The Wednesday Wellness threads are meant to encourage users to ask for and provide advice and motivation to improve their lives. It isn't intended as a 'containment thread' and any content which could go here could instead be posted in its own thread. You could post:
-
Requests for advice and / or encouragement. On basically any topic and for any scale of problem.
-
Updates to let us know how you are doing. This provides valuable feedback on past advice / encouragement and will hopefully make people feel a little more motivated to follow through. If you want to be reminded to post your update, see the post titled 'update reminders', below.
-
Advice. This can be in response to a request for advice or just something that you think could be generally useful for many people here.
-
Encouragement. Probably best directed at specific users, but if you feel like just encouraging people in general I don't think anyone is going to object. I don't think I really need to say this, but just to be clear; encouragement should have a generally positive tone and not shame people (if people feel that shame might be an effective tool for motivating people, please discuss this so we can form a group consensus on how to use it rather than just trying it).
Jump in the discussion.
No email address required.
Notes -
Out of sheer boredom, I decided to cough up £20 for a ChatGPT Plus subscription. I'd heard good (or at least interesting) things about o3 as a model, and wanted to check it out.
It certainly is a trip. o3 is the most busy LLM I've ever interacted with. You can think of it as a highly neurotic coworker, who will, without any prompting, submit a Monte Carlo simulation of weather patterns when you ask it if it's going to rain tomorrow. This can occasionally be handy, and just as frequently, hilarious.
While the Chain of Thought isn't entirely intact (or divulged to you, the user), it's entertaining to watch it scurry about searching things, writing small snippets of code and using its tools. It loves tools.
Google has been so unkind as to remove the version of Gemini (2.0 Flash Experimental) that produced images on AI Studio, in the UK, EU and a few other authoritarian shit holes (that is not an exclusive or). While OAI graciously does out an image or too a day to free users, getting far more mileage out of it doesn't hurt, but is still a curiosity.
I've tried out GPT 4.5, which is a rare mutant of a model, a relic of different times/research artifact more than a daily driver. So far, I've been whelmed, as it didn't do anything particularly impressive even at the task of writing fiction, which it's supposedly great at.
The elephant in the room is, of course, Google. You can use the best model on the planet, Gemini 2.5 Pro, entirely free on AI Studio, or via API. So it goes for the older models, unless removed due to regulatory concerns. If I wasn't curious, there's very little actual reason for me to subscribe to ChatGPT, and that's probably true even for you, the casual user. Even those too normie to have heard about AI Studio have access to the Gemini app, which is finally taking off, and even there, Google hits OAI where it hurts, in the wallet, by making their best models cheap/free.
It's a good time to be an AI aficionado. Don't pay for ChatGPT unless you really want to out of curiosity, there's no other good reason at present.
PS: Anthropic. A meme at this point. Claude does nothing better than the competition, and is practically unusable due to stringent rate limits on the free tier, and suffers from the same even for paying users (going off anecdotal evidence).
As a major, major fan of 3.5, it’s clear Anthropic have missed a step or two in the last six months. And the rate limits, holy cow, even the API randomly drops out frighteningly often. I bounce back and forth a little on my OAI sub, right now I have one because I’ve been doing laying around with the deep research a bit, and voice is handy on car rides or in the kitchen. I almost feel like o1 was better than o3 but needs more testing by me to be sure.
I need to spend more time with Google. Have you used much the feature where you like highlight and edit and interact with passages of your or its old responses? I didn’t quite grasp the mini demo they did so I’m not sure if that’s a genuinely new useful interaction mode or not.
I've occasionally edited my own responses, but never the LLMs. I wonder if that would be helpful for jailbreaks. At some points, when the context window wasn't enough, I'd delete unnecessary responses to clear up space, but with a 1M window now? Never necessary.
The ability to fork chats is clutch, I can tell you. I miss it on every other platform.
Do you usually fork because you're unsatisfied with the response, didn't want to clutter the context, felt you had a natural off-topic you wanted to explore somewhat sandboxed, or something else? I'm trying to imaging the typical use case where it would feel so essential.
Let's say I have a difficult task that needs a lot of comms with Gemini to get it on the same page. Then forks allow for easily throwing that into the context for a variant of that task. Or simply A/B testing.
The first time I made use of forks, it was when I was generating images with Flash, and got it dialed in. I then forked it so as to try alternative prompts and the effects of different details, with the easy ability to jump back and forth.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link