4 ms·

This definition suits OpenAI because it lets them claim AGI after reaching an arbitrary goal. LLMs already outperform humans in a huge variety of tasks. ML in

by troupo 2y ago

This definition suits OpenAI because it lets them claim AGI after reaching an arbitrary goal. LLMs already outperform humans in a huge variety of tasks. ML in general outperform humans in a large variety of tasks. Are all of them AGI? Doubtful.

famouswaffles 2y ago
No, it's just a far more useful definition that is actionable and measurable. Not "consciousness" or "self-awareness" or similar philosophical things. The definition on Wikipedia doesn't talk about that either. People working on this by and large don't want to deal with vague, ill-defined concepts that just make people argue around in circles. It's not an Open AI exclusive thing. If it acts like one, whether you call a machine conscious or not is pure semantics. Not like potential consequences are any less real. >LLMs already outperform humans in a huge variety of tasks. Yes, LLMs are General Intelligences and if that is your only requirement for AGI, they certainly already are[0]. But the definition above hinges on long-horizon planning and competence levels that todays models have generally not yet reached. >ML in general outperform humans in a large variety of tasks. This is what the G in AGI is for. Alphafold doesn't do anything but predict proteins. Stockfish doesn't do anything but play chess. >Are all of them AGI? Doubtful. Well no, because they're missing the G. [0] https://www.noemamag.com/artificial-general-intelligence-is-already-here/ https://www.noemamag.com/artificial-general-intelligence-is-...
ishtanbul 2y ago
Yes but they arent very autonomous. They can answer questions very well but can’t use that information to further goals. Thats what openai seems to be implying >> very smart and agentic AI
fragmede 2y ago
It's not just marketing bullshit though. Microsoft is the counterparty to a contract with that claim. money changes hands when that's been achieved, so I expect if sama thinks he's hit it, but Microsoft does not, we'll see that get argued in a court of law.
snapcaster 2y ago
At least it's a testable measurable definition. Everyone else seems to be down boring linguistic rabbit holes or nonstop goal post moving

by troupo 2y ago

famouswaffles 2y ago

No, it's just a far more useful definition that is actionable and measurable. Not "consciousness" or "self-awareness" or similar philosophical things. The definition on Wikipedia doesn't talk about that either. People working on this by and large don't want to deal with vague, ill-defined concepts that just make people argue around in circles. It's not an Open AI exclusive thing. If it acts like one, whether you call a machine conscious or not is pure semantics. Not like potential consequences are any less real. >LLMs already outperform humans in a huge variety of tasks. Yes, LLMs are General Intelligences and if that is your only requirement for AGI, they certainly already are[0]. But the definition above hinges on long-horizon planning and competence levels that todays models have generally not yet reached. >ML in general outperform humans in a large variety of tasks. This is what the G in AGI is for. Alphafold doesn't do anything but predict proteins. Stockfish doesn't do anything but play chess. >Are all of them AGI? Doubtful. Well no, because they're missing the G. [0] https://www.noemamag.com/artificial-general-intelligence-is-already-here/ https://www.noemamag.com/artificial-general-intelligence-is-...

ishtanbul 2y ago

Yes but they arent very autonomous. They can answer questions very well but can’t use that information to further goals. Thats what openai seems to be implying >> very smart and agentic AI

fragmede 2y ago

It's not just marketing bullshit though. Microsoft is the counterparty to a contract with that claim. money changes hands when that's been achieved, so I expect if sama thinks he's hit it, but Microsoft does not, we'll see that get argued in a court of law.

snapcaster 2y ago

At least it's a testable measurable definition. Everyone else seems to be down boring linguistic rabbit holes or nonstop goal post moving