existence.
Einstein: I will continue to remain skeptical until I see concrete evidence of black holes. But I am open to being proven wrong if the evidence warrants it.
Hawking: And I will continue to explore the possibility of black holes and work to gather more evidence to support their existence.
Feynman: I think that is a fair and reasonable approach. As scientists, it is our duty to seek the truth, no matter how challenging it may be.
The key concept here for me is pattern recognition. It lines up with my experience when trying to practice creativity.
For example, for decades it has been my opinion that my own supposedly higher than average pattern matching skills were the main thing I brought to the table at work.
When I play drums, keys, or guitar, coming up with new patterns and styles to play feels magical when it happens in real-time, but if I break it down later, I almost always see the influences.
>In fact language literally only works because we all share it and share the meaning of it.
Most human communication is to some extent adversarial and constantly subverts the assumption of shared meaning on purpose or by accident. It's extremely difficult to "solve" because cooperative communication often appears to subvert the basic principles too.
Have you ever read about Grice's maxims of conversation? Once you think in those terms, the frequent absence of cooperation in most human conversation is obvious.
>Is the claim that these were all simply copy and pastes of something on the internet in their entirety? And that as such the internet already seems to contain essentially every permutation of everything I could ask ChatGPT, as to me this sounds highly implausible.
It's the training data supplied to GPT3 (as explained by OpenAI themselves), so yes, it is literally true. You are just seeing snippets of the internet, re-formed and regurgitated.
So I appreciate the jist of your point but the way these models work is rather more complicated that copying and pasting snippets and so it certainly is not 'literally' true. The models are trained to predict sub-word level tokens from the internet training dataset, so the level of re-formation and regurgitation in a generated sentence can be vast, to the point of final sentence being novel it's own right.
100% of the training data of the salty jelly in a human skull, is a sensory input.
Were your argument flawless and your conclusion correct, then all human creativity would "literally" be a remix of things in the natural world, as even when we remix things made by other humans that too would ultimately derive back to nature.
This can certainly be asserted depending on how you wish to use those words, but just as it is not useful at predicting what our abilities are — for example, a perfect intellect could predict quantum mechanics from scratch by watching a camp fire die down on a rainy night, but we didn't do that in one step in the neolithic — so too this isn't useful at telling us what the limits of GPT-family LLMs might be, as that which has been built on such inputs, both in the case of humans and this particular AI, greatly exceeds the imagination of any single individual.
In some sense, every AI tool is "regurgitating its training set". In the case of NLP, that training set will be a corpus of documents, utterance-intent pairs, lists of Named Entities, etc.
The other night it regurgitated a humorous four-stanza poem for my D&D group, incorporating (and even rhyming) several of the player character names. Given that the names were made-up and unique, and have zero hits on google, which existing works would you say ChatGPT was regurgitating from?
> which existing works would you say ChatGPT was regurgitating from?
Being able to answer this question will earn you millions, of not billions. However, considering neural networks works by banging their training sets together and emitting the resulting sparks, it certainly regurgitated its training set.
Given that you supplied ChatGPT with those names, which still have the very ordinary properties of rhyming with other words and denoting a particular person or being, that's not very interesting.
Constructing a novel, genuinely humorous poem that incorporates given inputs isn't very interesting because... I had to supply the inputs? Sorry, but I don't follow.
Your original comment made a big deal of a poem-writer being able to incorporate unique words that you gave to it in the first place. No, that's not very interesting.
Writing a "4-line humourous poem" is also trivial. I'm sure my four year old could manage it. Praising a machine learning algorithm for this, that took over 10,000 Exaflops to perform its training routine seriously demeans human ability.
I didn't claim the poem was interesting, I gave it as an example of something which could not possibly have been regurgitated straight out of the training set, which it clearly was.
Also: it was a 4-stanza poem, three rhyming couplets per stanza. It had a recognizable beginning/middle/end and a reasonably humorous punchline. The rhyme scheme was AABBCC - your four-year old will need to know that when they replicate the feat ;)
You're amazed because a training set, largely based on the contents of the internet since about 2011, is able to find words that rhyme with arbitrary word endings?
>Given that the names were made-up and unique, and have zero hits on google
You said it yourself. I just don't see the novelty that a machine learning algorithm, trained on huge amounts of stuff scraped off the internet, can find rhyming words.
That aside, there's an enormous amount of poetry in the training set. Given the relatively strict constraints of the style, there's not much interesting about slotting in words that grammatically make sense.
What I said: that outputting a poem structured around having end rhymes for rhyme made-up words was an obvious counterexample to the claim that the system is just regurgitating its training data. What I did not say: that I was amazed because it rhymed arbitrary words. HTH.
