MattDaEskimo

18 ms·

121.
by MattDaEskimo 3y ago
Disclaimer: An opinion, beware. There is going to be a lot of spam, but the power of AI will provide unique curated content to each user. So websites, mostly gone. Content, mostly on social media and then provided to AI through API services
122.
by MattDaEskimo 3y ago
Agreed, to an extent. These platforms have powerful, refined algorithms to handle the swarm of submissions. So far there's a single metric of initiated conversations for rankings in the store. The Plugin Store had no metrics, and was r
123.
by MattDaEskimo 3y ago
Full permissions for a user is blatant negligence. For anyone who's never used Firebase before this is as simple as a single piece of logic that appears basically as: if authUserID is UserDirectoryID That simple.
124.
by MattDaEskimo 3y ago
This is still apples to oranges, but I'll bite. A summarized book can still entice a potential reader to purchase it. It's a form of advertising. Chewing up content and spinning it without any citations does not provide the origin
125.
by MattDaEskimo 3y ago
This is apples to oranges. You don't need permission as a human, to read content if it's freely available. It was explicitly made for us to consume with an expectation of returning value. Mainly advertising. Laws are (in theory) p
126.
by MattDaEskimo 3y ago
Your comparison to the Amazon mp3s doesn't match at all the current, unprecedented situation. The obvious issue is that copyrighted material was used without permission, and an opt-out feature was introduced without any way to remove a
127.
by MattDaEskimo 3y ago
I am speaking of the OPs implementation of RAG, not in general. The retrieval part can be expensive if an LLM is used to confirm that it is sufficient, and if it's decided to continue searching if it's not. OpenAIs retrieval is a
128.
by MattDaEskimo 3y ago
RAG conceptually is the solution for hallucinations. I'm being critical of the implementations used to achieve it and the lack of awareness for hallucinations. It's definitely not ready.... Yet.
129.
by MattDaEskimo 3y ago
The described technique of RAG is not only expensive, but also prone to hallucinations. I would have liked more discussion on hallucinations, which is the ultimate pitfall of LLMs. This is critical for discovery-based public-facing chatbots
130.
by MattDaEskimo 3y ago
I'd like to believe that lives are way too nuanced to properly train and evaluate a transformer this way. Also, I would wager that upon release, suddenly everybody says they eat healthy and exercise daily.
131.
by MattDaEskimo 3y ago
I am not a mechanic - only an enthusiast, but I know that a steering knuckle should never break the way it did in the photo unless there was some flaw in production.
132.
by MattDaEskimo 3y ago
The more I use and see GPT bots in the wild as public-facing chatbots, the less I see them actually being useful. What's the solution here? An intermediate classifier to catch irrelevant commands? Seems wasteful. It's almost like
133.
by MattDaEskimo 3y ago
If they are using GPT (They most likely are) you can report them as this goes against OpenAI's terms of service.
134.
by MattDaEskimo 3y ago
Not a surprise considering the recent events. Hope to see more information released in the meantime.
135.
by MattDaEskimo 3y ago
I don't think the post you quoted is implying that they should've closed their eyes. It makes much more sense to try a different browser first and see if the problem persists. Instead of test versions and extensions.
136.
by MattDaEskimo 3y ago
I'm thinking the same thing. It could be that Grammerly injects it's own loading spinner with the same filename into the HTML. I wish they tried to simply rename the file instead of remove it.
137.
by MattDaEskimo 3y ago
The new GPT-4 model has a context length of 120k. For consumers this equates to slightly more than $1/message input-only. If ChatGPT is using this model then it's more reasonable to assume that they are bleeding money and need to