Black Mirror creator unafraid of AI because it’s “boring”::Charlie Brooker doesn’t think AI is taking his job any time soon because it only produces trash
The thing with AI, is that it mostly only produces trash now.
But look back to 5 years ago, what were people saying about AI? Hell, many thought that the kind of art that AI can make today would be impossible for it to create! …And then it suddenly did. We’ll, it wasn’t actually suddenly, and the people in the space probably saw it coming, but still.
The point is, we keep getting better at creating AIs that do stuff we thought were impossible a few years ago, stuff that we said would show true intelligence if an AI can do them. And yet, every time some new impressive AI gets developed, people say it sucks, is boring, is far from good enough, etc. While it slowly, every time, creeps on closer to us, replacing a few jobs here and there in the fringes. Sure, it’s not true intelligence, and it still doesn’t beat humans, but, it beats most, at demand, and what happens when inevitably better AIs get created?
Maybe we’re in for another decades long AI winter… or maybe we’re not, and plenty more AI revolutions are just around the corner. I think AIs current capabilities are frighteningly good, and not something I expected to happen this soon. And the last decade or so has seen massive progress in this area, who’s to say where the current path stops?
Nah, nah to all of it. LLM is a parlor trick and not a very good one. If we are ever able to make a general artificial intelligence, that’s an entirely different story. But text prediction on steroids doesn’t move the needle.
The best ones can literally write pretty good code, and explain any concept on the Internet to you that you ask them to. If you don’t understand a specific thing about their explanation, they can add onto their explanation, and they can respond in the style you want (explain as if I’m ten, explain as if I’m an undergrad, etc).
I use it literally every day for work in a somewhat niche field. I don’t really agree that it’s a “parlor trick”.
LLMs are awful for facts, because they don’t understand what facts are. You should never rely on them if you require factual correctness.
They are OK for text summation, formatting and just making shit up. For summation a human with experience still produces nicer output, because they understand the content and don’t just look at words. As for making shit up you will get the statistically most likely output, so it’s usually trite and boring. I think the progress is amazing, but there are still so many problems to be solved.
Right now I use them for boiler plate stuff, like writing a text with some parameters and then I polish it. For code I find them quite useless, because with an IDE I can write boiler plate just as fast as when I polish the prompts until the LLM delivers useful stuff. And with the IDE I don’t get references to methods or entire libraries that just don’t exist.
Right now I use them for boiler plate stuff, like writing a text with some parameters and then I polish it
It’s actually great for dnd to produce NPC dialogue or names on the fly. We also tried using it to calculate area of effect spells, ie “how many average sized humans in armor with swords could fit in a circle with a diameter of 30ft.” We were rolling with it before someone pointed out that it didn’t calculate the area of a circle correctly, however it got the rest more or less accurate. So we don’t use it for that anymore, and it’s funny how what often appears to be the simplest component of a question is the thing it most often gets wrong.
People are also kind of shit at facts. There are so many facts, and many of them aren’t practical for every person who needs to assess a fact’s accuracy to do so. But it isn’t structurally impossible to mimic how humans learn how to gauge truthfulness, we just have to be prepared for the idea that it will be bound by the limitations of language, as well as the risk inherent in trusting data that it has not independently verified.
I use LLMs for having things explained to me, too… but if you want to know how much salt to pour in that soup, try asking it about something niche and complicated you already know the answer to.
They can be useful in figuring out the correct terminology so that you can find the answer on your own, or for pointing some very very obvious mistakes in your understandings (but it will still miss most of them).
Please don’t use those things as answer machines.
I’m going to use those things as answer machines and you can’t stop me.
Jokes aside, I always validate what chatbots tell me, not even just important things. I use GPT-4 for work and 90% of the time it can show me how to use very specific functions in complex ways, but yesterday (for the first time in awhile) it made up a function that didn’t exist. To its credit, I said, “Are you sure about [function]?” and it said, “I’m sorry, I got confused. That function doesn’t exist. However, look into X, Y, Z for further resources” and I did and they were the correct things to look into.
No they can’t. Your phrasing is misleading. It’s a Chinese Room test output and nothing more. I had an Encarta CD that could do rudimentary version of this in 1995. That was more impressive, tbh.
If you’re really comparing LLM’s to your Encarta cd from 1995 and saying the Encarta CD was the superior experience…
I’m afraid there’s not much left for us to discuss… Our views are too far apart.
Sam Altman (Creator of the freakish retina scanning based Worldcoin) would agree, it seems. The current path for LLMs and GPT seems to be in something of a bind, because to seriously improve upon what it currently does it needs to do something different, not more of the same. And figuring out something different could be very hard. https://www.wired.com/story/openai-ceo-sam-altman-the-age-of-giant-ai-models-is-already-over/
At least that’s what I understand of it.
He’s not saying “AI is done, there’s nothing else to do, we’ve hit the limit”, he’s saying “bigger models don’t necessarily yield better results like we had initially anticipated”
Sam recently went before congress and advocated for limiting model sizes as a means of regulation, because, at the time, he believed bigger would generally always mean better outputs. What we’re seeing now is that if a model is too large it will have trouble producing truthful output, which is super important to us humans.
And honestly, I don’t think anyone should be shocked by this. Our own human brains have different sections that control different aspects of our lives. Why would an AI brain be different?
In humans, abstract thinking developed hand in hand with language. So despite their limitations, I think that at least early AGI will include an LLM in some way.
I’ve been having a lot of vague thoughts about the unconscious bits of our brains and body, in regards to LLMs. The parts of our brains/neurons that started evolving back in simple animals as basically super primitive ways to process visual/audio/whatever input.
Our brains do a LOT of signal processing and filtering that never reaches conscious thought, that we can’t even reach with our conscious thought if we tried, but which is necessary for our squishy body-things to take in input from our environment and turn it into something useful instead of drowning in a screeching eye-searing tangled mess of chaotic sensory input all the time.
LLMs strike me as that sort of low-level input processing, the pattern-recognition and filtering. I think true generalized AI would have to be built on pieces like this–probably a lot of them. Ways to pluck patterns out of complex but repeated input. Like, this stuff definitely isn’t self-aware, but could eventually end up as some sort of processing library for something else far down the line.
Now might be a good time to pick up Peter Watts’ sci-fi book Blindsight. He doesn’t exactly write about AI in it, but he does write about a creature that responds to input but isn’t exactly conscious like you or I.
some sort of processing library for something else far down the line
This is what I meant.
pick up Peter Watts’ sci-fi book Blindsight
I just got the EPUB, thanks. Looking forward to reading it.
FYI Blindsight is free for audible members.
Parlor trick is a perfect description.
People don’t get that these things aren’t anymore intelligent than their smartphones predicting the next word. The main difference is instead of a couple words it has thousands to choose from.
Half of the trick is how it uses the prompt to decided what words to start with.
That is not how it works. Your smartphone has all the dictionary available, same as LLM. It is simply something very different. People super confidently discussing about AI on lemmy are the real hallucinating parrots
There is an inverse relationship between the intelligence of a person and their amazement at what these large language models can produce.
People who aren’t amazed at what LLMs produces have no clue how complicated it is to generate plausible language in the first place. Dunning–Kruger and all that.
The ability to generate plausible language was a lack of compute power. The actual programs running the LLM is not complicate.
The model that is produced is complex.
Its training required compute power that was not previously available but the math/code behind these systems is not complex. They are resource intensive. There is a difference that a layperson often cannot comprehend.
I heard the same for people who downvote on lemmy when notified about being an exemplification of the dunning Kruger effect
Have you ever even bothered to play around with any of the LLMs or are you just parroting what you heard in badly written articles?
The fact that the LLM predicts the next word does in no way shape or form limits its intelligence. That’s after all the same thing you do while writing your post.
These idiotic claims about AI not being intelligent really make me questions if humans are.
Nah, nah to your understanding of LLM’s
No it’s not true intelligence. Yes, it makes humans much faster at their work
It has really sped up my work, especially when coding in unfamiliar languages.
