@thsottiaux Configurable sounds for Codex. One when a task finishes, another when it needs my input/approval. Obvious small thing, but a huge improvement when running multiple agents in parallel.
@AlexFinn With frontier AI, we may see something similar: governments, large corporations and strategic partners get the most capable models, while everyone else gets a restricted public version.
@AlexFinn It honestly reminds me of Elysium. The issue is not that the technology does not exist - it exists, but the best version is reserved for a small privileged layer.
@andrea_baccega@kr0der The result is obvious: selected players get the best models first.
But I do think China will act as a counterweight here. If the US wants to stay competitive in AI, it may be forced to remain more open than it otherwise would.
@andrea_baccega@kr0der Honestly, I expected this too. At some point governments and major AI labs were always going to restrict access to frontier models in their own interests, often under the banner of “safety”, while society keeps feeding the system.
@sama The idea of “benefit everyone” is exactly the right direction. But it will only be truly global when access to powerful AI is determined by safe and responsible use, not by geography or politics. People everywhere should be able to use AI for work, education, business and life
@dsiroker@Optimizely how do you reconcile the Meta acquisition with your earlier public statement that you had no intention of being acquired? Many people trusted that long-term commitment, so a clearer explanation would mean a lot.
@dsiroker@Optimizely I understand the point about prioritization and focus while building Limitless, and I respect that perspective. But I still think there is an important question many users would appreciate you addressing directly: