Valquery

And just like that, AI User Group's inaugural conference is over. What a blast it was! Yesterday was Developer Day, the one nearest and dearest to my heart, and I saw a lot of rad things: - Lightning AI has maybe the best developer experience I've... experienced. Bonus points for being the only thing I was able to actually deploy and send to Peter Kunhardt during a workshop. I'd put the link here but I'm out of credits ?? - Grit is my favorite type of AI product: something that takes a tedious but attention-demanding task and intelligently automates it. Glad I accidentally was at the main stage for their presentation. - Qwiet AI I'd put in the same bucket, takes something that isn't at the top of most developers list of fun things to do and use AI to handle it intelligently. - I need to look into bringing OpenAI's evals library into the Valquery loop. Even if it means having non-JavaScript code in the app. All in all the vibes were really good. No one seemed to be trying to automate me out of a job, just automate me into doing a better job at my job. I'll take it.

Let's talk about GAR. I have been struggling to think of a better way to succinctly describe what Valquery does. It literally came to me while I was lying awake starting at the ceiling last week, and after a little informal soft launch at AI User Group's conference I think I've got it: GAR. RAG (retrieval-augmented generation) is all the (perhaps just most of?) the rage in the AI app space. Valquery does something similar, but there's an important difference. Rather than using reference materials to improve LLM responses, Valquery uses an LLM to ensure that the completeness and quality of the results returned from our search tool. We generate a text explanation of the search results while we're at it, and each of the results individually, but at the core we're using the LLM to serve search, rather than the other way around. Thus, GAR. What do you think?

Day 2 (Designer Day) of AI User Group's conference complete. Takeaways: - Lots of tools aiming to make prototyping and deploying web pages and apps easier for designers. The designs the apps spit out are... not the most tasteful... but having that skeleton there to start from seems like it will be helpful. - Voiceflow had a very cool demo. It reminded me of setting up a legacy call center management system, but actually good. - I can drink a pretty wild amount of fizzy water when I'm not paying attention. Tomorrow is Developer Day, my wheelhouse, gonna be a big one ??

Had a great time at Day 1 / Marketer Day at AI User Group's conference! A few takeaways, from my perspective as a founder: - Saw some very cool prospecting tools I am excited to take for a spin. Email marketing is not my specialty area so anything that can assist with that is super welcome. - Bestever AI apparently has an agent that checks for correct levels of color contrast in its ad generation? Incredible. - Might have to look into making a Zapier integration for Valquery, it would be nice to access to the zap ecosystem for ingesting data, at least as an option. Looking forward to Designer day tomorrow.