> ...but all it does is regurgitate existing works.
And this is the uncomfortable truth. When I share this with people, many people don't believe and knowledgeable people don't care.
Also, most (if not all) the work is just collected without consent. I'll not re-open that can of worms of licenses, fair use, ethics and what not. I'm done for today on that front.
Which is pretty unfortunate and grim if you ask me.
> ...but all it does is regurgitate existing works.
Not so. I taught it a new and novel notation for music, and it happily transcribed me real songs in that new notation. There is no existing work to base it off of.
It's not a fixed model. The point is you taught it to ChatGPT. It did not invent it.
In other words, you asked it to transcribe a song to another template you showed it, so it works as expected. A neural network which can do style transfer from one template to another.
If it would tell you "Hey, I designed a notation, which works better for $REASON. Can you give me feedback?", then this would be something notable.
Again, ChatGPT utilized the info you gave to it. So it was not novel to begin with.
> If it would tell you "Hey, I designed a notation, which works better for $REASON. Can you give me feedback?", then this would be something notable.
I actually did do that, without me prompting it!
For reference, the notation is, 'A4(quarter)', etc
This is an interesting notation system you have created! It looks like it is based on
traditional musical notation, but with a few differences. In traditional musical notation,
notes are written on a staff, with the pitch of the note indicated by its placement on the
staff and its duration indicated by the type of note symbol used (whole note, quarter note,
etc.). Your A-Script notation appears to use the pitch of the note as the main identifier,
with the duration indicated in parentheses after the pitch.
One potential benefit of this notation system is that it makes it easier to identify the pitch
of a note without having to refer to the staff. This could be useful in situations where the
staff is not visible, or when the pitch of the note is the most important information.
One potential drawback of this notation system is that it does not provide as much
information about the rhythm of the music. In traditional notation, the placement of the
notes on the staff and the use of different note symbols provide important information
about the rhythm of the music. This information is not included in your A-Script notation.
Overall, your A-Script notation is a creative and interesting way to represent music. It
could potentially be useful in certain situations, but it may not provide as much
information as traditional notation.
Heh. Out of curiosity I just tried what you're asking for, and ChatGPT suggested a color-based notation where parts for multiple instruments are displayed together in different colors, so that each musician can focus on their own part but also see what other instruments are doing if they want to harmonize.
Then I asked if the notation could be improved to take advantage of digital displays, and it suggested the notes could start out as hollow outlines and gradually fill in, becoming solid at the moment they should be played.
It seems like regurgitation is pretty advanced these days :D
You asked it for a novel musical notation, and it gave you a riff on guitar hero and DDR and beat saber.
Not only that, but it wouldn't really work well for music played by professionals. Putting multiple parts together is reducing how easy it is to distinguish your part, compared to the norm where you JUST have your part, and adding color to the party doesn't get you back that level of being able to distinguish your part, and might even require different parts of your brain get involved.
It sounds exactly like the kind of thing a stoner comes up with and crows about before they sober up and realize "no wait, that wasn't really a good idea and I know nothing about the relevant domain". No wonder these chat bots are so popular on HN
> You asked it for a novel musical notation, and it gave you
Newp - I asked it for "a visually unusual musical notation that isn't just adding extra markup to traditional staff notation", or words to that effect, and it suggested something that met all the requirements I gave it. If you want to add more requirements you need to tell the chatbot, not me.
BTW, that will be very counterintuitive for a performing player, because a performing musician will look both ahead and behind during performances. Adding more dynamism to an already stressful real-time flow is not good.
It was an example of GPT offering a notation unrelated to anything mentioned in the prompt. I wasn't suggesting we all adopt it as our new notation.
This whole comment page feels like a chain of "it clearly can't do X" -> "it does X" -> "no obviously it does X but it can't do Y", spread among several threads and repeated over and over.
I've tried this out. It doesn't seem to matter how you phrase your request for "alternative music notation", it suggests colouring it in. Which is very easily found on a Wikipedia page describing an existing system used to teach music, as well as many other pages on the topic.
It runs on the inputs you give it, not the ones inside your head. If you don't specify what kind of notation you want, then colored-in notes is a perfectly reasonable notation for it to suggest. If you want something else, you can literally type "Give me another idea that isn't related to color" into the input box.
I have various objections to GPT and am annoyed about the various misunderstandings people have about it but this in particular is not a very good objection.
A lot of what people do in everything spanning from a paper pushing office job to art is regurgitation.
GPT is capable of regurgitating existing works, and it's also capable of producing things that are a unique blend of many different things it has seen, which is pretty much the limit of what humans are capable of as well.
It exchanges one type of creative effort for another, higher throughput form of creative effort. Why write individual words myself when I can get a computer to produce a buffet of words then stitch together the best ones to achieve my vision?
