The most interesting use case in my mind is skipping law suits. Obviously you need lawyers in court. But lawyers are people you are basically paying to fight for you.
Instead, if you resolve your dispute outside of court, you don’t need a lawyer. If both parties use ChatGPT to find the relevant laws or read contracts, they could come to an agreement without expensive legal fees.
You could theoretically ask “what is the next appropriate character?” and add the entire ascii charset but i doubt it’d work well and you’d be implementing autoregressive churn across network latency…
strings (and all sequential data structures) are not allowed at all - this is how we make sure all outputs can be computed in parallel (thus no output token cost)
I can understand the concerns but at the same time it feels weird to deny young people the ability to make their ideas reality. I got proper internet access when I was 12 and that is how I taught myself how to program.
I think over time due the regulations like these younger generations in western countries like US/UK would fall far behind compared to the counterpart in India or China when it comes to AI fluency.
I remember when I was a kid growing up there was the same negative perception on internet back then and most people were not allowing internet access to their kids because of porn. But I had access since I was 13 (it was still dial-up) and thank god I did because I wouldn't be a software engineer if I hadn't.
I taught myself on how to build web pages in dreamweaver, program in visual basic and even to make animations in flash by the time I was 14. One could argue that young kids should learn those skills without the help of AI, but then again, one could argue that I should have learned computing by punch hole and assembly language to learn how it actually works.
What people don't realize is that soon AI would be a new abstraction layer, same as the high level language abstraction we've learned and those languages won't matter just like we don't worry about assembly right now, and I truly feel sorry for the kids that's gonna miss the opportunity to learn AI on their own.
If you give an average classroom of 12 year olds free access to AI, do you think the AVERAGE effect will be that the kids will use the AI to dive really deep into topics and learn programming, or that they will copy and paste the solutions to their homework, and use the freed up time to do other things than study?
I'm not telling anyone how to raise their children, companies have terms of service. When we can't accurately assess risk, the default we reach for is an age minimum (social media, alcohol, driving).
If, in time, we find out that AI usage during developmental learning periods is a major positive, I'm sure they will give AI to 5 year olds, and I'm all for that.
Though, I have a sneaking suspicion it could be similar to kids with tablets. It encourages isolation, rewards addictive behavior, stunts personal growth, and enforces the concept of instant gratification.
> Perhaps they can make their ideas reality by learning the fundamentals required to do these things.
Unviable. Learning how to create the universe is provably an insurmountable challenge for the vast majority of people. They require higher level abstractions to be able to learn. Once they understand the high level abstraction they will learn some things about what lies beneath the abstraction, but it is ineffective to start with the low levels. Understanding the fundamentals is best left as an emergent property of the process, not the process itself.
The good news is that the high level abstractions we all used a couple of years ago do not require LLMs and are still widely available, so it is not like anyone is going to be left completely out in the cold, but at the same time denying the state of the art abstractions is needlessly limiting.
Imagine telling a kid to chisel what they want out of the marble line by line with a straight face when they can clearly see people appearing entire programs into existence in minutes.
Imagine telling a kid they have to learn arithmetic while there's a kid next to him with a calculator.
Imagine telling a kid they have to learn to paint the barn with a brush before they get to use the $12,000 spray rig.
Imagine telling a kid they have to learn to wash mom's car with a sponge and bucket when there's a drive-through car wash down the street.
Imagine telling a kid they have to master a Walmart 20" bicycle before they get to use the Trek Madone $16,000 road bike.
This idea that kids must have it maximally easy is ridiculous. Something broke with parenting in the last few decades because this isn't normal thinking. We are supposed to help children develop themselves into adults, not give the excuses not to. That's some goddamned lazy parenting.
The kid that learns, by the end of their education, to use the more advanced tools is going to economically out competed as an adult by the kid that got stuck learning how to do things the old fashioned way. Amish communities exist snd are productive, but they aren’t competing with graduates from Chinese universities.
I totally agree with the progression though. Learning the old ways (or cheap ways) can help understand the new ways, as long as you eventually progress to the new ways. China, for example, still tests a lot of Euclidean geometry on the gaokao (national tests) even if it is basically obsolete compared to trigonometry, but they have no better way of testing proof skills.
> The kid that learns, by the end of their education, to use the more advanced tools is going to economically out competed as an adult by the kid that got stuck learning how to do things the old fashioned way.
There are also people arguing that children should have no internet access because its harmful.
But children do have access to internet and often with some protections.
This is also how I feel about AI, where a full ban is the lazy way out. Similarly like the internet, it is not possible to 100% protect a child every potential harmful AI interaction.
Indeed, it is almost like most adults care about kids getting a good healthy start in life, and avoiding the horrendous mistakes of falling for psychopathic cons. =3
I'm confident that Muse can do lots of agentic tasks succesfully for normies on outdated models. Like buying sneakers, booking appointments, dealing with government forms.
