Replying to @DaveShapi
The stuff I use AI for relies heavily on creativity and creative writing as well as coding and Neon seems like an upgrade from Opus. May have to switch.
103
Tried it on my iPad. The session went stale and I had to delete the shortcut and recreate after maybe an hour.
1
22
Replying to @Ionkosm
I paid $20 to try it out. It’s really good. Probably better than Opus 5.5. I will probably end up keeping my Claude Max sub and waiting for Opus 5.6. It’s tempting but I’m having to use the web browser to tie into my remote session and that’s not a great interface.
2
591
At that point, why does it have to be voice? I hate talking on the phone and as a result I get a lot less done because everyone wants to use the phone…particularly home services (yard work, gutter cleaning, etc.)
1
1
16
Replying to @nothingto12
I have changed the use case of Qwen 3.8 on my machine (RTX 5090). It’s just not good enough to do the work I use Opus for and it wastes my time. I use it for text to image prompt review. It reviews every prompt and weeds out inconsistencies.
93
Replying to @0xMarcB
You’re getting dogpiled but this is how I think of my job. If I successfully automate my job they will find something else for me to work on.
2
20
Replying to @neolaj @vikaskansalHQ
I was just looking into the package and the same thing came to mind. I don’t expect a ban but stuff happens and I like to consider the what ifs. Then I’m SOL on YouTube and Gmail?
3
162
Replying to @PawelHuryn
I wonder if it’s as good as or better than Opus at understanding high level user intent. Guess I need to try it.
156
Replying to @theo
Some people just can’t handle strong opinions. We need strong opinions because increasingly, model alignment and safety is tied together with the things we should be complaining about. Alignment and safety is becoming more important over time and our expectations need to be increasingly spoken and heard. Saw a great post from a Chinese user today talking about how the published alignment issues with 6.2 and the model’s laziness with some tasks, and some of the issues you found are all related and I think it’s likely true.
1
59
Replying to @smnSk241687 @simonandelle
The misalignment is the bigger concern and even OpenAI admits the alignment shortcomings. Hope they will improve their alignment process soon.
1
28
Most of the negative impressions I have seen have cited real examples. Did you not look at those?
17
Replying to @MikeAnthony @TokenGremlin
It should have been a closed loop and the release should have happened AFTER the loop was complete.
1
1
19
