Hacker Newsnew | past | comments | ask | show | jobs | submit | leokennis's commentslogin

Honda seems to love pulling out of things just when they are about to succeed. Both in F1 and apparently now also in EV's.

I admire and applaud Kagi for what they are trying to be.

But on the AI front, the Assistant is simply worse than using for example Gemini or ChatGPT directly. It is slower, it cannot generate images etc.


That's why I don't think anybody is going to pay for the Assistant alone.

On the other hand, it doesn't monetize any of the data you give. Six of one, half dozen of the other, yeah?

Re: image gen... it's a search engine. Why would I need my refrigerator to toast my bread?


The privacy buffer between the LLM providers and the user is a large part of the appeal. Having it hooked up to a search engine outside of the data broker space makes it uniquely attractive.

Given that window dressing, having toasted bread in a kitchen that isn't selling my data is something I want.


Because at least for my own usage of Google the LLM started out as an interactive search with substantially better context filtering that could tune the results to my desired technical level. However I promptly started just having it explain the subject matter to me rather than spending 30+ minutes consulting various docs and forum posts because it makes for an excellent secretary/tutor combo provided you vigilantly watch for misinformation.

So to answer your question, while charts might not be particularly useful for a search engine a tutor certainly benefits from them.


The research assistant should be able to generate images

I see the slowness and now the news that they might remove the assistant for the $10 plans as evidence of how costly it is to run LLMs and by extension how unsustainable it must be for OpenAi, anthropic, Microsoft etc to be offering such performance for free or very low prices. Surely something has to give soon.

It is not worse I would say. It uses neutral system prompt by default, whereas Gemini and ChatGPT will please you too much to mislead you badly. Also the base search is much batter. You can control the search while with Gemini, for example, you can't.

Keep in mind that kagi offers a wide range of models, not just one. I wouldn't want to have multiple subscriptions (for chatgpt, anthropic, gemini etc.)

This is also true if you use something like OpenRouter, and it will almost always cheaper or betterter (excluding Kagi Search).

I love Kagi-I can't imagine going back to any other search engine-but it isn’t competitive when it comes to LLMs. In its defense, that’s largely because others are bleeding money.


In my experience that is true, but I get the answers I need much faster than any kagi search with or without their ai integration from Gemini, ChatGPT or Claude. I was rooting for them, but they just seemed to far behind the leading llms for search. Same with perplexity. Could never figure out why it needs to exist. If I want citations, I’ll just ask one of the LLMs I mentioned to provide them.

Also, there news app is pretty underwhelming.


Every time some site or person tries to make me feel bad for using AdGuard DNS, ad blockers etc. I read an article like this and I feel fine.

I see three options:

1. Show me reasonable ads and I will disable ad blocking

2. Do the crap described in this article and don't complain when I arm myself against it

3. Do a hard paywall and no ads; force me to pay to see your content


It's my device. I decide what I download, execute and display on my device. A website is free to offer me to download an ad and I am free to decline that offer. Demanding me to download anything on my device or even worse execute someone else's programs [JS] and claiming that I have a moral obligation to do so is deeply creepy.

I was about to comment this too.

For someone that explicitly states:

> I don’t listen to pop songs. I prefer music of the 500-year tradition (...)

And who apparently wants to stream music, it is wild he's not subscribed to Apple Music Classical which exactly circumvents all complaints in this article...


I have a 2020 MacBook Air M1 with 16GB RAM - for development work, there is 0 reason to upgrade it. All day battery, silent, small, no lag...


At least in the ChatGPT app, you can set some "personality" traits. Like the style (more or les warm or enthusiastic, use more or less lists and emojis) and the tone.

I have mine set to efficient, using less warmth, less enthusiasm, less headers and lists and less emoji's. Combined with sensible personal instructions (don't placate me, don't flatter me, be professional, tell me if I'm wrong, tell me if you don't know etc.), I see none of the "that's not crazy, that's commitment!" or "here is the no-frills rundown" BS.


> They will agree to make some high margin simple to assemble thing in the US to appease government

I chuckled out loud at the huge-ass-safety-hazard-in-any-manufacturing-environment US flag thumb tacked to the factory wall. It's all wafer thin gold leaf to appease the toddler in command.

https://www.apple.com/newsroom/images/2026/02/apple-accelera...


This seems like a very nice project to learn stuff, including hardware design/implementation.

Are there similar options on the market for "set and forget" people like me?


Love those weather icons. Personality in software design is underrated.


Thanks! I intentionaly made the weather symbols somewhat "childlike" to give it some personality and also make it obvious that it's a custom device, and not some off the shelf gadget. Works well as a conversation starter!


With so many apps introducing either paywalls (requiring either login or circumvention measures) or terrible RSS feeds (with content missing, images missing etc.) I have found it necessary to use a feed reader that you can configure per-feed to open either:

- The feed item (read the XML)

- The site fulltext

- The original site (in case of login required)

For me that app is https://www.lireapp.com/


Is there some way to create this kind of experience without having to change RSS readers? Is there a service that allows you to easily create RSS feeds for websites without them? I'd rather go with a more unix "do one thing and do it well" philosophy for something like this.


There’s rss-bridge which is in the ballpark


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: