Host of Vector Podcast: https://xitter.cf/t.co/CogWF5e53n Director of Search, @aiven_io
Earth
Joined May 2011
- Tweets4.3K
- Following730
- Followers1.3K
- Likes2.8K
Pinned Tweet
It was exciting to give the keynote on stage of the OpenSearchCon in San Jose last week together with Tony Piazza, Senior Solutions Architect at @aiven_io to discuss how we run managed OpenSearch for everyone everywhere and our upstream contributions around migration tooling linkedin.com/feed/update/urn…
Dmitry Kan retweeted
Weekend hack: HackerTube. I was annoyed with YouTube's recommendations so I had Claude filter down 43M comments over 20 years of HN to 6,454 videos that keep being recommended. *Significantly* better viewing experience, ~2 hours with Opus 5.5:
hackertube.org/
- Excellent recs without any logins (state is kept in local browser session)
- Thumbnail is a representative frame from middle of the video instead of the super duper A/B test hyper-optimized user uploads
- Underlying data from the comment analysis as one zip file, point your Claude at it and ask for your own version
- HackerTube Longs ✊
A bit more on the process:
I use Claude Code (desktop app for Mac), Opus 5.5 on Medium effort as my default. I only talk to Claude and have it do everything, the only "manual work" I helped with in tandem was in the browser (reminiscent of MenuGen): 1) I had to create some kind of API key from Google Cloud UI to query YouTube for video metadata, 2) I booted up a VPS box for serving and 3) I bought the domain name. Once the box was up I gave Claude the IP and told it update, secure, harden, deploy and optimize the code, and it did the whole thing.
Many awesome projects that would have normally taken weeks/months even can now be few-shotted for breakfast. What else are we not thinking of.
What sometimes creates pressure in life: is sifting through non-quality stuff. Our modern setup globally is: get something non-quality fast. That is why, if you value quality, you need time filtering and need enormous patience to get to the quality. But don't give up.
The best way to get interviewed (and tested) is by being stopped at the conference in San Jose and speak unscripted. You get to say what actually excites you about the work you are doing, and you can't really hide your emotion or blame jet lag ;) linkedin.com/feed/update/urn…
Dmitry Kan retweeted
How come Russia isn’t being held accountable for violating the “energy ceasefire” President Trump said he wanted Russia and Ukraine to abide by?
Russia fired 39 ballistic missiles into Ukraine last night, most of which targeted Ukriane’s energy facilities.
Putin is the Hitler of our time.
Really excited to partner with you guys!
Today @itsklipy is powering 10 billion searches monthly. Isn’t that cool?
Just imagine how many conversations, reactions, and moments KLIPY is powering every single day. And as our distribution scales, so do the new products we’re building on top of it, including AI Games, GenAI via API, KLIPY Ads Manager, and more.
We’re super excited to keep bringing new expression experiences to developers and users around the world.
Thanks for the shoutout at OpenSearchCon, @DmitryKan and the team at @aiven_io. We’re very excited to be working with you on this!
Dmitry Kan retweeted
Live now at OpenSearchCon North America: Dmitry Kan and Tony Piazza, Aiven, on their ambition for OpenSearch everywhere, for everyone.
Join now > bit.ly/4rrhnuZ
One positive thing LLMs as coding agents did: they removed the fear of building. It is not about wether you can build it, you can (in some approximation). It is why this matters, who will use it, and then of course making it dependable long term.
Excited (and terified) to be invited to give a keynote at the OpenSearchCon in San Jose this month with Tony Piazza (who has been on a roll contributing upstream).
🎙️ We’re excited to welcome @DmitryKan, Product Director, Search, at Aiven to the #OpenSearchCon North America keynote lineup. Since 2010 he's specialized in search (Lucene, Solr, Elasticsearch), co-owned TomTom's planet-scale map search, and hosts the Vector Podcast. Don’t miss his talk!
Sept 22-24 | San Jose, CA
Register: bit.ly/3NXhWNN
Thanks @srbhr_ for writing this coverage of my episode with @FlaxSearch vectorpodcast.com/blog/charl…
Call me old (programmer), but abstracting a CLI call through an LLM is too much. Why burn tokens for a direct tool call I know how to use?
I just published The difference between then and now medium.com/p/the-difference-…
I'm burning token quota with a single agent quite fast. Are you saying, with multiple agents you will burn 10x faster? What's the point?
45-th episode on Vector Podcast: with @FlaxSearch - finally! We recorded live on @berlinbuzzwords 2026 - lots of really good advise for the guy who has seen it all in Search: youtube.com/watch?v=m8D0iBo0…
Sometimes I think: would I be better coding everything by hand? At least I'd know where I want to go exactly :) Out of this: Fable does not really work for me, taking too long. GPT-* models are more economical in time and lines of code modified (not an ad, just an observation)
Dmitry Kan retweeted
Really is a joy to get hands on with code, forcing it into shape, so that you know, for certain
(a) The architecture has meaningful e2e behavioral tests you trust will test something real
(b) Remove hidden agent-injected complexity, duplication, structurally stupid architectural choices
(c) Develop visceral, hands-on opinions about what makes maintainers lives easy / hard (agent or human)
Then I love begin going to town with the agent to stamp out e2e tests and other bits of functionality that conform to patterns I've established
Dmitry Kan retweeted
2 billion downloads. 15,000+ community members. 1M+ monthly visitors.
OpenSearch is scaling as a community-driven foundation for search, analytics, and AI.
In this episode of Front Page, Bianca Lewis, Executive Director of the OpenSearch Software Foundation, discusses open governance, hybrid search, and how OpenSearch helps organizations build AI applications more efficiently.
Watch: bit.ly/44eQi3u
The fact that LLM is trying all these, on 3rd iteration of trying to make it work, says something about fundamentals
It was a real fun to present (as is probably evident from our talk recording) with Hans Kristian Flaatten 🕊️🍉 on stage of the OpenSearchCon in Prague - a proud moment for us at @aiven_io to have the case study up on the OpenSearch blog.
When the Norwegian Labor and Welfare Administration (NAV) set out to migrate from Elasticsearch and modernize its observability platform, it needed a solution that could support national-scale workloads without disrupting critical services. Learn how NAV partnered with member company Aiven and used OpenSearch to achieve a zero-downtime migration while preserving developer workflows and operational reliability.
Read the full case study: bit.ly/4e3Fnzp
New episode on Vector Podcast with Julien Nioche, creator of StormCrawler, who currently focuses on GreenOps and his open source framework SPRUCE: youtube.com/watch?v=-IY7s00g…
