Hello everyone, does anyone use third-party platforms like NanoGPT? I used to have a subscription that was enough for AI usage for a month, but then I tried Claude 3.7 Sonnet Thinking and 3.7 Sonnet through API, and it consumes a lot of tokens and money accordingly. It seems expensive and more cost-effective to use it directly or through Perplexity, which has a subscription system where you know for sure that any AI will be sufficient for a month. After all, it’s not always possible to create a new chat and communicate with Claude in a new chat, and I often need to edit different texts or generate content, as well as analyze data, which I like Claude for.
What do you use, and what would you recommend in my case? Is there something to replace NanoGPT that’s also private but uses a subscription system? Or if not subscription-based, how can I solve this problem?
I use NanoGPT and it works really well for me given all that it offers. I don’t use it often, may be a couple of searches a day at most and sometimes not even that. But even then, I have found the pricing to be good and better - atleast for me and my use case and frequency.
What exactly are you trying to accomplish here? Can you clarify? What’s your usage like?
Searching for information on the internet, solving a particular problem, assistance in creating cover letters for job applications (automation), brainstorming in a single chat to reach a solution, learning something new from the world, solving technical problems, finding out how something works, how to configure something, conducting some kind of analysis.
Yeah, NanoGPT works well. If pricing an issue based on frequency of use, then its really upto you. But for what NanoGPT offers, it is much better than what you can find elsewhere.
$19 for 4 days, often for my purposes I don’t create a new chat and the old one comes up and rereads. I like the new Sonnet more than GPT-4o, but it takes a lot of tokens via API, so I’m considering switching to a subscription mode with Perplexity, or directly to Claude.
There are situations where it’s necessary to use it in a single chat to get a relevant result or to get an answer that satisfies me, because the emotional tone of the text or prompt affects what the result will be, and it also depends on the task at hand.
I use it thoughtfully and understand that it adds to my productivity, automates routine tasks, and greatly simplifies life, saves time, plus I engage my critical thinking and make many decisions myself. I don’t think it makes me less intelligent; on the contrary, I develop faster thanks to the emergence of AI.
Or do you think that AI makes people less intelligent and that people need more independence, data analysis, and independent information searching on the internet?
Seems like you’re using it properly and thoughtfully. That’s good.
But how I meant was, while AI does seem to help with use case such as yours - over time, a person’s inherent capacity/ability/skillset of doing things by themselves by thinking for and using their own inherent creative and critical thinking skills will degrade. Human intelligence and the skills around are perishable. And hence, extreme dependence on AI for things won’t lead to anything good long term.
AI also takes away your capacity to evaluate info for what it is - if AI is directly given you the answer. And in today’s world, truth is much needed and reliance and blind belief on AI is not. But the corporate world has different priorities it seems.
But how can I not use AI if I need to use it to be competitive in the IT field, for example? Plus, my future profession assumes that I should use AI. If I don’t use it, I will have a lower salary, I won’t be as valuable to businesses, and I will lag behind other people and develop more slowly.
Additionally, many IT professions are transforming, and some will start disappearing, and people in the world need to adapt to this world. What do you think? Is AI really that bad for humanity and the brain?
By the way, have you read any research about this?
I’m not saying and nowhere did I say you should not use it. All I expressed was my opinion that it must not be heavily relied upon at all times at risk of losing the skills you must have or develop than relinquish control to those areas to an AI for your work.
Of course, develop AI skills along with your own to do the same (as much and wherever you can).
I am a much much lighter user of AI than you are (and I also try to avoid the proprietary models like Claude). My interaction with LLMs has been mostly split between:
Small open source models hosted on my local network (using ollama and openwebui via docker), hopefully larger models in a few years.
If you are a very heavy user, I expect you will run into limits or caps with #2 and possibly #3.
We endorse Privacy by Design. As such, your conversations are private to you and will not be shared with anyone, including model authors, for any purpose, including for research or model training purposes. You conversation data will only be stored to let you access past conversations. You can click on the Delete icon to delete any past conversation at any moment. Link to full privacy policy↩︎
Everywhere they advertise AI and show how people do everything faster with it and become more successful. And I thought, is it possible to remain competitive if you hardly use it, and instead use AI mainly for searching information on the internet.
But I have been using it very actively for several years, and my acquaintances around me also use it constantly, but I haven’t thought about the consequences for me, for my development
It depends on how you’re thinking of and defining competitive here. And for what purpose and area you want to remain and be competitive.
There’s no easy or a simple answer to this question.
Also,
Yeah, I would consider this a problem. This is what I was referring to. I don’t think this is a good thing. This basically means you’re trusting and believing all that AI responds you with and question it very little.
You can continue researching and reading up on it. I already shared links and a couple of research papers. That should give you enough idea on how be better and what not to do.
But if I were you, I’d break this addictive cycle to only use AI for everything.
