I. From GPT-4 to AGI: Counting the OOMs
SITUATIONAL AWARENESS - The Decade Aheadsituational-awareness.ai
I. From GPT-4 to AGI: Counting the OOMs
How did we get to the doorstep of the next leap in prosperity?
In three words: deep learning worked.
In 15 words: deep learning worked, got predictably better with scale, and we dedicated increasing resources to it.
That’s really it; humanity discovered an algorithm that could really, truly learn any distribution of data (or really, the underlying “ru... See more
In three words: deep learning worked.
In 15 words: deep learning worked, got predictably better with scale, and we dedicated increasing resources to it.
That’s really it; humanity discovered an algorithm that could really, truly learn any distribution of data (or really, the underlying “ru... See more
samaltman.com • The Intelligence Age
Ok, so the rumor about OpenAI Orion is that it's "100 times more powerful than GPT-4".
Let's forget that this is ill-defined, and a rumor. It implies gruesomely superhuman generalist ability.
The tech media describes this potentiality like so:
"The model’s development builds on lessons from OpenAI’s past efforts to make interactions with AI more relia
... See moreo, alternative pathways to building Type-2 reasoning-capable AI systems, likely using neurosymbolic approaches, have become much more attractive. People like Gary Marcus have argued for neurosymbolic approaches for decades. Such approaches combine the pattern recognition of neural nets, like LLMs, with symbolic reasoning’s logic and rules. Vinod Kh... See more
Azeem Azhar • 🧠 AI’s $100bn question: The scaling ceiling
“From this point on, the intelligence of LLMs… will only continue to improve. Human intelligence will not.”
OpenAI’s paper on using AI to debug AI code