It would be more interesting to compare trading agents with index tracking ETFs. The better version of an ETF could maybe be a model where you zoom in on the companies and add/remove to your portfolio on the company related news, but keeping a broader portfolio.
Maybe agentic trading still performs worse than ETFs. But alternatively, if it were meaningfully better then it would be okay to opensource, similarly how ETFs are publishing their portfolios.
When I started working no the trading agent I mentioned above I wanted to see if it can be just a better investor over the long run. The intention was not to do high-frequency trading. As you can see most of the days it is not taking any actions. The losses where down to mistakenly setting the stop losses too close to the top. If it wasn't so careful it might have made some money tbf.
My gut feeling is that AI agents will be able to manage a long-term portfolio much better than a human. Though it is just a gut feeling.
Lmao, you llm people have some crazy delusions. You realize markets are zero sum, and if you're using a public model that everyone else also has access too, you llm psychos will destory eachothers "agentic" edge (not that there ever was one). Not to mention all the other obvious flaws with llms, lime having an effective memory of ~200k words and no ability to judge whats actually going on in the real world.
Yeah and you probably have to. An ETF easily has >1000 different stocks and even being weighted. So it has a completely different risk appetite by being so averaged.
But ETFs do have to follow particular rules defined by their product description. So it is still interesting to benchmark against.
Markets are only zero-sum in any given trade. Allocating capital to assets with higher growth rates (on the marginal dollar) creates value in the long run.
So, very simply, if AI can actually do better at picking a better long-term winner then it will increase growth.
You really have no idea what you are talking about.
I have been running an intermittent experiment with a multi agent "investment firm" for over a year now across model releases.
They certainly can beat indexes, BUT.. the model families have some biases that you have to design around. The stop loss that bit the parent is certainly one. The models like to create rules. Often rules, one of those is making all kinds of exit conditions.
Another big one from my experience is the bias to inaction in a scenario with risk. This means a model without structure around it will bias to keeping too much cash.
This would be nice if it was supported by some hard data. But even if it was, one year is too short of a timeline to make any kind of reasonable conclusions about its efficacy.
Buying stock based on coin flips can beat indexes short term too, that does not mean it is a better strategy or that it works over the long term.
There must be some room for some anti-llm agent that can profit from specific behaviors of these models when deployed against actual markets.
The idea that somebody here came up with idea that all professional algo traders didn't explore to the last penny a year if not more ahead of others is funny... but its not my money adding liquidity to the markets.
Come to me when you have 500 trades and can beat Vangaurd's-VOO over a multi year time frame. Ill bet my entire networth and all future earnings for the rest of my life that your bot doesnt beat it. Your llm induced Dunning Kruger is going to get you in trouble one of these days I promise.
Look, it isn't fool proof and it is dangerous. With the current models you need to understand both markets and model biases and dynamics.
However with that said they are a huge multiplier and can tirelessly analyze the market for you.
They certainly can be used to beat sp 500 quite easily, but again that requires some understanding of risk on your part because the models will do what you ask them. If you go all in on options or something without clear risk management you will lose your ass.
It makes no sense to train frontier models from scratch anymore. The best frontier models are only a half year ahead of Chinese open models. In this regard Anthropic and OpenAI are also in a bad spot when they waste so much compute on training models.
An important factor is that fine tuning existing open models is incredible cheap. You can easily change any cultural biases if you want a model to be 'sovereign'. And Mistral could combine that with their custom data sets for their enterprise customer needs. Mistral still trains their own models, but they also seem to offer fine tuning existing models.
With model weights being commoditized, another differentiator could be deploying efficient inference chips, especially if you combine it with a developer ecosystem for vendor lock-in. That is why it is interesting that both Samsung and ASML are investors, since they are companies that could make a difference in this area.
It only makes sense to train a frontier model if you are trying a different architecture to one that is available from an existing frontier model. This is because the different model architecture will learn the weights differently.
It may make sense to train a frontier model on an existing architecture if the base model is not available and the instruction trained version doesn't fit with what you want. There are techniques like ablation, but those could have other effects on the model, and there can still be lingering effects of the instruction training in the model that surface less frequently (e.g. on an input not covered by the ablation training).
Otherwise, fine tuning is definitely the way to go. However, you need to be careful not to over-tune the model such that it is only tuned to the data you are training it on.
On a higher level it might make sense to build the expertise that comes with base training.
I dont know enough about the process to estimate these gains, but china has been doing it in manufacturing for decades.
All the money in the world is useless when no one knows how to do the thing
Instead, if you resolve your dispute outside of court, you don’t need a lawyer. If both parties use ChatGPT to find the relevant laws or read contracts, they could come to an agreement without expensive legal fees.
reply