Original post unavailable.
Sentiment
Many users agreed with the Arcade CEO that permission layers and human oversight limit enterprise AI agents more than raw intelligence, while others highlighted risks like legal and compliance issues with full autonomy.
Pos
66.7%
Neg
33.3%
14 comments with sentiment.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Related links
Posts from X
Most Activity
Most Activity
VIEWS8.2KBOOKMARKS19LIKES58RETWEETS4REPLIES21
AI agents are getting smarter fast. But intelligence is not what stops them inside real companies. The harder question is: What are they allowed to do - and who can prove they followed the rules? In our new conversation, Arcade CEO Alex Salazar (@TheMostlyGreat) explains why permissions are the real enterprise bottleneck, why Human + Agent is beating full autonomy, and why his team is already trying to live in a post-browser world. YouTube link in comments as well! Really cool conversation we had with him. My partner Peter and I have already conducted some cool interviews with experts in the AI field and will continue to do so. So feel free to follow us on our channels!
@TheMostlyGreat (yes, I wore that tshirt, iykyk)
AI agents are getting smarter fast. But intelligence is not what stops them inside real companies. The harder question is: What are they allowed to do - and who can prove they followed the rules? In our new conversation, Arcade CEO Alex Salazar (@TheMostlyGreat) explains why permissions are the real enterprise bottleneck, why Human + Agent is beating full autonomy, and why his team is already trying to live in a post-browser world. YouTube link in comments as well! Really cool conversation we had with him. My partner Peter and I have already conducted some cool interviews with experts in the AI field and will continue to do so. So feel free to follow us on our channels!
@TheMostlyGreat YouTube: https://www.youtube.com/watch?v=eka3kyDRygI
AI agents are getting smarter fast. But intelligence is not what stops them inside real companies. The harder question is: What are they allowed to do - and who can prove they followed the rules? In our new conversation, Arcade CEO Alex Salazar (@TheMostlyGreat) explains why permissions are the real enterprise bottleneck, why Human + Agent is beating full autonomy, and why his team is already trying to live in a post-browser world. YouTube link in comments as well! Really cool conversation we had with him. My partner Peter and I have already conducted some cool interviews with experts in the AI field and will continue to do so. So feel free to follow us on our channels!
@vishalsingh2972 @TheMostlyGreat Added chapters and timestamps on YouTube
@kimmonismus @TheMostlyGreat Awesome but I wish this had timestamps 😕
@kimmonismus @TheMostlyGreat Yup currently watching it there ❤️
@vishalsingh2972 @TheMostlyGreat cool! Thank you :) And thanks for bringing it to my attention
@vishalsingh2972 @TheMostlyGreat hmmm good call. Havent thought about it. Will add some!
@kimmonismus @TheMostlyGreat yeah, permissioning is where enterprise agents start getting complicated
@kimmonismus @TheMostlyGreat 'Human + Agent beats full autonomy' is the whole thesis. My one push on 'post-browser': the browser is where that proof is cheapest. You watch the agent act where you already live, and grab the wheel the moment it drifts. Hard to rebuild that legibility behind an API.
we need the intelligence because it's the next crossing point for all of us. we are now spending our time teaching these agents what we want to do because they aren't smart enough to do it on their own like a super smart human would we need more intelligence, and we are getting it, what a time to be alive!
@kimmonismus @TheMostlyGreat @kimmonismus it's wild how often companies forget the second part. rules are key, but who's checking? that's the real challenge.
@kimmonismus @TheMostlyGreat Figuring out who can audit the agent's steps might matter more than raw IQ.
@kimmonismus @TheMostlyGreat The next wave of AI will be defined by secure execution and verifiable actions, not just better reasoning.
@kimmonismus @TheMostlyGreat Full autonomy sounds cool until legal and compliance show up.
@kimmonismus @TheMostlyGreat The content turned out really cool!
@kimmonismus @TheMostlyGreat permissions and the audit trail are the real bottleneck not the ai brains themselves especially on chain
Load more posts
@sunglassesface @TheMostlyGreat :D :D :D