Maybe I need to start looking at SMS/MMS proper, but that's putting a lot of information in the clear and giving up a lot of features.

I used to think that a small payment could accomplish the same thing, but X selling blue check marks proved that doesn’t help much. Well, at least it’s a much weaker signal than the previous curated version.

The challenge is any barrier to entry high enough to discourage motivated spammers is also high enough to discourage casual users. That disrupts the network effects you’ve traditionally needed to bootstrap a social website.

If I was trying to get a new social site off the ground right now, I would try:

1) secure a good brand from the pre-AI era. Twitter, Digg, Friendster, MySpace. Something that motivates a first look.

2) Require third party identity verification on sign up, configured so the social site is never the custodian of PII, though require enough demographics to support high-value advertising later. Verification is free to the user, ideally provide multiple verification options- one US and one EU at minimum.

3) Target a few core communities and invest. Find the people who moderate historically great subreddits, were active in twitter communities during the good years, etc. get them in your platform. Maybe even pay them.

That should be enough to tell you if it’s going to work or not.

Of course, they could also be unrealized potential just waiting for someone to believe in them and mentor them to greatness. If you drop them too quickly you miss out on wonderful, loyal, capable people.

The trick is being able to tell the difference quickly, make the best choice you can, and move on.

> The polar journey, like other zeppelin flights, was largely financed by stamp collectors; Graf Zeppelin carried approximately 50,000 letters sent by philatelists, and made a water-landing to exchange mail with the Soviet icebreaker Malygin, which itself carried a large quantity of mail sent by stamp collectors.

> Prioritize recall over precision.

Have you tried stemming your regex? That would help you catch messages where a different form of your word appeared. For example instead of “story” you look for “stor” which catches “stories” as well.

Then you might think, could we do an even better job by figuring out the general semantic intent of the query and history? Let’s project them into a semantic vector space! That’s an embedding.

Then you want to query that, which means you need a vector database. So now we can take the query, embed it, query the vector DB with that embedding and retrieve the N closest history documents. You can use that to augment the generation of the response to your prompt.

This is RAG.

Anyway, interesting to see different degrees of sophistication here. Certainly a handful of naive regex are very snappy.

There’s probably a hybrid approach where you use sophisticated NLP and embedding techniques to robustly define topics, then train a regex to approximate that well.

I’ve been using RapidMLX for this. The integrated speed tests matter because the quality of the backend is a moving target and the quantization / MLX format conversion also matter. It’s not enough to say “oh use this model family with X parameters” you have to add the architecture specific quantization too.

https://github.com/raullenchai/Rapid-MLX

https://youtu.be/f3u57jkwBFE?si=FJuxZfmc-i7EkPlx

“Built on Python, Salt is an event-driven automation tool and framework to deploy, configure, and manage complex IT systems. Use Salt to automate common infrastructure administration tasks and ensure that all the components of your infrastructure are operating in a consistent desired state.”

https://docs.saltproject.io/en/latest/topics/about_salt_proj...

This is a way for web services to make your computer complete large amounts of compute at their behest. Tokens have value. There will be incentive for bad actors to use your local LLM for their own purposes, much like hostile crypto mining payloads.

This is an obvious target for prompt injection attacks and other malicious remote code execution. In many ways, model prompts ARE programs. The browser / local device would need to provide an LLM with the same sandbox guarantees as the rest of the browser. Can they be trusted to do that? Does anyone understand this well enough to do that with confidence?

I’m a big fan of local models, but I would be very cautious about letting random websites call the model I’m hosting on my local machine with open source software.

It’s not enough for a chat-first research agent, but it’s definitely enough to unlock features that rely on natural language understanding. Seems like a small thing compared to Claude/ChatGPT and the general hype, but still magic in its own context.