MemoryCache: Augmenting local AI with browser data • Hacker News Board

Re: MemoryCache: Augmenting local AI with browser data

#71

Post by thepra » Tue, Dec 12, 2023, 9:42 PM UTC

conradev wrote: post #63

Jayakumark wrote: post #2

Was just talking about this on reddit like two days ago Instead of data going to models, we need models come to our data which is stored locally and stay locally. While there are many OSS for Loading personal data, they dont do images or videos. In the future everyone may get their own Model but for now tech is there but product/OSS is missing for everyone to get their own QLORA or RAG or Summarizer. Not just message…

> Instead of data going to models, we need models come to our data which is stored locally and stay locally. We are building this over at https://software.inc ! We collect data about you (from your computer and the internet) into a local database and then teach models how to use it. The models can either be local or cloud-based, and we can route requests based on the sensitivity of the data or the capabilities needed…

As fare as I can see it's just a MacOS image, nothing is happening

Re: MemoryCache: Augmenting local AI with browser data

#72

Post by voakbasda » Tue, Dec 12, 2023, 9:42 PM UTC

gardenhedge wrote: post #69

conradev wrote: post #63

Earlier quoted context omitted.

> Instead of data going to models, we need models come to our data which is stored locally and stay locally. We are building this over at https://software.inc ! We collect data about you (from your computer and the internet) into a local database and then teach models how to use it. The models can either be local or cloud-based, and we can route requests based on the sensitivity of the data or the capabilities needed…

Wow, nice domain. I'd work there for the name alone haha.

Am I cynical thinking the opposite? I can’t imagine they got that domain for a song. Spending a pile of cash on vanity such as that is a real turn off for me; it signals more flash than bang. Am I wrong to think this?

Re: MemoryCache: Augmenting local AI with browser data

#73

Post by herval » Tue, Dec 12, 2023, 9:43 PM UTC

conradev wrote: post #63

Jayakumark wrote: post #2

Was just talking about this on reddit like two days ago Instead of data going to models, we need models come to our data which is stored locally and stay locally. While there are many OSS for Loading personal data, they dont do images or videos. In the future everyone may get their own Model but for now tech is there but product/OSS is missing for everyone to get their own QLORA or RAG or Summarizer. Not just message…

> Instead of data going to models, we need models come to our data which is stored locally and stay locally. We are building this over at https://software.inc ! We collect data about you (from your computer and the internet) into a local database and then teach models how to use it. The models can either be local or cloud-based, and we can route requests based on the sensitivity of the data or the capabilities needed…

site's pretty funny, but would likely be more useful with more information and less clicking-around-nostalgia 8-)

Re: MemoryCache: Augmenting local AI with browser data

#74

Post by tesdinger » Tue, Dec 12, 2023, 9:46 PM UTC

I wish they would fix basic features such as downloading pictures on Firefox for Android. Often long pressing the image on your screen opens a context menu that does not allow download, only following the link associated with the image.

Re: MemoryCache: Augmenting local AI with browser data

#75

Post by huy77 » Tue, Dec 12, 2023, 9:46 PM UTC

so this is how growth hacking look like, building a landing page for a imaginary product to test market-fit idea?

Re: MemoryCache: Augmenting local AI with browser data

#76

Post by smith7018 » Tue, Dec 12, 2023, 9:47 PM UTC

herval wrote: post #73

conradev wrote: post #63

Earlier quoted context omitted.

> Instead of data going to models, we need models come to our data which is stored locally and stay locally. We are building this over at https://software.inc ! We collect data about you (from your computer and the internet) into a local database and then teach models how to use it. The models can either be local or cloud-based, and we can route requests based on the sensitivity of the data or the capabilities needed…

site's pretty funny, but would likely be more useful with more information and less clicking-around-nostalgia 8-)

That's because the company is more or less in stealth/investigatory mode. It's the same team that built Workflow which was acquired by Apple and then turned into Shortcuts.

Re: MemoryCache: Augmenting local AI with browser data

#77

Post by rhn_mk1 » Tue, Dec 12, 2023, 10:00 PM UTC

linsomniac wrote: post #6

I would sure love a way to "chat" with my browsing history and page content. Is there any way to automatically save off pages that I've visited for later processing? I looked a decade or more ago and didn't really find a good solution.

Check out Recoll with Recoll-WE https://addons.mozilla.org/en-US/firefox/addon/recoll-we/

Re: MemoryCache: Augmenting local AI with browser data

#78

Post by ChrisArchitect » Tue, Dec 12, 2023, 10:14 PM UTC

Could barely get a sense of what any of this meant from the shared link.

Went back a bit further/to the official site:

> MemoryCache is an experimental development project to turn a local desktop environment into an on-device AI agent.

Okayy...

And this from November

Introducing Memory Cache

https://memorycache.ai/developer-blog/2023/11/06/introducing...

Re: MemoryCache: Augmenting local AI with browser data

#79

Post by atomicUpdate » Tue, Dec 12, 2023, 10:31 PM UTC

CollinEMac wrote: post #10

I'm confused by the example they gave. > What is the meaning of a life well-lived? Is the response to this based on browser data? Based on the description I was expecting queries more like: > What was the name of that pdf I downloaded yesterday? > What are my top 3 most visited sites? > What type of content do I generally interact with?

One thing you'll see from a lot of these LLM examples and demos are intentionally subjective queries, so they can't be judged on pass/fail criteria.

For example, you'll see things like "where should I visit in Japan?" or "how should I plan a bachelor party?", because they are a huge variety of answers that are all "correct", regardless of how much you disagree with them. There is also a huge number of examples from them to draw from, especially compared to something as specific as your browsing history.

Re: MemoryCache: Augmenting local AI with browser data

#80

Post by pradn » Tue, Dec 12, 2023, 10:42 PM UTC

Jayakumark wrote: post #2

Was just talking about this on reddit like two days ago Instead of data going to models, we need models come to our data which is stored locally and stay locally. While there are many OSS for Loading personal data, they dont do images or videos. In the future everyone may get their own Model but for now tech is there but product/OSS is missing for everyone to get their own QLORA or RAG or Summarizer. Not just message…

Yes, should have local models in addition to remote models. Remote ones are always going to be more capable, and we shouldn't throw that away. Augmentation is orthogonal - you can augment either of these with your own data.