If ChatGPT gets an artifact window equivalent that can actually run the html code it writes, I might try it again but I’m so used to creating simple educational html web apps and running them on my phone for my kids natively in Claude.
Great points about the examples used. Strikes me that Anthropic thinks more about user experience and use cases while OpenAI cares more about the technical aspects of the api, but even for developers, they need inspiration and affordances for whats possible.
It’s out in the UK overnight. Been using it for a project today. Pretty good. My use case was for coming up with options for sales territory management tools for a salesforce setup. Pretty good results.
I think the 12 days thing is really about recapturing the public interest in openai and the perception that ooenai is leading the pack. It's about dominating the news cycle, not about supporting developers, and I can see why that would be annoying for you guys.
OpenAI needs to front GPT o1 with a really dumb but very lightweight model. So while it is thinking, it'll say stuff like "Huh." "Yeah, that's an interesting question." "Just to be clear, you want to know the number of 'r's in the word 'strawberry', right?" "Let me think a moment..." "Hmmm - I think..." And maybe it can scan the internal thought process to get some intermediate commentary like "Ah - I see, so to count the number of 'r's in Strawberry, I can just write down each letter and increment a count if that letter was an 'r'. " " 'S' - not an 'r', so the count is zero..." Finally: "Ok, I've done it - there are three 'r's in Strawberry".
i hate the new model... so cold, no memory of how i've trained it already... and they want $200/mo for a cold robot? was elon the project lead? crap...
I think the increased ability to solve math problems is more important on a deep level than you guys are thinking about. First of all, as developers you should be keenly aware that the quality of the process is different than the quality of the answer. There are analogies in math to brute forcing a problem versus finding a closed form solution. And when you talk about AI that can be used to build AI, coming up with slightly more performant probabilistic computation used over a hundred billion parameters or tokens, you could see how a little bit better math might go a long way.
They will have terms that allow them to switch users off or move them to the API pricing on a discretionary basis. Like the phone companies do for data plans or stream companies have in their terms and conditions. They they have half decent lawyers it’ll be in there.
I need o1 without rate limits. The thing is brilliant, but I'm always concerned about hitting my limit. o1 combined with my personal GPTs would be perfect.
![Noob To Max With DRAGON REWORK In Blox Fruits [FULL MOVIE]](http://i.ytimg.com/vi/LnBMOinoOvA/mqdefault.jpg)
If ChatGPT gets an artifact window equivalent that can actually run the html code it writes, I might try it again but I’m so used to creating simple educational html web apps and running them on my phone for my kids natively in Claude.
Great points about the examples used. Strikes me that Anthropic thinks more about user experience and use cases while OpenAI cares more about the technical aspects of the api, but even for developers, they need inspiration and affordances for whats possible.
It’s out in the UK overnight. Been using it for a project today. Pretty good. My use case was for coming up with options for sales territory management tools for a salesforce setup. Pretty good results.
I think the 12 days thing is really about recapturing the public interest in openai and the perception that ooenai is leading the pack. It's about dominating the news cycle, not about supporting developers, and I can see why that would be annoying for you guys.
you mean the 20% bigger flavor explosion wasn't a real thing? 😕
OpenAI needs to front GPT o1 with a really dumb but very lightweight model.
So while it is thinking, it'll say stuff like "Huh." "Yeah, that's an interesting question." "Just to be clear, you want to know the number of 'r's in the word 'strawberry', right?" "Let me think a moment..." "Hmmm - I think..." And maybe it can scan the internal thought process to get some intermediate commentary like "Ah - I see, so to count the number of 'r's in Strawberry, I can just write down each letter and increment a count if that letter was an 'r'. " " 'S' - not an 'r', so the count is zero..."
Finally: "Ok, I've done it - there are three 'r's in Strawberry".
I think they may be planning exactly that. Spin up the question in 4o then allow you to send to o1 if needed.
Hey guys, I subscribed to Simtheory Pro plan but now can't past the "Select a plan to continue" page. Can't reset password. Any help!
Please contact support@simtheory.ai
It is So Amazing. 😅😅😅
i hate the new model... so cold, no memory of how i've trained it already...
and they want $200/mo for a cold robot? was elon the project lead?
crap...
I think the increased ability to solve math problems is more important on a deep level than you guys are thinking about. First of all, as developers you should be keenly aware that the quality of the process is different than the quality of the answer. There are analogies in math to brute forcing a problem versus finding a closed form solution. And when you talk about AI that can be used to build AI, coming up with slightly more performant probabilistic computation used over a hundred billion parameters or tokens, you could see how a little bit better math might go a long way.
I have it with a common subscription, and I'm in Canada. Yet to use it as I just checked it based off this video title.
Guys, how does OpenAI keep from people from exploiting the “unlimited” access feature of the pro tier? This confuses me
I guess it's like a gym membership, they don't expect everyone to turn up everyday...
@@thisdayinaiexactly, most people only go to chatgpt when they need rather than hanging around there all day long
They will have terms that allow them to switch users off or move them to the API pricing on a discretionary basis. Like the phone companies do for data plans or stream companies have in their terms and conditions. They they have half decent lawyers it’ll be in there.
I left Claude and gone to API. I think it will be cheaper and on demand without limits. We will see
There not competing with perplexity. They see themselves as competing with Google search.
I would pay up to 100 per month for unlimited AVM with all features like in the launch stream. Not interested in anything else.
I was totally unimpressed. I couldn't even watch the whole thing.
I need o1 without rate limits. The thing is brilliant, but I'm always concerned about hitting my limit. o1 combined with my personal GPTs would be perfect.