Wow I’ve never seen enshittification mentioned by a politician. Glad to hear it’s getting inside the Overton Window.

I agree that with the current state of tools around LLMs, this is very unadvisable. But I think we can develop the right ones.

• We can have tools that can generate the context/info submitters need to understand what has been done, explain the choices they are making, discuss edge cases and so on. This includes taking screenshots as the submitter is using the app, testing period (require X amount of time of the submitter actually using their feature and smoothening out the experience)

• We can have tools at the repo level that can scan and analyze the effect. It can also isolate the different submitted features in order to allow others to toggle them or modify them if they’re not to their liking. Similarly, you can have lots of LLMs impersonate typical users and try the modifications to make sure they work. Putting humans in the loop at different appropriate times.

People are submitting LLM generated code they don’t understand right now. How do we protect repos? How do we welcome these contributions while lowering risk? I think with the right engineering effort, this can be done.

Unless you have a balanced diet that anticipates your workouts and gives you the proper amount of sodium, potassium and magnesium. Sports drinks are just selling you those at a big premium. Stick with water. Eat a banana.

Isn’t it the opposite then? Since your windows will have vertical scrolls, it makes sense to tile them horizontally in order to maximize vertical space for each window, imo.

Thanks! I had actually gotten confused by the Create Post interface and accidentally did not post the URL to the blog post heh. I fixed it now

There are lots of people who could use them. Schools, libraries, poor people.

What do you mean? I follow a lot of hashtags on Mastodon. Won’t I be seeing a lot of Threads content if I’m on a server federated with them without explicitly opting into that?

It was just an example. The same can happen at the Mastodon-level instead of the Fediverse-level. Since there is some desired interop (e.g. between Mastodon and Lemmy), services do influence each other in their feature set.

I’m not sure what you mean by “a lot of what people are worried about Threads doing has already been done by Mastodon”. Do you mean that the decisions that Mastodon make influence the rest of the Fediverse? If so, let’s make sure we understand the difference here: Threads has a much more hostile disposition. Mastodon seems to have incentives aligned with the rest of the Fediverse services, and probably deserves the benefit of the doubt; Facebook has abused that benefit time and time again.

This is more a question of tolerance. We know Facebook is NOT tolerant of competitors, of the open web, of free software, etc. They cannot survive as a megacorp without a level of assurance and control that they can’t have if they’re “just another fediserver”. They WILL try to wrangle control. They WILL try to eat us all up. Why let the fox in the henhouse when you already know it’s a fox?

That doesn’t actually fix the issue. If Facebook is trying to set itself up like Chrome with the webplatform, or GTalk with XMPP, then they will drive the feature set of ActivityPub, whether you’re federated with them or not.

Hypothetical example:

Want to see this picture/video from someone on Threads? You need Facebook’s proprietary picture format, which has DRM baked in it. Even if you don’t federate, Mastodon, Lemmy, etc now have to take energy away from their work to adopt the proprietary picture format. It depends on the proportion Threads takes on the network and how they can leverage that position to put pressure.

Threads currently has voice notes. Should all ActivityPub services support that? If so, do we adhere to Threads’ standard or not?

I was under the impression that there were resources in that area that the US currently has privileged access to because of their alliances there. So they have a stake in making their allies come out on top.

I’m not sure what you mean. GPTs also allow you to add datasets, and external APIs. Both of which can be used to supply a spiritual pov.

No.

It seems like they have an automatic lab that tested 58 of them and 41 were successfully synthesized. So 70% success

It does more than that, it magnifies, feeds and perpetuates them. It’s not just simple exposition.

What kind of advice were you looking for if not this?

Essentially, you don’t ask them to use their internal knowledge. In fact, you explicitly ask them not to. The technique is generally referred to as Retrieval Augmented Generation. You take the context/user input and you retrieve relevant information from the net/your DB/vector DB/whatever, and you give it to an LLM with how to transform this information (summarize, answer a question, etc).

So you try as much as you can to “ground” the LLM with knowledge that you trust, and to only use this information to perform the task.

So you get a system that can do a really good job at transforming the data you have into the right shape for the task(s) you need to perform, without requiring your LLM to act as a source of information, only a great data massager.

I’ve been using LLMs pretty extensively in a professional capacity and with the proper grounding work it becomes very useful and reliable.

LLMs on their own is not the world changing tech, LLMs+grounding (what is now being called a Cognitive Architecture), that’s the world changing tech. So while LLMs can be vulnerable to bullshitting, there is a lot of work around them that can qualitatively change their performance.

“It has already started to be a problem with the current LLMs that have exhausted most easily reached sources of content on the internet and are now feeding off LLM-generated content, which has resulted in a sharp drop in quality.”

Do you have any sources to back that claim? LLMs are rising in quality, not dropping, afaik.

Or flatpak