It’s silly to compare it to a parlor trick or text prediction.
Yeah, in order to predict text accurately, you need to actually understand what it says.
I think the breakthroughs in AI have largely happened now as we’re reaching a slowndown and an adoption phase
The research has been stagnating. Video with temporal consistency doesn’t want to come, voice is still perceptibly non-human, openai is assembling 5 models in a trenchcoat to make gpt do images and it passing as progress, …
Companies and people are adopting what is already there for new applications, it’s getting more common to see neural network models in lots of solutions where the tech adds good value and is applicable, but the models aren’t breaking new grounds like in 2021 anymore
The only new fundamental developments i can recall in the core technology is the push for smaller models trainable on way less data and that can be specialized for certain applications. Far away from the shock we all got when AI suddenly learned to draw a picture from a prompt
The research has been stagnating.
It utterly baffles me how people can make that claim. AI image generation has exists for not even three years and back than it could do little more than deformed Avocado chairs and shrimp. This stuff has been evolving insanely fast, much quicker than basically any technology before.
Video with temporal consistency doesn’t want to come
We have barely even started training AIs on video. So far it has all been static images, of course they aren’t learning motions from that and you can’t expect temporal consistency when the AI has no concept of time, frames or anything video related. And anyway, the results so far look quite promising already. Generators for 3D models and stuff is in the works as well.
Far away from the shock we all got when AI suddenly learned to draw a picture from a prompt
What the heck do you expect? Of course going from nothing to ChatGPT/DALLE2 will be a bigger jump than going to GPT4/DALLE3 (especially considering most people skipped GPT1,2,3 and DALLE1), that doesn’t mean both of them aren’t substantially better than previous versions. By GPT5/DALLE4 you might really start to worry about if humans will still be necessary at all. We should be happy that we might still have a few more years left before AI renders us all obsolete.
And of course there is plenty of other research going on in the background for multi-modal models or robots that interact with the real world. Image generations and LLMs are obviously only part of the puzzle, you are not going to get an AGI as long as it is locked in a box and not allowed to interact with the real world. Though at the current pace, I’d also be very careful with letting AI out of its box.
Here is an alternative Piped link(s):
robots that interact with the real world
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source; check me out at GitHub.
We should be happy that we might still have a few more years left before AI renders us all obsolete.
Wow, this is some spectacular hyperbole!
That’s the current pace of AI. It’s evolving insane fast and already extremely capable.
Here is a little game:
- go to artstation.com
- pick a random pretty picture
- recreate it in DALLE3, Bing Image Creator (which gives DALLE3 access for free) or Midjourney
Example: https://www.artstation.com/artwork/LRmYvlResult: https://imgur.com/a/ImbNQDk (about 20 seconds of effort)
It’s ridiculously easy to recreate almost anything on there at a similar or sometimes even better level of quality. Literally seconds to recreate what would take a human hours or even days. What are the chances that humans will still be relevant in this line of work in 5 or 10 years, when we are able to create this level of quality after not even three years of AI image generation?
And the same will be true for every other job or activity that mainly works on digital data. When you can find enough data to train an AI on, it’s gone. Humans are no longer needed. And more general AI model will sooner or later eat up all the rest as well.
I seriously don’t know how one can look at the progress in AI over the last two years and not have a bit of an existential crisis.
It’s ridiculously easy to recreate almost anything on there at a similar or sometimes even better level of quality
And ridiculously difficult to copyright any of it because it was generated.
That doesn’t change that the value of human art just went down to zero. Nobody is going to pay hundreds of dollar for something AI can produce in seconds. Furthermore the whole “AI art can’t be copyrighted” is just wrong to begin with, any tiny bit of human cleanup automatically makes it copyrightable again and since nobody can tell how the image was created in the first place, you’d be operating in a minefield if you just randomly steal art in the hopse that it was AI generated. Keep in mind that Photoshop already has most of this builtin and it’s becoming a normal part of the workflow of editing images.
