Technology

This is a classic feedback problem: you use a microphone to amplify your voice, but If the mic picks up the amplified sound it creates audio feedback + a sharply increasing wail.

I can't imagine what LLM feed back 'sounds like', but a guarantee you it ain't pretty.

I've heard this theory. Feels like unrealistic hopeful wishes of people who want AI to fail.

LLM processing will be a huge tool for pruning and labeling training sets. Humans can sample and validate the work. These better training sets will produce better LLMs.

Who cares is a chunk of text was written by a human or not? Plenty of humans are shit writers who believe illogical or clearly incorrect things. The idea that human origin text is superior is a fantasy. chatGPT is a better writer than 80% of humans todat. In 10 years LLMs will be better than 99.9% of humans. There is no poison to be avoided.

chatGPT has an apparent style when used in the default mode, but you can already get away from that with simple prompt tweaks. This whole thing is a non-issue.

You have to remember one thing, writing or speaking of a language is not a fixed scientific law or math formula that will stay true through out history. A living language is always moving and evolving in most of its components, be a vocabulary, grammar, or even meaning of words/phrases. We are just entering an era where AI generated content someone might feel appealing and follow that style, compare to copy a contemporary popular writer.

I think it could end up being a problem that we face in the future, but probably not an insurmountable one.

For one, I suspect that clean data sources will always be available, though it could become a lot more expensive to obtain. As an extreme example, you could always source your data by recording in-person conversations.

Also, as AI improves, I'm guessing it will be able to handle bad data more gracefully, and that it should be able to train to the same effectiveness while using a smaller dataset.

Well thankfully we have tons of archived conversations before ChatGPT on archive.org.

Many forums with hundreds of thousands of posts are just waiting to be scraped.

I suspect the quality LLM development teams will pursue the same in-depth data sourcing & cleaning techniques that quality ML researchers are developing today. Or rather, they'll do something similar in principle to mitigate this issue.

I still agree with your conclusions. It will be a bigger consideration and less scrupulous teams will be more effected.

There has already been jokes of AI being used to create well crafted correspondence, then another AI translating that into a short summary.

I think you are going to see AI as something people lean on more to talk to others, and that is going to create its own language where AI talks to AI.

thanks, it's terrifying

Thanks, I hate it

There's usually a context difference that might might be significant. People don't write the same way way for an email, like they would a letter, text message, or tweet.

They might write more like an LLM for things like essays and reports, but your usual writing is probably still fine. Then classics that inspire people to write are still around, and I doubt that they would be supplanted by an LLM any time soon.

We might start being in trouble if people start republishing books with them, but that's unlikely to to happen any time soon, considering the current state of copyright around AI works.

I’m not sure this is true. They could be trained based on published works prior to a certain date as the formal writing style, eg Project Gutenberg, then layer on the recent internet to better capture modern stylistic trends.

Ultimately, the models will always require fine tuning, and selecting which data set you use for early training has a very large impact on the overall performance of the model. Additional knowledge and trendiness can be learned after the fact.

There are some in the research community that agree with your take: THE CURSE OF RECURSION: TRAINING ON GENERATED DATA MAKES MODELS FORGET

Basically the long and short of that paper is that LLMs are inherently biased towards likely responses. The more their training set is LLM generated, and thus contains that bias, the less the LLM will be able to produce unlikely responses, over time degrading the model quality throughout successive generations.

However, I tend to think this viewpoint is probably missing something important. Can you train a new LLM on today's internet? Probably not, at least without some heavy cleaning. Can you train a multimodal model on video, audio, the chat logs of people talking to it, and even other better LLMs? Yes, and you will get a much higher quality model and likely won't get the same model collapse implied by the paper.

This is more or less what OpenAI has done. All the conversations with 100M+ users are saved and used to further train the AI. Their latest GPT4 is also trained on video and image recognition, and they have also been exploring ways for LLMs to train new ones, especially to aid in alignment of these models.

Another recent example is Orca, a fine tune of the open source llama model, which is trained by GPT-3.5 and GPT-4 as teachers, and retains ~90% of GPT-3.5's performance though it uses a factor of 10 less parameters.

This sounds like what an ai would write /s

I think that while LLMs are going to get worse, the AI software will get better to the point of strong AI, and it will do a lot of "apple-esque" changes to mass produced speech that will ultimately be for the better.... The cynical possibility is that it will further taint human dialogue even though it could provide a better way.

I don't believe this theory 100%, however it is true to some extent. At some point, ai language will plateau out and simply won't get better. Once it's at it's max and has little to learn, they will be so human-like it won't matter if it's learning from itself. The percentage of influence would be so infinitesimal it practically won't matter. At that point it wouldn't be necessary to learn anymore, anyway.