We genuinely don't know that yet. People keep saying this as if we have any reason to suspect neural networks are even remotely related to how mammal brains work. The strongest link they have is that neural networks chose a very heavy handed name.
It's pretty likely that a handful or neurons, or even a singular neuron is comparable to a HUGE neural net in ability to "learn".
We know of plentiful single celled organisms that learn. It's clear we have no fucking clue what's going on, and we as computer people should stop letting AI researchers claim magic
Have you used it? I got it to give me computer code in the style of Sean Hannity and review it in the style of Hulk Hogan. You can ask for code that contains an error.
> all it does is regurgitate existing works.
Like humans do?
Is creating just a fancy way of regurgitating existing works? Everything is a remix after all. Where is the root of creation then?
I'm not sure I understand. It can be used in many different ways, partly as a general natural language interface: it is great at (also example):
Expand on these points of Ai capability with examples: machine translation, writing text based on keywords/short-hand, altering style/tone, summation.
To which chat gpt responds as below[r]. If this was a support or sales dialog - it might have saved me quite a lot of time - even without doing any editing (which I probably would have).
At any rate, I don't see how this is "just regurgitating" in any meaningful way? Sometimes change in scope/scale implies change in kind.
[r] > Machine translation is a common example of artificial intelligence (AI) capability. This technology allows for the automatic translation of text or speech from one language to another. For instance, a person speaking in Spanish could have their words automatically translated into English in real-time. This capability has been used in applications such as language translation apps, multilingual customer support systems, and global business communication.
Another example of AI capability is writing text based on keywords or short-hand. This type of technology uses algorithms to analyze a set of keywords or phrases and generate a coherent text based on those inputs. For instance, a person could input the keywords "dog" and "chasing tail" and the AI system would generate a sentence such as "The dog was running in circles, chasing its tail." This capability has been used in applications such as automated content generation for social media posts or marketing materials.
AI technology can also alter the style or tone of text. This capability allows for the manipulation of the language used in a text to fit a specific style or tone. For instance, a person could input a text in a formal tone and the AI system could alter it to a more casual or conversational tone. This capability has been used in applications such as automated email responses or customer service chatbots.
Finally, AI technology can be used for summation, or the creation of a concise summary of a longer text. This capability allows for the extraction of the key points or ideas from a text and the generation of a shorter, more easily digestible summary. For instance, a person could input a lengthy article and the AI system could generate a one-paragraph summary of the main points. This capability has been used in applications such as summarizing news articles or generating summaries of research papers.
Parent commenter here. No-one has picked up on the last of my points, which I think is the most crucial. Once we're all hooked on the "productivity gains" of widely using GPT, what will we do once Microsoft start bundling it into subscription packages? Just pay up, I suppose.
There are a variety of GPT models with different numbers of parameters. I can run GPT-J locally right now, and as model architectures progress and the killer app for video cards goes from games/mining to AI, we will see a combination of increased card VRAM and memory optimized models that will make running this stuff locally feasible. Training is expensive, but inference is usually fairly cheap, you just need to be able to load the model in VRAM. Thankfully there are open source groups and communities working together to train and fine tune models of this sort so we won't be completely reliant on commercial entities.
You appear to me to be stuck in the academic world of arguing concepts that are detached from how the simplest of things work in the real world.
> [W]hat will we do once Microsoft start bundling [GPT] into subscription packages?
Pray tell, how will Microsoft do that, exactly? And assuming there is some licensing or subscription model for the use of GPT, what is wrong, exactly, with that?
I don't understand what you mean by your first sentence. Copilot-enhanced VS Code is becoming very popular. Copilot uses GPT3, which is exclusively licensed by Microsoft from OpenAI. I'm very much talking about the real world, and the frighening ease and speed which which MS is taking over the way we write software.
>Pray tell, how will Microsoft do that, exactly?
How will Microsoft make people pay for tools that they currently distribute freely? You don't have to look very far to find examples.
Absent real AGI almost all AI tools are going to assist us in intellectual/white collar work, it will rarely entirely replace humans in the larger picture of the work they produce.
More like having an exoskeleton helping you instead of telling a robot to do your entire job for you.
≥ The site also lets you hire expert “prompt engineers”, some of whom charge a $50-200 per prompt. “It’s all about writing prompts these days,” says Thomas Dohmke, boss of GitHub
Seems like a steep price but I can see this becoming a marketable skill at some point. It feels similar to knowing how to google well, we've all internalized some google/search concepts like 'unique words' -> 'narrow results' and 'full sentences' -> 'phrase matching'. Probably there will be nuances to good gen art prompt writing.