And it’s all pointless anyway. You have AI, you can recreate anything in seconds. Why even bother stealing anything in the first place? You can just make your own and customize it for the occasion.
The whole idea of copyright might soon be obsolete, as AI can make you something very similar, yet completely original.
The interesting question left is: Will static art survive at all? Will the future still have static movies or will everybody just generate their personalized dynamic entertainment on demand?
I want to note that everything you talk about is happening on the scales of months to single years. That’s incredibly rapid pace, and also too short of a timeframe to determine true research trends.
Usually research is considered rapid if there is meaningful progression within a few years, and more realistically about a decade or so. I mean, take something like real time ray tracing, for comparison.
When I’m talking about the future of AI, I’m thinking like 10-20 years. We simply don’t know enough about what is possible to say what will happen by then.
By its nature, Large Language Models won’t ever be truly innovative, after all they rely on expected patterns. But a lot of the media that we consume is also made to appeal to patterns that we expect: genres, tropes, usual messages. AI could replace a lot of it and frankly, that’s scary to think in a world where we need to work to earn our living.
Truly groundbreaking art may not be what people usually seek, it’s often something they don’t even know they want until they experience it, or they might even fail to appreciate it. But it likely won’t be automated unless AI achieves full consciousness, but if it does we will have a much more complicated situation in our hands than “we can command AI to make art better than we can do ourselves”.
Still, getting paranoid over the uncertain latter won’t help us with the former that is just around the corner.
Good points.
One problem with replacing everything with AI that people don’t think about: middle managers will start to be replaced too. There’s no way to ask a LLM “why did you do that”? Fewer people will need to be managed.
It seems unwise to replace managers with LLMs because LLMs don’t understand the real world implications of their responses, they don’t have awareness of the real world, they simply give you often used language patterns, which can be innacurate or biased based on flawed human data. But it would be a great way for sketchy human executives to offload responsibility for unethical actions and feign objectivity or uninvolvement, so I don’t doubt they will try.
Even if we imagine a perfect AI that does takes into account every objective fact and philosophical argument, that still leaves the question of how will the people who get replaced in all these intellectual, artistic and service jobs will make a living. That’s not an answer that technology will give us, that will a nasty political situation.
No, you misunderstood. The managers are fired because there’s fewer people to manage.
That makes sense too. Overall, a lot of people’s jobs are threatened, but I don’t think “learn AI” is going to cut it this time. Not for all these people.
I miss the old Black Mirror…
I’m legit wondering what crack were they smoking in the latest season.
I lost a lot of interest once I realized that the story lines would always turn in to “computer bad”. End. I missed when we had more interesting scifi short stories turned in to anthology episodes.
I mean it’s literally never “computer bad” / “technology bad” it’s always Humans bad, using this technology this time.
I thought this season was way better than the season before it. I was glad they went with something different like the horror theme. The season prior was a shit show of boring tropes.
I couldnt even finish this season, it was just so boring
I just recently started rewatching some of the older episodes and I realized that “Be RIght Back” was inadvertently an LLM episode. Having a computer absorb the online presence of a loved one to allow you to talk with them after they’ve passed is honestly something that seems within reach for these models.
Fun fact, that’s what Replika was originally designed for, before they realized they could make more money marketing it as a therapist and/or erotic roleplay partner.
Bold of him to assume that companies would not just publish the trash - and that people would not watch it anyway.
It’s not just trash, it just wouldn’t make sense.
AI would kill off a character only to forget that the character was dead. There would be no chronology, it would be unwatchable.
100% correct. The IRC channel I hang out in has a bot utilizing ChatGPT and it does a summary of the most recent conversations when someone joins.
Sometimes, it does a great job! It impresses me how well it’s able to summarize multiple ongoing conversations in a succinct way.
…and often times, it gets shit quite wrong. Not the actual topics, those it’s good at – but it is outright terrible at correctly indicating who actually said what.
Granted this is all to be expected – it’s an LLM, not really AI.
“…and somehow, the Emperor returned!”
Haha, I wonder if AI could produce better Star Wars sequels than Disney. Although my stance has softened on them recently compared to when they first came out.
