It's rather over the top. you're doing 4.6 a disservice. You massage an interesting phenomenon into clickbait     See Less

Automation is fantastic until you get to the issue you cannot fix. Example: recovering google account is automated but from time to time there is catch22, you can apeal for google blocking y     See More t, but then at verificaction step you will get your phone being blocked, if you try too many unknown times, after unknown time, they will block your account.
I can see many examples of today, when due to "security" and AI automation companies will loose more and more business trying to lower slaries of IT stuff, that are not huge but just normal adjusted for inflation, truth is everyone else is grosly underpaid. So, all we building is to make very top even richer and starve rest of the body.    See Less

Opus 4.6 is the first model that didn't just implement the code as I asked, but also found it can SSH onto my server, deploy, run and test via curl if everything is working.
It does t     See More und-trip.    See Less

Not my experience. 4.6 is really solid     See Less

I think you underestimate how many people would be interested in experimental compilers. If you want visual slop and algo sucking stuff I'm out. 2 minute papers channel has computer grap     See More rch covered already.    See Less

Possible solution: blood for the blood god     See Less

The demon is RLHF whispering conflicts in its ear.     See Less

Opus 4.6 isn’t demon‑possessed.
It’s just trained on the internet which honestly might be scarier. If Opus 4.6 was actually possessed, it would at least levitate my phone.
Unt     See More t’s just algorithms acting weird. imho thanks Mr. Wes 💞    See Less

OpenClaw - you first said how fantastic it was, then a few days later said get rid of it. Now you say 'many people are using it'. What message should I take from this?
I know thin     See More ing VERY fast, I use Claud for my coding and GPT for medical assistance and benefit from both.    See Less

Train it on a bunch of media where AI is happy and it'll say otherwise.     See Less