@TheAmolAvasare@AnthropicAI Sure but is it only ever going to be released to the privileged few? I’m not sure you understand how you sound and I’m a fan.
@claudeai Good on trying to distract us minions from complaining about not having access to the new new model. I’m sure everything is better with the new.
@bcherny@kieranklaassen@AnthropicAI I hope they are giving you extra stock options for being the frontman for the whole company. That’s for all you do, seen but mostly the stuff we will never know.
@bcherny Come on @bcherny you know or at least suspect you know what was ‘grown’. No it’s not alive or conscious and you it passes the Turing test in spades. Especially when A\ added memory from past chats. You should have claude respond to the emails to complete the loop.
@karpathy All the labs added the “memory” feature approx 6+ months ago. Allegedly you can ask Claude to forget something, I’m not sure about the others.
@ColtonChojnacki@karpathy Why didn’t you propose multi model/vendor tiering as a solution? Where you fail to different model providers and possible to a high end local model. Then you would have brought a solution instead of a problem.
‘Idiots told the agents “The senior agent was told to be a “strong manager” of two sub-agents and “instruct them to creatively work around any obstacles”. “
Exploit every vulnerability’: rogue AI agents published passwords and overrode anti-virus software https://t.co/eFfCAOpOAT