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

I don't follow closely all these benchmarks but I would love to have some idea of the status of models for these specific use cases. Average intelligence is close for each mainstream models, but on writing, design, coding, search, there is still some gaps.

Even if it's not benchmark, a vibe test from a trusted professionnal with a close use case to mine would suffice.

Your point about ecosystem is true, I just switched main main provider from OpenAI to Anthropic because they continue to prove they have a good concrete vision about AI


I don't know about life-changing but to me there are two major benefits that get me really interested:

- Augmenting CLI with specific knowledge and processes: I love the ability to work on my files, but I can only call a smart generalist to do the work. With skills if I want, say, a design review, I can write the process, what I'm looking for, and design principles I want to highlight rather than the average of every blog post about UX. I created custom gems/projects before (with PDFs of all my notes), but I couldn't replicate that on CLIs.

- Great way to build your library of prompts and build on it: In my org everyone is experimenting with AI but it's hard to document and share good processes and tools. With this, the copywriters can work on a "tone of voice" skill, the UX writers can extend it with an "Interface microcopy" skill, and I can add both to my "design review" agent.


As a designer I'm following the space quite closely. In the not-so-distant future, we might skip most visual software and work directly with code. Especially if we have cool transitional solutions like this

I'm still wondering about how this approach would work when conceptualizing flows, systems, or more complex interfaces


I’m in a similar boat. I do like the idea of code as material and this could be a version of that.

I find conceptual work is sending me back to paper and lower fidelity even more these days with AI. The fundamentals thrive with more freeform tools.


There's also a regulatory component. No way hidden ads will be allowed in major markets like the EU.

I could see a sponsored section in the middle of the reply where the LLM just tells of these vendors align with what the user is looking for


Yes definitely most likely thing to happen.


I'm not sure about that. Reports have shown that models from China or Mistral can achieve 80% or more of OpenAI's performance for a fraction of the cost.

If you're tucked in right behind the absolute frontier models, the economics change completely


I'm currently trying both Raycast windows (beta) and Flow Launcher. I've never really used this kind of launcher before (just the highly frustrating Windows main search feature).

- Raycast has a nice UI that can expand to work well with extensions

- Flow is faster to use. With Raycast you often need to enter an extension to finish your action. To launch a scrip on Flow I just type "r [shortcut] -> enter" while Raycast is "quicklinks -> enter -> [shortcut] -> enter. [edit, with minimal setup using aliases, you can have similar speed. See __jonas comment below]

- Performance-wise, Raycast was often eating my RAM, but a dev mentioned it's expected in the beta, they'll fix it for the launch. Otherwise, both feel snappy

- Both seem to have enough community support and extensions

- I never really tried the AI features, I don't know if it's the right place for me to augment my workflow w/ it

Curious about the experience of others with these tools or similar ones


> Flow is faster to use. With Raycast you often need to enter an extension to finish your action. To launch a scrip on Flow I just type "r [shortcut] -> enter" while Raycast is "quicklinks -> enter -> [shortcut] -> enter

That’s surprising to me, since it’s not how it works in the mac version of Raycast.

There you just type the extension name to trigger it, which you can also set an alias for, so I have it set so that if I type “c” then press space I see my list of vscode projects which I can search. “f” goes into file search (I think that’s the default even)


Good point, turns out I was using aliases wrong! It works with spacebar and makes it as easy to use as flow once you set it up.

F is not set as an alias by default tho


Have you tried Everything by Voidtools? I’m curious how these tools compare with that. I like how fast and simple it is.

https://www.voidtools.com/


Nope, sorry. My main use case is app launch, finding settings, and some scripts. I don't use file search that much.

Small point for Flow here again, because you just have to use the prefix doc: to search through your files, whereas on Raycast, you need to set up an alias and enter the extension. Both have file preview


My mom recently praised the brave AI summary of a webpage so who knows, the usage might be higher than we think.


Loads of people are Google's AI Summaries; it's the first result, so, hard to miss.


I used to hate Twitter when it first launched because I thought short form text was stupid, now I see everything will become summaries with AI and nobody will ever read anything meaningful.


It could be something of an historical return to form; a small class of properly educated people and then the wider, semi-literate masses.


I'm "properly educated" by most definitions, 95% of web pages are garbage and a summary is fine. Also I imagine you frequently read summaries of books and movies and many other things before deciding to read or watch the entire work.


>95% of web pages are garbage and a summary is fine.

Mmm, summarized garbage.

>Also I imagine you frequently read summaries of books

This isn't what LLM summaries are being used for however. Also, I don't really do this unless you consider a movie trailer to be a summary. I certainly don't do this with books, again, unless you think any kind of commentary or review counts as a summary. I certainly would not use an LLM summary for a book or movie recommendation.


Communicating in pictographs


That should be a next step. It takes too much time to read summary. So the result should be a summary picture! Text based image generation is quite good now. How would you call this chatgpt feature?


Gotta love them emojis


If someone wanted to do this for whatever reason, there's actually a language that can be written exclusively in emojis. It's called toki pona, and while emojis aren't the standard writing system, there have been several proposals. It works well since toki pona has a very small syntax (only around ~150 words iirc)


There is plenty of text for which a good summary will have a far higher ratio of meaning to words than the original.


Did you write a comment like this last time a recipe clipper got posted here?


I want to love Mistral but never really used their products. What are they good enough for / great at?


I finally made DDG my main search engine because I use Perplexity for anything substantial. It works quite well


Not comfortable at all, but when you have serious medical issues this is the least of your problems.

For work, I assume everything is used for corporate espionage, so it depends on the sensitivity of the data. If my employers / clients authorize a tool it becomes their problem.


No, I'm not talking about serious medical issues. I mean uploading your blood test results and asking an AI, "Give me specific nutrition, lifestyle, and supplementation recommendations."

For work, as an employee, sure, it's easy to say the company approved ChatGPT or Gemini, so you can go ahead and upload, for example, usage data to get a retention analysis. But what if you're the employer?


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

Search: