I don’t mess with code autocomplete, Cursor, agents or any of that stuff. I’ve got subscriptions to 2 platforms that give me access to a bunch of different models and I just ask whatever model I need directly, copy/paste the context it needs. On that note, AI search engines like Perplexity genuinely bring zero value to my workflow. I’d rather do the searching myself and feed it the relevant context, feels like it misleads me more often than it helps. I actually have a Perplexity sub (got it free) and haven’t touched their web search in like 4 months.

I’ve thought about the environmental impact and taken steps to minimize my usage. That’s actually one reason I avoid Cursor, agents, and AI web search - feels super wasteful and I’m not convinced it’s sustainable long-term. I guess I just like being in control, you know? I also try using smaller open source models when I can, even if they’re not as powerful.

My go-to models right now for daily use (easiest to hardest tasks): Llama 4 Scout -> DeepSeek v3.1 -> DeepSeek v3.1 (thinking) -> Gemini 2.5 Pro / Claude 4 Sonnet (thinking) -> GPT 5 (thinking). Sometimes I’ll throw in other models like Gemini 2.5 Flash but mostly stick to these.

By the way I would recommend trying out t3.chat ( that’s one of the platforms that I use). Cost 8 USD / month and is made by Theo pretty happy with it for the price. The UI is honestly its strongest point.

For how I actually use AI, I wrote a more detailed answer in another thread about AI usage. Have a read

A bit of a tangent but I wonder if this is becoming more common these days. The “shut-in” phenomena, japanese call them hikikomori. News headlines say it is but i wonder how well they represent reality.

There is a fifth way of using AI: Ask AI to hint the problems of your text or suggest rules to look up so you can can “solve” it yourself.

The problem of over-reliance on AI isn’t anything new - we’ve always learned by struggling through problems ourselves. It’s like playing a puzzle game - if you just go look up the solutions instead of trying things yourself, not only do you lose the point of playing the game by reducing it to a series of bothersome tasks that just need to get done to get something at the end, but eventually you find yourself out of depth as you didn’t develop the proper understanding of the puzzles.

Because of this, I’ve been gravitating more towards the 4th and 5th ways of using AI for things that matter to me, things I need or want to understand deeply.

I try not to rush through things, to enjoy the process and instead of just asking for an answer to a question, I’m starting to ask it: ** How can I find the answer myself? What materials would an experienced person in this field look up in order to solve this problem?** And similar variations of these types of questions. The main idea is: I instruct it not to give a solution or code right away but instead to explore the problem together with me and teach me how to fish instead of giving me the fish. If I give him some part of the documentation and he gets an insight, I ask: How did this part of the docs help you get to the conclusion of that? How did you know what to look for ? And so on. Basically I assume the AI is a more experienced person next to me and we’re trying to pair program. He doesn’t know the solution from the back of his mind but he can easily “find it” and we’re walking through it together.

This shift happened because AI kept missing the mark on my questions - partly because I work with relatively niche tools, partly because when you’re learning, you don’t know what context is even relevant to give it and if you give irrelevant context usually you end up misleading it.

And it’s actually surprisingly fun and enjoyable to work on my problems now. There’s this shift of not seeing the problems as something to be solved but as something that needs to be understood, a game that needs to be played if you will. Obviously it takes longer as the article pointed out learning takes time.

I do some dev and linux stuff and I still found it a bit hard to justify the cost. Like it’s not a bad deal, however i did find myself going through days not really using it properly and just doing simple searches like “product x documentation” or just “site:” searches that i can do with simply DDG or Ecosia (Google). I think it’s cuz a lot of the internet is converging into larger sites rather than being scattered all over and i do find myself relying a lot more on LLM’s. If you actually test out their search results you will get something like 97% there if you do a Google search and a Bing search to follow up or just use searx.

What has helped me completely stop using them is finding sites with valuable information for my use case and creating a manual “lens” something similar to what Kagi offers. It’s really simple and works better than theirs. I have this pasted into my obsidian note for linux (example):

site:archlinux.org OR site:endeavouros.com OR site:reddit.com/r/arch OR site:reddit.com/r/archlinux OR site:reddit.com/r/endeavouros OR site:manjaro.org OR site:reddit.com/r/hyprland OR site:hypr.land

Everytime i search for anything arch linux related so i paste this in and enter the search phrase at the start.

I use t3.chat for my LLM needs (you get way more for the price that you pay, including premium models) so I don’t need their LLM’s. I havent found LLM web search to work well for my use case so don’t need that.

I use DDG as my daily driver and then switch to Ecosia (google wrapper) if i cant find stuff. DDG is a nice search engine overall, fills in that “gap” left by Kagi.

I use zen browser and have created search engine keywords like @ecosia, @ddg, @wolfram so i can CTRL + T and “@ecosia search term” easily.

Overall, I feel like they’re not offering anything groundbreaking but their whole package is nice, i see the appeal especially if you browse on mobile. Whether you find them useful or not will depend on your use case. For those 10$ + tax I personally don’t. The Russia support is disgusting though.

And what do you think Kagi uses? They rely heavily on the same Bing and Google. If you played around with them you will see most of their searches are identical to what you would get if you did a bing search and then follow up with a google search, you’re not getting much more. The fact of the matter is that there are only 2 real choices of search engines. Google and Bing. There’s a massive gap between them and every other independent search index.

Everyone should be aware of the fact that Kagi supports Russia by buying Yandex index API. They excuse themselves with neutrality. Well, DuckDuckGo for some reason took the L and dropped them after the Ukraine invasion, so make your own conclusions here. I think it’s easy to come up with all sorts of justifications for your actions and I cancelled my sub a couple of months back. They do have a decent product, though but it’s also pretty expensive.