petervandijck

11 ms·

31.
by petervandijck 3y ago
There is more out there: https://arxiv.org/pdf/2308.13507.pdf https://assets.amazon.science/ca/33/5d25245f4e34b3eadac7e983...
32.
by petervandijck 3y ago
I found this here https://ar5iv.labs.arxiv.org/html/2212.07769
33.
Ask HN: Research on LLMs asking clarifying questions
1 points by petervandijck 3y ago|2 comments
34.
by petervandijck 3y ago
An additional (and I was quite surprised by this) trick is to ask an LLM to reformulate the user prompt “to be more concise and precise”, then run a vector similarity search against that, which (in our experience) leads to better matches. T
35.
by petervandijck 3y ago
I built an app to give us analytics on an LLM we are building for a client. It focuses specifically on “what people are talking about”. It’s literally one call to Gemini in terms of LLM usage… https://simplyanalyze.ai/ (We
36.
by petervandijck 3y ago
(Do you remember this btw https://petervandijck.com/xfml/spec/1.0.html )
37.
by petervandijck 3y ago
Simon! Still loving your blog posts about this stuff. Thank you for doing that. Agreed that not everyone lives in the terminal, but you know.
38.
by petervandijck 3y ago
It's "Retrieval Augmented Fine Tuning". The related blog post is interesting: https://gorilla.cs.berkeley.edu/blogs/9_raft.html
39.
by petervandijck 3y ago
why NexusRaven specifically, what has your experience been?
40.
Ask HN: What are you building LLM/RAG chatbots with
7 points by petervandijck 3y ago|9 comments
41.
by petervandijck 3y ago
Ongoing product development with migrations, tweaking indexes etc. becomes really hard. Every small database tweak now has to be deployed over 1000s of databases.
42.
by petervandijck 3y ago
Love that you're tackling this and congrats on the launch. Feedback: the synthetic data demo shows nicely what that piece does, but the page is really messy, it would be nice to have that demo cleanly on a dedicated page: https:/
43.
by petervandijck 3y ago
I don't think it's standard at all. Hand out equity for cash, not for "advice". Advice is not work, and it's not cash. They take zero risk, they get no equity.
44.
by petervandijck 3y ago
Improve probably, but speed also matter, I worry a lot about chaining lots of LLM calls together, each of them takes a bunch of seconds, and then the experience becomes just really slow.
45.
by petervandijck 3y ago
Just generally a lot of abstractions that seem sometimes overwrought, and seem to hide details.
46.
by petervandijck 3y ago
For example, for us, we are building an LLM chatbot that pulls in the data of a technical book publisher. They have 20 years of technical books, and 20 years of videotaped conference talks. Hard: - We're using LangChain, which isn&#x27
47.
Who is building LLM Chatbots, and what issues are you running into?
16 points by petervandijck 3y ago|11 comments
48.
by petervandijck 3y ago
We are working on a way to see what people are talking about in your LLM chatbots. It’s basically conversation analytics. It’s early but the developer preview is out, you can create a free account and get analytics going if you have a live
49.
by petervandijck 3y ago
BTW, as a rule, advisors get no equity. If they believe in this company and want to help it succeed, they can invest in it with a safe, which gets them equity rights and a strong incentive to help you succeed.
50.
by petervandijck 3y ago
This is exactly right. If you agreed to give him more than 2 or 3 % equity as an advisor, you have possibly killed your startup before it has a chance to take off. Advisors should honestly get no equity ever. Talk to him again and talk this
51.
by petervandijck 3y ago
Additionally, if you have a fairly specific use case for something like this, we’d love to hear it. We’re very early and still trying to figure out what people need of a product like this. Our first customer is in publishing, and is particu
52.
Show HN: Analytics for LLM Chatbots(simplyanalyze.ai)
3 points by petervandijck 3y ago|1 comments
53.
Y Combinator website gives a 500 error(ycombinator.com)
3 points by petervandijck 3y ago|0 comments
54.
Why do prompts even work at all?
3 points by petervandijck 4y ago|4 comments
55.
by petervandijck 4y ago
Learn SQL and use the ORM is usually the correct answer. Let it print out the SQL queries and how long they take (ORMs do that out of the box), then check for indexes and other standard issues. If there’s some unusual complex query to be
56.
by petervandijck 4y ago
I'm launching climatelink, it's basically trying to be the Techmeme for climate. The problem we’re trying to solve is that today, really great climate change insights are just hard to find. Most good reporting lives behind paywall
57.
Climatelink.io (launching) tries to solve Don't Look Up(climatelink.io)
2 points by petervandijck 5y ago|1 comments
58.
by petervandijck 5y ago
We're launching climatelink.io today. Climate change is complex, scary, real and the mother of all collective action problems. We're building a social network focused exclusively on saving the planet. In my own climate journey, a
59.
by petervandijck 6y ago
Google storage. Dropbox. Dashlane. Netflix.
60.
by petervandijck 7y ago
Spend time understanding your current traffic and users. - For the traffic: do the users just find it on Google Search, and then leave immediately? What are they looking for? - For the requests: is it all from 1 or 2 users? Are there many u

More ›