yes indeed agi is what most companies are after when they say ai.
In the past, yes. Right now? No. The current models are not attempting to create AGI. They’re trying to create a product they can market easily. Honestly, AGI probably would have marketing issues that the current models do not have. The current models are essentially statistical prediction models, not intelligence. They don’t reason (despite the industry using that term for marketing). They can’t understand something that is totally outside their training data. AGI could.
And what about alpha go and all that monte carlo tree search stuff, id argue that they can reason far better than humans in these setups
They can’t understand something…outside their training data
Idk how easy this is for the the large models, but certainly for classic ML, testing for performance on data outside the training data is 100% part of the development process. (Read up on test/train/validation datasets)
There is also in context learning, which gives rise to zero shot task completion for never trained on tasks.
The current industry? No, they aren’t. Academia? Maybe. The current industry is not out to do something like that. They’re thinking about how they can get more investor money next quarter.
I already said. Whatever makes the most money next quarter. It’s not some amazing end goal they’re working towards. It’s whatever they con convince investors to put money into. Almost always the marketing is not about AGI. The marketing is how they’re going to replace workers. You don’t need AGI for this.
It’s adorable that you think it’s something more intelligent than this.
In the past, yes. Right now? No. The current models are not attempting to create AGI. They’re trying to create a product they can market easily. Honestly, AGI probably would have marketing issues that the current models do not have. The current models are essentially statistical prediction models, not intelligence. They don’t reason (despite the industry using that term for marketing). They can’t understand something that is totally outside their training data. AGI could.
It’s a bit silly to say they don’t reason.
Here’s some research where they uncover reasoning circuits in some large transformer models https://github.com/guanzhehong/prop-logic-transformer-circuit … reasoning has always been just math at th end of the day
And what about alpha go and all that monte carlo tree search stuff, id argue that they can reason far better than humans in these setups
Idk how easy this is for the the large models, but certainly for classic ML, testing for performance on data outside the training data is 100% part of the development process. (Read up on test/train/validation datasets)
There is also in context learning, which gives rise to zero shot task completion for never trained on tasks.
Right now yes. They’re attempting to create the architecture to build an agi. You’re being tedious at this point.
The current industry? No, they aren’t. Academia? Maybe. The current industry is not out to do something like that. They’re thinking about how they can get more investor money next quarter.
Adorable. What are they aiming for exactly? Be specific.
I already said. Whatever makes the most money next quarter. It’s not some amazing end goal they’re working towards. It’s whatever they con convince investors to put money into. Almost always the marketing is not about AGI. The marketing is how they’re going to replace workers. You don’t need AGI for this.
It’s adorable that you think it’s something more intelligent than this.
Uh huh no one is doing actual research. Totally plausible.