To be honest, I just think it was poorly worded. I am about 90 percent certain what he meant was that when life gets too busy and you are unable to talk to your kids about these things, you can easily get caught up by using data sources like your calendar.
So at the very least, you have an idea of what happened, and then maybe the next day you can talk to them about it (because it will make a world of difference to the kids that you remembered). I think the way that the tweet ended up coming out, though, was a little tone deaf.
Like all those complaints from people constantly burning through their subscription quota, the prompt matters, what you're trying to do matters, and how you approach it matters.
I wouldn't use Luna and ThreeJS to build a Roman Colosseum (or whatever that example is), but I've replaced almost all my light to moderate coding and agentic tasks with it. It's ridiculously fast, works flawlessly, and barely touches my usage limits.
I know that entire industries exist just to quantify productivity gains on these tools, but they're non-deterministic by default, and then on top of that, everyone uses them so differently that broad, black-and-white claims just don't hold up.
I think when we hear advice like use Luna Max because it's the new Sol Medium. We have to take that nuance into account :)
@kirtandopamine Alternative take: I hope tech companies never reach a point where fun stuff like this disappears. It's what made Google great for so long. Hope you don't take that the wrong way <3
I think Sam could have possibly framed that a little bit better. I could see the use case he's describing makes sense if talking directly to your kids is genuinely what you want to do, but if it didn't work out that way for a day because of how busy you are or any number of things that happen in the daily life of a parent, this is at least a way to fall back on something, again, just for that moment.
Unfortunately, I think it came out more like he was asking ChatGPT to do the parenting for him and given how many people are eagerly awaiting to jump on him for any AI transgression. I don't think it's too surprising. He got the reaction he did.
Opus just feels like it likes to hear itself talk. It's really obnoxious. Unfortunately for me, it got so bad that I think that's going to be the reason I ended up cancelling my Claude subscription. I was just expecting so much more, and having to beat my model into submission with a bloated agents.md file just to get it to pipe down isn't my idea of a good time.
@abue_ammar Just to clarify, I assumed you might be open to it because I can see the app itself is developed with AI assistance. I meant no offense and I hope it wasn't taken that way! <3
Just to give you a brutally honest critique: the tone of your post and the follow‑up, and the way that he replied, are two very, very different levels of dignity to me...and it's not in your favor. It's not that I don't think you're making a point about being genuine, especially when you hold such a high position.
But I feel like you could do this without being so emotionally charged. There's almost always a better way of communicating.
@Mads_Hartmann But the charging block is white, and unless you're looking at the charging block, it shouldn't matter? I don't think Apple makes a non-white charging block.
If you picked up a Creator Micro 2 because you missed the Codex Micro 2, you need to make sure that the firmware is updated to the latest as of today, which is 0.6.1, and then you need to follow these instructions to get it working with the ChatGPT app.
I figured it out. Leaving it here for anyone else. I had to use the Input app (https://t.co/NTeIZ9rzQ2) to add a layer, and then set that layer explicitly to "ChatGPT Layer".
@work_louder if I can make a suggestion, maybe consider smoothing this out a little bit for those of us who use ChatGPT. Maybe consider seeing if you can get the @ChatGPT official app to take care of this automatically, or at least suggest it if ChatGPT can detect the macro pad, but there are no valid key map files for it.
I don't really mind because I would consider myself fairly tech savvy. But I can easily see others who might not have the same patience about something they dropped several hundred dollars on. :)
I figured it out. Leaving it here for anyone else. I had to use the Input app (https://t.co/NTeIZ9rzQ2) to add a layer, and then set that layer explicitly to "ChatGPT Layer".
@work_louder if I can make a suggestion, maybe consider smoothing this out a little bit for those of us who use ChatGPT. Maybe consider seeing if you can get the @ChatGPT official app to take care of this automatically, or at least suggest it if ChatGPT can detect the macro pad, but there are no valid key map files for it.
I don't really mind because I would consider myself fairly tech savvy. But I can easily see others who might not have the same patience about something they dropped several hundred dollars on. :)
@work_louder I'm a little confused. I got a Creater Micro 2 (since I missed the Codex Micro drop), and I was under the impression that it was just going to be plug and play. But then I saw some post that said there's some kind of beta firmware that gets it to show up properly in ChatGPT.
Can you help me out here?