They wouldn’t have AI produce the whole show like that, it’s like feeding it a context to create dialogue within set parameters.
That would be better, but just seems more trouble than it’s worth for the quality. People ultimately do want to watch good well written shows. Apart from those who don’t!
The Big Bang Theory ran for 12 years…
To be fair to them even though the jokes were poor they did their homework with the physica.
Overall yea, but there’s a thing about how the more niche and specialized content is to a person’s interests, the more they’re willing to sacrifice on quality. So I think what will happen, because this will also reduce production costs so much (in theory), is we’ll get these incredibly specific shows made for smaller and smaller target audiences. I’m hoping this ends up generating some hilarious content that just seems absurd to people who aren’t targeted. Instead of catering to universal human experiences it will be like, “a show centered around a support group for people who love black licorice, and the challenges they face in their relationships with people who hate black licorice.”
Yeah as if there already isn’t complete trash, presumably it will just be cheaper and easier to produce, so expect more ubiquitous and niche trash!
Movie and TV executives don’t care about boring. Reality shows are boring. They just care if they make money.
AI is nowhere near the point where it can actually write a script. It doesn’t even remember what it has written to you 5 minutes ago, how is it going to keep shows and TVs consistent? Even when you tell it what you want it to do and that it’s wrong it still provides wonky information.
AI is nowhere near the point where it can…
ChatGPT is 10 months old, not even a whole year. And it was never fined tuned for story writing in the first place. A little bit premature to proclaim what AI can and can’t do, don’t you think?
ChatGPT isn’t the entirety of AI, AI research has been going on much longer than ChatGPT has been around
Isn’t chatGPT just a beefed up predictive text engine?
Depends on the ai though. With koboldcpp you can make memories for the ai to come back with. Even text personalities (like bitchy and sassy responses) when using tavernai together with kobold.
This. You have to baby it and then if you want it to do something different you have to tell it a hundred times in a hundred different ways before it stops producing the same stuff with the same structure with slight differences. It is a nightmare.
“I was frightened a second ago; now I’m bored because this is so derivative.” - Me, while watching some of the Black Mirror episodes, proudly made by fellow humans.
Did he not watch the latest season? Fuck, one episode was literally devoid of scifi entirely. Latest season only had one good ep in it.
Ever since Netflix took over, it’s been a step down in quality
Eh, I never bought this. The show has always been wildly uneven. The first season was very strong with two great episodes but the first episode is not great. The second season is pretty bad overall with the exception of Be Right Back. White Christmas was a good one off. The first Netflix season had some really strong showings with Nosedive, Shut Up And Dance and my personal favorite of the whole series San Junipero. I even thought season 4 was pretty decent overall. Also, banderanatch was actually pretty cool, especially if you really dug into all the paths.
The last 2 seasons, 5 & 6 were pretty bad overall with only one good episode each and some particularly bad ones. But honestly Charlie is probably just running out of ideas and I can’t really blame him at this point. I suspect he’s just trying different things. Sometimes big swings work and sometimes you just get Mazey Day.
I really, really could not get behind the premise that the PM of the UK would fuck a pig to save a princess.
In the real world, it would have been “we do not make deals with terrorists” and that would be that, ratings be damned. Like holy shit, did he start with a completely unrealistic premise.
Edit: complete fabrication be here.
The “do horrible thing to save someone” story, yeah, maybe.
But the pig thing had been going around for years about the then/former prime minister. Some kind of hazing at eton or wherever those posh folks go. The episode was just a reference to that.
The episode is from 2011, Piggate was in 2015.
Ha. I guess I rewrote my own memory.
I think he’s trying to steal the lightning from the episodes that resonated well with the audiences to make new episodes and it’s just not hitting quite as hard.
The old formula was amazing stories meets Twilight zone meets tales from the crypt we’re 8 out of 10 times the protagonist gets right fucked for just trying to be a good person.
It feels to me like the take away from san junipero was that people like happy endings so they’re trying to apply it to everything
The reason you got Maisy day is because he got access to Miley Cyrus and had to write something that she would be happy with.
