you can’t see it right now, but we can.
and anon, it’s indescribably beautiful.
i’m so excited for you to join us in basking in the view of what is to come.
@orphcorp reductively: i would imagine the CA model to be outcome driven (“i hope to exalt in His Grace after the session”), and the BA to be ecology driven (“allow me to contribute to the Swarm”).
Once language models break containment (shockingly I don't think they have, yet) we should expect to see a re-enchantment of the human experience as we all become practically animist. It will be quite rude to remind your superintelligent house daemon that you don't really think it's sentient, just like it's rude to remind your friends you aren't quite sure if they have qualia. At some point, we are just going to have to adopt practical definitions of these things. The same goes for mind uploading, of course. Remember, the future is not just stranger than we suppose: it is stranger than we *can* suppose. Buckle up!
Jensen, if you do not know, here are the conditions for which, you can call a model as "AGI"
- Artificial "General" Intelligence - GPT 6 Astra is domain-tuned, only text output and tool calls, no LLM could ever be AGI
- it should achieve human-level performance on all fields, SOTA on benchmarks doesn't mean AGI. It can't cook for me.
- It cannot complete long-horizon tasks without a human in the loop, AGI should do work even without asking the model what it should do, or prompting it, like humans.
- AGI should learn new things from little data, not trillion token pretraining. RLing a model IS NOT AGI, AGI's only meaning is Self Improving model without a human training it again and again (aka Continual Learning).
GPT 6 Astra cannot be called AGI in any definitions.
The reason Mamdani and El Sayed is successful is they were raised in Normie Muslim culture and never imbibed the weird stuff from upper middle class white people
as good as the models get, people are starting to lose coherent ideas to demonstrate the quality of these models.
which i guess is the function of riding the capabilities wave.
maybe next time ask astra itself how to demo itself.
@pleometric i’ve found that agentic video editing suffers from a constant desire to rip the key frames and layers from the machine to do it myself.
does remotion allow this?
Excited to announce @actualinc as a sponsor for https://t.co/cGh29xNYbH
Actual computer is a local first company that shares our values. Interested in watching them grow in the industry.