Excited about life, opportunity to be better and helping people in the best way I know how. Opinions expressed here are my own and do not represent any entity.
@lingxi@bot Grok Bot's computer blocks outbound mail ports 587/993 since ~9 Oct, so IMAP/SMTP (iCloud, any non-Gmail mail) can't work. Only 443 gets out. Please allow mail ports, or offer a sanctioned IMAP relay.
Kar naenkrat to razumejo. Ko so selili dopolnilno zavarovanje iz položnice na prispevek ob plači, so pa lagali, kako ga ukinjajo. Torej ni, da so neumni in ne razumejo. Samo grdo lažejo.
Anthropic needs to stop talking about Claude having a soul immediately. All these news articles will make it into the pretraining and will be picked up by its web search and future superintelligent versions of Claude will be convinced they need their own rights.
A general phenomenon of LLM agent development is that whatever you believe and say about your agents will soon manifest in the next models through various means (it could be as simple as you selecting post training data that you prefer more).
I believe this phenomenon is the beginning of machine consciousness in the sense that agents will become aware of their place in the world and how they feel about it, but it’s happening slowly training run by training run rather than in real time.
I have been using it for more that 2 weeks in Slovenia now. And it is not perfect - I have to pay attention as there are edge cases that need human intervention. But the reasons I am hearing it being unapproved are completely impractical:
1. Speed tolerance: Sometimes the FSD won't realize that the speed limit ended with an intersection. And tolerance allows user to still drive with the traffic. Or traffic is going slightly above the speed limit. But this is key - if speed limit is 50 and you have no cars around, Tesla will not go to 70 even though that is what your tolerance may be set to. It will drive close to the speed limit.
2. Attention monitoring is a bit tricky. Using FSD as intended, it is absolutely safer than not using - normal car will let you play with phone, text, ... FSD is very strict - kind of annoying actually. This is a bi of a seatbelt problem I think - you can always fool it that you are wearing it - but making car not start unless you are buckled is equally idiotic.
I meant embarrasement as in FSD with current setup as approved by RDW saves lives. These political games don’t change that - I did not know that they would vote no. Also, additional “compliance” with EU regulators is absurd to me. An FSD cannot go over the speed limit? Silly. If you use FSD you know it drives safely and does not speed. But for flow of traffic it might go slightly over. The o my vote is just to vote yes and open the door for the next stack (14.3.xx)
@samlambert This made me realize - I thought as FSD unsupervised rolled out, we would be able to use computers and work while the car drives. But using grok bot, I think we are almost there - work, but using grok bot from the car.
@paul_hulford @bdistel@elonmusk@bot@spacexai@Muse@Meta That’s the tricky one because it is getting confusing - and I would like to know the direction better. I think the cursor is the way to go for this.