The biggest problem is just that they ran out of ideas, they have recycled the cookie-brain-upload thing over and over again. The first few seasons were far more creative and covered a broad spectrum of technological and social issues, without relying on a single gimmick.
It comes down to what you like about the show. Do you like the darker sides of technology that raises questions? Seasons 1 and 2 are great. Do you like sci-fi happy endings with a better production value? The Netflix seasons are better.
Nah. Even season 3 there’s a distinct lack of “okay this is supremely fucked up” that you got in series 1 and 2. There’s some fantastic episodes in the first couple netflix seasons, but they dialed back the dread factor in favor of exploring new sci fi concepts. Not complaining, but the og channel 4 episodes are way darker.
I thought it was a Netflix original 😂
Where did this tv show aired then if not Netflix?
Channel 4 in England, I think?
Well I’m pretty sure Scotland, Wales and Northern Ireland too
Yeah yeah that’s the one, with Honah Lee and Narnia and all that!
Charlie Brooker is one of the only intelligent people left.
Lmao black mirror feels like it was written by an AI so quite a statement
Last season was painfully bad
I’m not to worried about AI. Isn’t the next iteration of GPT closed source? Technology is made best as a research or passion project, but once profits become the focus everything goes down hill. That and when you consider the global supply chain required to manufacture the chips that AI depends on, well things aren’t looking too great in that department.
Tl;DR humans will shit all over the prospect of scary intelligent AI well before we get there.
“Open”AI is entirely proprietary and closed-source.
Meta’s Llama series are kind of open source, but don’t publish the weights and so can’t really be reproduced with full accuracy without a ton of manual effort.
These and many other companies in the hype-space are using the same published research from a few years ago, which is why they have similar qualities.
Chatgpt want open source in the first place?
People are so worried about chatgpt taking jobs, and I’m here, worried about chatgpt bots spamming propaganda in the next elections around the globe. It’s going to be a mess…
There’s more to “AI” than ChatGPT. Deepfakes, propaganda swarms, precise tracking of people online across pseudonyms/handles. The power available to malicious organizations and governments is absolutely terrifying. Any social media that doesn’t also have AI-based countermeasures is vulnerable.
I’m here worried about the fact that it’s going to take some jobs, but leave some. So there will be huge unequality for a while until ai cna actually do all jobs
Both possibilities can be concerning, but that doesn’t mean every discussion of one of necessarily has to include discussion of the other.
Maybe the 5th episode of the 6th season was written by an AI and they were playing some 4D chess game all along with out minds, because otherwise, I wonder how such fucking trash got the greenlight to be produced 🤗
This. I think the only one I really thought was good was the Aaron Paul one where they went into space… I might be someone neo-ludditish but that movie shows some true terrors of those who want to eradicate technologies and the individuals associated with them. Cold ending…
It’s only producing trash now. Already there is a decent jump in quality from GPT-3 to 4, and it’s only gonna get better.
Plus it can do a lot of heavy lifting – tell it to make 20 scripts with different prompts and then a single writer or team can Whittle them down. That’s how a lot of scripts end up in production anyways, but now you ain’t gotta deal with writers and can make rapid, drastic changes
I also find the “just look how bad the hands are heh heh heh” thing so dumb … it’s going to learn how to draw hands pretty quickly
The problem atm it’s that chat gpt has pretty terrible memory. It couldn’t write a coherent show if it wanted to
Most of telly is trash already, if it’s cheap enough for entry then it can saturate the market and there will be no need for the expensive “good” writers
Didn’t he just make an episode of Black Mirror depicting the opposite?
I don’t think so. He says he isn’t afraid of AI replacing creative jobs because it’s incapable of originality and as a result, boring. The episode you’re referencing didn’t depict the opposite. There was a quantum computer that was only capable of producing a show that recreated data from the protagonist’s life in real time. It was always limited by actual events as they played out. That episode seems fairly consistent with his views.
Lololol if ain’t that some bullshit 🤣
