pull down to refresh

What's the Best Consumer Application of AI?
246 sats \ 9 comments \ @Coinsreporter 20h AI
Just read an article from WSJ about the same question. However it didn't make much sense to me as it was mainly about setting a narrative. Anyone who thinks rationally can easily count a few of AI consumer applications that really are aimed at and are able to enhance human efficiency or solving problems.
So, have your say Stackers....
What do you think is the best consumer application of AI?
Home Automation 0.0%
Vehicle Automation 0.0%
NPCs in Gaming0.0%
Virtual Assistants 28.6%
Image/Video Generation 28.6%
Vibe Coding 28.6%
Forecasting 0.0%
Medical Diagnostics 0.0%
Curing Lonliness0.0%
Something else14.3%
7 votes \ 1 day left
write
preview
related posts
100 sats \ 1 reply \ @Car 3h
Music is by far making the most advancements. This next generation of kids will grow up with AI generated artists or who knows maybe become them.
Say hello to transhumanist artists. Damn Bowie was right again.
reply
21 sats \ 0 replies \ @optimism 1h
"It's just a tool though, isn't it"? "No, it's an alien lifeform"
Love this.
reply
20 sats \ 5 replies \ @justin_shocknet 18h
Most of these things are just algorithms
So that means the correct answer is an interface for algorithms
reply
0 sats \ 4 replies \ @Coinsreporter OP 18h
I agree. I'm just more curious to know it from the angle of consumers or to say it in other terms what's the best use case of AI for consumers?
reply
0 sats \ 3 replies \ @justin_shocknet 18h
The unified interface seems to be it, since a chat bot can manipulate APIs for literally anything, it's becomes the chat bot expands access to otherwise niche algos that couldn't stand alone as there own tool
It's the interface for programs the same way search engines are for the web
reply
17 sats \ 2 replies \ @optimism 13h
One of the lessons I learned from making the pylint MCP is that its more efficient to just run pylint and feed the results to an LLM than it is to make it make the call. And I expect that letting an LLM write the query parameters to an API call but then executing the call programmatically is more efficient too.
The trick right now if you don't know how to code is to let the LLM write the code to call the API, and integrate it with calling the LLM and then the API, and then execute that chain.
If you know how to code, just code it. Much faster.
reply
100 sats \ 1 reply \ @justin_shocknet 13h
Yea things like gpt, cursor and Gemini are doing a lot of tricks behind the scenes to make so much functionality available in their bots
I think MCP will eventually enable similarly useful bots to be more distributed and open, but that still has to be built brick by brick
reply
17 sats \ 0 replies \ @optimism 13h
Here's what I think is useful in MCP today for most people: search & apply. Like the old school google one-shot ("I'm feeling lucky" button) but with a cognitive part in the middle to determine the top result by actually reading the link.
So you let <chat agent> query the search engine, then let <search ranking agent> rank the results properly (zhe algo is awfully good at determining relevance, to my surprise and dismay) and then let <chat agent but without memory> answer the original question based on the provided results.
The difference between what current search engines do in their LLM suggestion top box and this is that they don't read the linked content, but the indexed meta tags. Which is why i really like my news summarizer because it gives meaningful summaries rather than SEO-gameable results.
reply on another page
20 sats \ 0 replies \ @optimism 19h
I chose "something else" because it's mostly tooling for me, not so much "assistant", more "assistance". Examples:
  1. Search relevance / ordering
  2. Text summarization
  3. Transposition (i.e. image/video to text, or since I did the research a bit this week, I also like it for audio transpose)
reply