VHRanger

6 ms·

31.
by VHRanger 11mo ago
What ChatGPT / Claude features do you use that we don't support? We have an MCP server I can give you access to for search immediately. Down the line a search API and chat completions API to our assistant in the pipeline.
32.
by VHRanger 11mo ago
All of this will depend on the settings on the model (reasoning effort, temperature, top_k,etc) as well. Which is why you should have benchmarks that are a bit broader generally (>10 questions for a personal setup) otherwise you overfit
33.
by VHRanger 11mo ago
That's a fair comment. Just to give my point of view: I'm head of ML here, but I'm choosing to work here for the impact I believe I can have. I could work somewhere else. As for the net positive effect, the point of my essay
34.
by VHRanger 11mo ago
We're not running on openrouter, that would break the privacy policy. We get specific deals with providers and use different ones for production models. We do train smaller scale stuff like query classification models (not trained on u
35.
by VHRanger 11mo ago
I don't think we share them across accounts, no, but we do use your personal kagi search config in assistant searches.
36.
by VHRanger 11mo ago
Quick assistant is a managed experience, so we can add features to it in a controlled way we can't for all the models we otherwise support at once. For now Quick assistant has a "fast path" answer for simple queries. We can&#
37.
by VHRanger 11mo ago
I can't really do anything with the recommendation you're making. The recommendation you made worked from your personal preference as an axiom. The fact is that the APIs in search cost vastly more than the LLMs used in quick answe
38.
by VHRanger 11mo ago
Our stuff is profitable. Actually if you use LLMs sized responsibility to the task it's cheaper than a lot of APIs for the final product. The expensive LLMs are expensive, but the cheap ones are cheaper than other infrastructure in som
39.
by VHRanger 11mo ago
Ah yes we have some benchmarks on this sort of misguided prompt trap, so it should perform well on this
40.
by VHRanger 11mo ago
We're explicitly conscious of the bullshit problem in AI and we try to focus on only building tools we find useful. See position statement on the matter yesterday: https://blog.kagi.com/llms
41.
by VHRanger 11mo ago
We're building tools that we find useful, and we hope others find it too. See notes on our view of LLMs and their flaws: https://blog.kagi.com/llms
42.
by VHRanger 11mo ago
(Kagi staff here) Generally we do particularly better on product research queries [1] than other categories, because most poor review sites are full of trackers and other stuff we downrank. However there aren't public benchmarks for us
43.
by VHRanger 11mo ago
I'd say the most useful part for me is appending ? / !quick / !research directly from the browser search bar to a query
44.
by VHRanger 11mo ago
You'll want de-censored models like cydonia for that -- can be found on openrouter, or through something like msty
45.
by VHRanger 11mo ago
It's not -- this was posted literally yesterday as a position statement on the matter (see early paragraphs in OP): https://blog.kagi.com/llms Kagi is treating LLMs as potentially useful tools to be used with their def
46.
by VHRanger 11mo ago
In the original riddle the boy's dad dies in the accident before the boy arrives at the hospital.
47.
by VHRanger 11mo ago
Keep feedback loops short and critical output to be verified by humans short. So this means that outputted answers in something like Kagi Assistant shouldn't be like those "Deep Research" report products where humans inevitab
48.
by VHRanger 11mo ago
Ah, might have been the temperature settings on the API I used. It seems to pass it on high reasoning and temperature=1.0 but it failed when I was writing the comment with different settings (copy pasting the string into an open command lin
49.
by VHRanger 11mo ago
LLMs can be useful as a tool, you shouldn't "delegate" work mindlessly to them. I don't "delegate" work to my nail gun or dishwasher, I work with the tool to achieve better productivity than without. When viewe
50.
by VHRanger 11mo ago
Right, this is why I (author here) close the article mentioning that product design needs to keep the humans in the loop for these models to be useful. If the product is designed assuming humans will turn their brain off while using it, the
51.
by VHRanger 11mo ago
Hi, author here! The hyperactivation traps (formal name: misguided attention puzzles) are mostly used as a rhetorical device in my post to show how LLMs come up to a verbal response by a different process than humans in an entertaining mann
52.
by VHRanger 11mo ago
LLMs can't generate knowledge - they don't have a concept of truth. They're very useful for research tasks, however, especially when the application is built to enforce citation behavior
53.
by VHRanger 11mo ago
It's a classic riddle from the late 20th century when surgeons were rarely female.
54.
by VHRanger 11mo ago
Hi, author here! One issue with private LLM tests (including gotcha questions) is that they take time to design and once public, they become irrelevant. So I'm wary of sharing too many in a public blog. The surgeon dog was well known i
55.
by VHRanger 11mo ago
Kagi assistant is effectively a superset of other LLM chat apps. Has access to kagi search which is a also a superset of search backends for the assistant
56.
by VHRanger 11mo ago
Thanks for this response by the way, It's useful knowledge.
57.
by VHRanger 11mo ago
Do you plan to expose daft as a backend in ibis? That would be the best way to smoothly test it out and transition workloads from other engines for codebases in my teams.
58.
by VHRanger 11mo ago
Yes we were aware of that when building it. Image slop is directly detectable by a model, but web page slop is necessarily a multi-signal system (page format, who posted it, link structure, content,...) So having AI images in a webpage is j
59.
by VHRanger 11mo ago
Correct, hence slopstop leveraging other signals than just the content
60.
by VHRanger 11mo ago
There's already methods that attempt that. It works for images because diffusion models leave artifacts, but doesn't work so well for text. Text is an incredibly information dense data format. The diffusion artifacts kind of sneak

More ›