deleted by creator

And your point is wrong because you keep boiling it down to simple black and white.

The Nobel prize is not purely political and is not devoid of merit.

The world is not full of binary systems. It’s made of multi variable systems where multiple influences can be true at the same time.

If you want to make a point about why accurately predicting the structure of hundreds of thousands of proteins doesn’t deserve the Nobel in chemistry then I’m all ears. Please tell us all exactly why you think their prize was political and not meritocratic, and why predicting protein structures automatically is not important?

Because if you can’t answer that very specific question, then you weren’t making a point relevant to the conversation, you were making a snide generalization to hear yourself speak.

Thank you for finally spewing out the point you wanted to make from the jump. It’s irrelevant in the context of the original discussion, but you got to hear yourself talk.

Lmao, it’s binary cause you say it’s binary.

Bro grow up. The world is not black and white. Literally not a single award on the planet is meritocratic if you insist on dealing in absolutes. Every award is awarded by some committee and there is some room left for human judgement, which leaves room for human bias, which makes it not perfectly meritocratic.

If you want to go an unhinged rant that no one wants to listen to then email the nobel association directly, don’t waste federated server time.

Lol if you rigidly define things binarily in a way that doesn’t reflect real world systems, then sure they’re binary.

This is false, it’s not a binary system. The prize is both.

I mean I agree that it’s probably vastly overvalued as a whole, the leap between current LLM capabilities and an actual trusted engineer is pretty big and it seems like a lot of people are valuing them at the level of engineer capabilities.

But the caveats are that simulated neural networks are a technological avenue that theoretically could get there eventually (probably, there’s still a lot of unknowns about cognition, but large AI models are starting to approach the scale of neurons in the human brain and as far as we can tell there’s no quantum magic involved in cognition, just synapses firing which neural networks can simulate).

And the other caveat is like the bear trash can analogy… the whold park ranger story where they said that it’s impossible to make a bear-proof trash can because there’s significant overlap between the smartest bears and the dumbest humans.

Now I don’t think AI is even that close to bear level in terms of general intelligence, but a lot of current jobs don’t require that much intelligence to do them, we just have people doing them because theres some inherent problem or step in the process that’s semantic or fuzzy pattern matching based and computers / traditional software just previously couldn’t do it, so we have humans doing stuff like processing applications where they’re just mindlessly reading, looking for a few keywords and stamping. There are a lot of industries where AI could literally be the key algorithm needed to fully automated the industry, or radically minimize the number of human workers needed.

Crypto was like ‘hey that decentralized database implementation is pretty cool’, in what situations would that be useful? And the answer was basically just ‘laundering money’.

Neural network algorithms on the other hand present possible optimizations for a truly massive number of tasks in society that were otherwise unautomatable.

Skill issue.

Try making a cohesive reasoned argument about the topic at hand instead of just staying ‘nuh uh’

Google doesn’t produce truth either, that doesn’t mean it’s not useful for finding information.

Again, you’re hung up on the idea that it either has to give you a perfect answer or its useless, but the situation is not binary.

I’m not falling for any trick about reasoning. I’m pointing out it doesn’t matter whether or not it’s using reasoning or not using reasoning if more often than not it arrives at the right answer, which it does.

I’d argue, that it sometimes adds complexity to an already fragile system.

You don’t have to argue that, I think thats inarguably true. But more complexity doesn’t inherently mean worse.

Automatic braking and collision avoidance systems in cars add complexity, but they also objectively make cars safer. Same with controls on the steering wheel, they add complexity because you now often have two places for things to be controlled and increasingly have to rely on drive by wire systems, but HOTAS interfaces (Hands On Throttle And Stick) help to keep you focused on the road and make the overall system of driving safer. While semantic modelling and control systems absolutely can make things less safe, if done well they can also actually let a robot or machine act in more human ways (like detecting that they’re injuring someone and stopping for instance).

Direct control over systems without unreliable interfaces, semantic translation layer, computer vision dependancy etc serves the same tasks without additional risks and computational overheads.

But in this case, Waymo is still having to do that. They’re still running their sensor data through incredibly complex machine learning models that are somewhat black boxes and producing semantic understandings of the world around it, and then act on those models of the world. The primary difference with Waymo and Tesla isn’t about complexity or direct control of systems, but that Tesla is relying on camera data which is significantly worse than the human eye / brain, whereas Waymo and everyone else is supplementing their limited camera data with sensors like Lidar and Sonar that can see in ways and situations humans can’t and that lets them compensate.

That and that Waymo is actually a serious engineering company that takes responsibility seriously, takes far fewer risks, and is far more thorough about failure analysis, redundancy, etc.

I don’t misunderstand how they work at all.

Quite frankly what you’re saying doesn’t matter in the context of my point. It literally does not matter whatsoever that they are not logic based but language based, as long as they produce helpful results, and they inarguably do. You are making the same types of point that my middle school librarians made about Wikipedia. You’re getting hung up on how it works, and since that’s different than how previous information sources work, you’re declaring that they cannot be trusted and ignoring the fact that regardless of how they work, they are still right most of the time.

As I said, it is far faster to ask copilot web a question about salesforce and verify its answers, then it is to try and manually search through their nightmarish docs. Same goes for numerous other things.

Everyone seems so caught up in the idea that it’s just a fancy text prediction machine and fail to consider what it means about our intelligence that those text prediction machines are reasonably correct so much. All anthropological research has suggested that language is a core part of why humans are so intelligent, yet everyone clowns on a language based collection of simulated neurons like it can’t have anything remotely to do with intelligence.

… Y’know, everyone sucking off “AI” while it’s still wrong a great number of times kinda makes sense, when you factor in how fucking stupid most people are…

Everyone who clowns on AI for being wrong sometimes sound like my hysterical middle school teachers talking about how you can’t trust wikipedia because anyone could edit it.

There are lots of systems that know they will inherently have errors, and need to rely on error correction mechanisms to accomodate. The computer RAM in satellites is constantly being bombarded with cosmic rays and having bits randomly flipped, but it can account for this by having error correcting memory. At a simplified level, Error Correcting functions like this: when you write data you can write your output to three bits instead of just one, then to read a single bit, you instead check all three and discard any outliers. ECC memory actually uses more complicated math so that it only has to store 8 error correcting bits for every 64 normal ones, but that is the general principle of error correction.

Similarly, quantum computers have been proven to have inherent fluctuations and unpredictability in their results due to the underlying nature of quantum mechanics. But they are still so much faster at solving certain problems that you can run them multiple times, discard the outliers, and still get your answer orders of magnitude faster than a classical computer.

AI being wrong sometimes is like this and this is why not everyone thinks it’s a huge deal. Copilot web can still parse and search the nightmarish spider web of salesforce docs and give me an answer orders of magnitude faster than I can, even using Google. It doesn’t matter if it’s occasionally wrong and it’s answers require me to double check them when it’s that much faster to give each answer.

People over use blocking like crazy.

I constantly see people blocking others just for making a point they disagree with. Rather than actually think through the logic and reasoning of what the other person is saying they go ‘oh I have no counter point to that, that must mean that you’re arguing in bad faith, blocked’.

The internet is already an inherent filter bubble, you don’t need to accelerate that. Most people would benefit from spending more time deeply considering that they might be wrong in ways they can’t fully comprehend, then they would blocking people who fervently disagree with them.

“hur durr AI bad”

Read the fucking link. It literally won the Nobel prize.

They should. That’s how automation works. We should be building a society that doesn’t require as much work, not insisting on doing work that machines could do when we don’t want to.

You keep saying y’all and it’s telling.

Learn how to communicate with people, not the simplified boxes you put them in.

When you’re ready to have a conversation instead of just hearing yourself regurgitate mindless internet grandstanding I’m here.

LLM is what usually sold as AI nowadays. Convential ML is boring and too normal, not as exciting as a thing that processes your words and gives some responses, almost as if it’s sentient.

To be fair, that’s because there are a lot of automation situations where having semantic understanding of a situation can be extremely helpful in guiding action over a ML model that is not semantically aware.

The reason that AI video generation and out painting is so good for instance it that it’s analyzing a picture and dividing it into human concepts using language and then using language to guide how those things can realistically move and change, and then applying actual image generation. Stuff like Waymo’s self driving systems aren’t being run through LLMs but they are machine learning models operating on extremely similar principles to build a semantic understanding of the driving world.

I am treating you like a child because you refuse to use your brain.

No you’re doing so because you started doom scrolling before you had coffee and now you’re trying to justify your uncalled for rudeness.

You gave me one obscure

It literally won the nobel prize.

very early stage example

It is not early stage, predicting the structures of those proteins has already actively changed the course of biomedical science. This isn’t early stage research that need fleshing out, this is peer reviewed published research that has caused entire labs and teams to completely change what they’re doing and how.

that isn’t even connected to the overall rise in value of LLMs and other forms of AI

It is in that it uses the same underlying type of algorithms and is literally from the same team that developed the “T” in ChatGPT.

So you are claiming the next real AI revolution is justtttt around the corner with a totally new technology you swear?

I have not claimed that, I said that AI algorithms are likely to be part of our climate solutions and our ability to serve more people with less manual labour. They help to solve entirely new classes of problems and can do so far more efficiently than years of human labour.

Rage out about tech bubbles and hype bros if you want. Last time it was crypto, streaming before that, apps and mobile before that, social before that, the internet before that, etc etc. Hype bubbles come and go, sometimes the underlying technology is actually useful though.