We aren't doomed to write like ai, different themes or stories require different nuances. It's artistic. But it depends on the medium. Sure, resume's, cover letters, memos, emails and whatever may become robotic (aren't they already?) But creative stories won't, to a great extent.

What I predict is that they'll try to implement data filters to avoid feedback loops, but there will be an enshittification process for AI too.

What might put the nail into the coffin much quicker isn't the feedback loop, but trying to monetize the whole thing. I think it's only a matter of time until OpenAI will try to get money off of it, like putting certain features behind a paywall, especially those that professionals might use.

I believe I read/heard somewhere that future AI training will take place using less data and will potentially pay field experts to better curate signal from noise

Slightly unrelated, but I was just talking with a friend about how we're going to have similar issues with young artists trying to copy ai. As is, many young artists will turn to cartoons instead of real life when starting out. Their work is a bastardization of a bastardization, with serious flaws in anatomy, gravity, light, and depth. They go on to call those mistakes their "style" and point to other artists making those same mistakes to normalize them. Since "style" isn't something they think they need to improve on, they may become good artists overall while having severe, glaring holes in their skillet that any professional can see. You can sometimes even tell when someone started out because "90s anime" or "10s cartoon network" made specific stylistic choices that changed over time.

So I think ai is going to cause similar problems. Newbies will copy what looks pretty to the untrained eye and learn an ai based style. Then when they become more popular they'll be fed into ai as reference material and perpetuate the problem. Even worse is actual professionals may turn to ai instead of real life references or a desk mannequin. Then their skills may degrade because they rely too much on improper tools. (I've already seen this becoming an issue with photoshopped reference photos.)

Anyways, that's my $0.02

My take is that "L"LMs are already old news. I think targeted or limited data-set language models are going to be the next wave.

I think this partly because very few people can do LLMs at the scale of Microsoft and Google so I think smaller firms and people in their garage are going to aim their sights on smaller targeted data sets with a eye towards factual accuracy.

And then maybe link them/daisy chain them together. I hope there is this unix philosophy for models where they do one thing well but you can 'pipe' data from one to another.

AI might in the medium-term change our vernacular but it won't be for the worse, generally, and most people won't feel much of a change in most contexts.

I liken it to the invention of the steamshovel excavator. Now your average worker doesn't need massive muscles to get work done quickly, but it's not like shovels went away for good, it's still used in parts of projects.

You are right in assuming there will be a symbyosis between AI generated text and human generated text, but jumping from there to assume that we will be using solely AI generated text is wrong, in my opinion.

AI generated content is not good enough on its own, despite what OpenAI marketing team wants you to think. No quality content is made by simply prompting chatGPT. Not just in writing, but in any field of knowledge, actually. Using chatGPT without some level of domain and fact checking on the subject you are prompting is a sure way to get screwed, as some lawyer in the USA will tell you.

But going back to writing specifically, what we will see at first is actually an improvement on the overall quality of human generated writing, with AI offering a solution to the mechanical and usually boring side of writing good content, such as eloquence, syntax, clarity, etc.

Then, what we will also begin to notice is the more frequent use of what I like to call shitstorming.

Shitstorming consisting in prompting a LLM model to bring up ideas, drafts and opinions on subjects you want to write about, and have some understanding on. What you will receive in response will be a biased, somewhat lacking content, which will either inspire you to modify and refactor in a way that it makes sense, or make you so angry that you will have to write something better in response to it. Writer's block will become a thing of the past.

There are others aspects and nuances to this symbiosis, but to avoid going longer on an already long post, I would conclude by saying that this evolution will be a loop that will keep improving LLMs, while also improving human writing simply because we will continuously look for ways to make the content better, and more original.

The bad side is that, for those that don't know how to use the tool, the amount of lacking content and standardized communication will indeed flood the internet, but this will only serve to contrast original content to the point where we will immediatly recognize the two apart, much like we do with advertising nowdays.

Future AIs might be trained manually over centuries by groups of experts.

I think there is going to be some sort of local minima of quality when the humans and AI both train the next AI. But then the quality will likely start raising again as we figure out better cost functions. Current cost functions don’t just optimize output to be exactly like the training data. They allow for some variability like word order (as long as grammatically correct) and synonyms, but that’s about it. Maybe we’ll discover a better cost function later?

As for good human generated data for training and building AIs? It’s like wood from trees. We’ve gone through the “just cut down the nearest tree, it won’t matter, they’re everywhere” period. Soon we’ll enter a data farming period, just like with managed de-forestation, and with the value of task-specific data and LLMs now being obvious, we’re probably already there.

Hmmm … maybe that’s why big social are ratcheting up the prices for their APIs??!!

Honestly, it’s a little creepy how tangible is a Matrix like scenario, without the apocalyptic war part that is. Machines feeding of of our data and thinking (which was, IIRC, the original premise, not energy).

Yeah, basically a photocopy of a photocopy of a photocopy…