crosen99

9 ms·

31.
by crosen99 4y ago
I don’t believe autodiff is finding the difference in that sense. It’s finding derivatives.
32.
by crosen99 4y ago
A small difference between the pattern you describe and the one of the inquiry is where responsibility lies for retrieving and incorporating the augmentation. You describe the pattern where an orchestration layer sits in front of the model,
33.
by crosen99 4y ago
That's a good share. OPs project and video are very interesting to me, although the "Introduction to Checkvist" explanation page serving as all in one demo/documentation/trial was for me a much more efficient way t
34.
by crosen99 4y ago
For sure, I just edited my comment to mention LangChain, which starts to get at that capability. Would be nice to see LangChain integrate with Trustfall similar to how it can integrate with Wolfram Alpha.
35.
by crosen99 4y ago
This is a nifty tool. It’s existence alongside the emerging LLMs reminds me of the two diametrically opposed approaches to harnessing it all: 1. Store the knowledge in a highly structured way and interrogate it with a precise and rigorous
36.
by crosen99 4y ago
Nice list. What helps me: “A preoccupation with the means is a lack of commitment to the end.” In this context, the means is knowledge. And for me, the end is building things. So, I try always to be building something (in my work or avocati
37.
by crosen99 4y ago
> Here's a Hacker News lookalike [1] for Machine Learning. You'll see that the latest post was 11 days ago, and the one before that 29 days ago and the one before 54. It looks like the main feed of DataTau ( https://d
38.
by crosen99 4y ago
That’s good info. Curious if you have any guidance about how big a document chunk should be to get a good match with a natural language query. For context, I’m building a question answering app using ada-02 that pulls in around 100 pages of
39.
by crosen99 4y ago
> Having said all of that, they do have a fine-tuning endpoint and I am guessing if you find the right parameters and give it a lot of properly formatted training data then it will be able to do an okay job. Unfortunately, the fine-tunin
40.
by crosen99 4y ago
You’re saying that if you do something you otherwise don’t want to do because you have the discipline, then you are not doing what you ultimately desired. Then why are you doing it at all? The only thing that directs you to “be disciplined
41.
by crosen99 4y ago
I think motivation vs discipline is a false dichotomy. I’d argue that discipline is simply a form of motivation that emerges when the pain of avoidance can be adequately internalized. IOW, what we call discipline is just a conditioning that
42.
by crosen99 4y ago
Our understanding is so incomplete that we have to posit the presence of dark matter and dark energy - comprising the majority of the universe - to explain the parts we think we do understand. With so much we cannot account for, it seems un
43.
by crosen99 4y ago
Here’s a company doing that: https://www.usemeru.com/
44.
by crosen99 4y ago
I see your blog referencing what you call the “semantic search answers” pattern. I’ve seen this elsewhere as Retrieval Augmented Generation or Data Augmented Generation. A few libraries, like LangChain, have support for this.
45.
by crosen99 4y ago
This is what you said originally: "We only need a single counter example to show that Othello-GPT does not have a systemic understanding of the rules, only a statistical inference of them." I don't believe this statement to
46.
by crosen99 4y ago
I'm not sure why you say these claims are unrelated. The article claims to have found the existence of concepts embedded in the trained parameters of a model. (You called this having a "systemic understanding of the rules."
47.
by crosen99 4y ago
> Just because a move has never been played, doesn't make it illegal, only untried. Well, I think this is nickelpro's point. Namely that the model "does not have a systemic understanding of the rules, only a statistical i
48.
by crosen99 4y ago
It’s a lot of work and it’s messy, but the multiplier on productivity is too large to ignore. At some point all this will get easier, but for now it’s do or die.
49.
by crosen99 4y ago
"Neural Networks and Deep Learning", by Michael Nielsen http://neuralnetworksanddeeplearning.com (full text) The first chapter walks through a neural network that recognizes handwritten digits implemented in a little o
50.
by crosen99 4y ago
In the short run, I’d think it more likely and useful for it to produce sources of factual statements rather than qualify or quantify its own confidence. For one thing, a source is something the user could verify. A confidence value would
51.
by crosen99 4y ago
It’s only a gut feeling, based largely on the predictions and accounts I’ve seen from those with actual first hand knowledge around the reasonable questions that you pose. My framework for thinking about this at a technical level is that sh
52.
by crosen99 4y ago
Yes, it can be. But apart from the accuracy improvements I mentioned, which may make this a much less severe issue in fast order, there's also the notion that users will become more adept at navigating such perils, just as users devel
53.
by crosen99 4y ago
Those who object to or are skeptical about ChatGPT for search on the basis of accuracy issues or content rights are missing some crucial points. Accuracy issues will improve. Mostly likely the improvement necessary for LLMs to answer a vas
54.
by crosen99 4y ago
Chat interfaces won't completely replace existing knowledge sources anymore than Web sites have completely replaced books. But books can't answer questions a reader has about it's text, or summarize themselves in the context
55.
by crosen99 4y ago
Well, you need one level of readiness for life and death apps and a whole other level for entertainment apps and then there’s everything in between. One can debate just how much of the market falls into the safer range of the spectrum, but
56.
by crosen99 4y ago
Getting you to your product page more directly doesn’t eliminate marketing opportunities. It just changes them. When you land quickly on your product page you can still be peppered with other buying opportunities. And in the cases of admin
57.
by crosen99 4y ago
Amazon is already close to this sort of interface from a UX perspective because the main feature on the Amazon www homepage is a search field. But, here are a couple of difference that the user would see at chat.amazon.com compared to the
58.
by crosen99 4y ago
No, the thought is not about putting a chatbot on a subdomain. It's about two things: 1. The notion that chat.company.com becomes the single entry point to a company that can effectively launch your journey more efficiently than www o
59.
by crosen99 4y ago
No, that’s not what I want. It could be a new single all encompassing entry point for any customer/persona to most efficiently get either the information they want or the application/person they need to handle a hard transaction.
60.
by crosen99 4y ago
Thoughts after reading the comments so far: I acknowledge ChatGPT grade models are very expensive to operate and that there are still real performance limitations/issues that make it not yet appropriate for autonomous handling of hard

More ›