Separately from the other thread—the little evidence I’m aware of (bing chat, sparks of AGI, absence of evidence on safety) suggests that Microsoft is bad on safety. I’m surprised they were included.
Edit: and I weakly think their capabilities aren’t near the frontier, except for their access to OpenAI’s stuff.
Separately from the other thread—the little evidence I’m aware of (bing chat, sparks of AGI, absence of evidence on safety) suggests that Microsoft is bad on safety. I’m surprised they were included.
Edit: and I weakly think their capabilities aren’t near the frontier, except for their access to OpenAI’s stuff.