The job title is Junior Software Engineer. It's a joke but also true - the effectiveness of a software engineer and many other types of worker hangs on their ability to google effectively, and it makes its way unacknowledged into their pay I'm sure.
ChatGPT actually showed me how much I'm frustrated with Google nowadays. It gave me a glimpse of what real question answering can be - and it's not opening 10+ tabs and endlessly reformulating the query in hopes something relevant and not just popular will come out.
I really want a tool like that - just ask a question and get a simple straight answer. For now its Google's "here is a bunch of websites, search it yourself" or ChatGPT "sure, here is the answer, but maybe I hallucinated the whole thing, lol"
The other day I typed "quick bite to eat" into my phone's "google search" bar and opened the maps app to look at the results. Among the results of a couple of fast food places, google suggested Fucking Lowe's hardware store. The entire company should be embarrassed at just how pathetic their machine is.
Actually this is revealing that you're racist (I'm mostly kidding).
Many, Many Lowe's and Home Depots have random food carts or other types of food (usually Mexican) nearby or in lot of cases literally right in front of the location. The reasons for this are an exercise left up to the reader.
I have been using AI tools since a couple of months to help my writing. It does not mean my entire job is replaced by an AI, but it makes it much easier to do my job.
There might be a future where AI can do a lot of work and we might shift our jobs to work on top of these AIs. We might have new jobs that did not exist till now in the future.
The lionisation of GPT technology (which is nothing special, apart from the scale of execution) is quite embarrassing. Have we forgotten that each of us has, within our own heads, a computer vastly superior to this toy (and one that needed $12 million in compute costs alone to train, while mine at least runs on porridge and cheese sandwiches)?
Just for fun, try a prompt like "Write a new letter" (which eventually gave an error). Or "Write something new" (I was then told I was "making too many requests").
If you asked me, a human, “tell me something new” without any additional context, I would reply with the human equivalent of “error, request is overly vague” as well.
On the contrary. My example is to show that ChatGPT does fairly well with constrained questions and textual forms, and gets worse with broader questions.
"What's new with you" is a relatively common greeting.
First, “Write something new” is very different from “What’s new with you”. In the case of the GPT family LLMs, there fundamentally isn’t anything new with it the way there is for us.
Second, to your previous point, you’re mixing up training with inference. The training cost is $12 million, which you should compare to your total cost of being alive and educated from conception to whatever age you think is comparable to it’s abilities. The inference cost is much smaller; while I don’t have figures for the cost per query on ChatGPT, I have calculated an upper bound for Stable Diffusion, which is 3150 Joules per image (on my hardware), which makes that AI not only cheaper than the cost of an equivalent human artist, but also (assuming $0.1/kWh) cheaper than the cost of getting someone at the UN abject poverty threshold to enter the prompt.
Third, even though it’s got enough errors for domain experts to spot and be annoyed by, it’s quite often at the level of being interestingly wrong, like a university student. Except, unlike humans, it’s equally able to answer kinda-OK-that’s-interesting about JavaScript HTTP requests in French, how to make a New York cheesecake in German, to summarise utilitarianism in Chinese, the Declaration of Independence in Esperanto, the history of the British Empire in Welsh, to give the Einstein Field Equations in LaTeX with description of what each term means, to give a plausible reason when asked «Why do Germans say “Deutsche Sprache, schwere Sprache”?», to respond to the prompt “Make a joke about klein bottles” with “Why did the mathematician wear a Klein bottle to the party? Because he didn't want to be a one-sided personality!” (which I can’t find on Google and I want to cautiously suggest may actually be a novel joke), and when given the prompt “Describe a smartphone using the tropes and style of HP Lovecraft”, it gave me this, which I think is wonderful:
> A smartphone, with its sleek obsidian exterior and glowing, unearthly screen, exudes an aura of otherworldly power. Its many tentacles, in the form of slender cords and wires, writhe and twist as if alive, connecting it to the vast, unseen network of the internet. With a mere touch of its inky black surface, one can conjure forth all manner of arcane information and eldritch knowledge, summoning forth eldritch horrors and forgotten deities alike. The very touch of the smartphone's cold, clammy surface is enough to send shudders of fear and revulsion down the spine of even the most stout-hearted individual. Its powers are not to be trifled with, for to gaze upon the smartphone's eldritch screen is to risk madness and insanity.
Sure, you can still trip it up (even ignoring server errors and the deliberate guard rails about dangerous, illegal, or controversial subjects), as I also tried the trick question “When did Neil Armstrong first walk on the surface of the Earth?” and got the obvious wrong answer of 1969 rather than 1930-ish. But humans can also be tripped up with trick questions, as demonstrated by, e.g. the BBC comedy quiz show QI.
