David Tippett
@taidesu.bsky.social
Open Source enthusiast and builder of search systems. Currently building search at GitHub. Former DA for open-source OpenSearch @AWS. Opinions are my own.
Time to fight fire with fire. Even worse AI generated PR reviews.
February 9, 2026 at 5:10 PM
Time to fight fire with fire. Even worse AI generated PR reviews.
Me ships a really cool update that makes issues and prs pages faster.
GitHub:
GitHub:
February 9, 2026 at 4:44 PM
Me ships a really cool update that makes issues and prs pages faster.
GitHub:
GitHub:
The right way to think about it is to say that Refresh writes segments to disk. Translog "stamps" them as being official (Lucene commit).
If the Translog reaches it's flush size it'll trigger a refresh before committing the segment.
They're related but not as an either or but a both and
If the Translog reaches it's flush size it'll trigger a refresh before committing the segment.
They're related but not as an either or but a both and
February 6, 2026 at 3:31 PM
The right way to think about it is to say that Refresh writes segments to disk. Translog "stamps" them as being official (Lucene commit).
If the Translog reaches it's flush size it'll trigger a refresh before committing the segment.
They're related but not as an either or but a both and
If the Translog reaches it's flush size it'll trigger a refresh before committing the segment.
They're related but not as an either or but a both and
I just realized my understanding of the Elasticsearch/OpenSearch refresh + translog operations was INCORRECT.
I'd always had the mental model that refresh or translog flush would write segments. That's partially true...
I'd always had the mental model that refresh or translog flush would write segments. That's partially true...
February 6, 2026 at 3:31 PM
I just realized my understanding of the Elasticsearch/OpenSearch refresh + translog operations was INCORRECT.
I'd always had the mental model that refresh or translog flush would write segments. That's partially true...
I'd always had the mental model that refresh or translog flush would write segments. That's partially true...
Now with copilot instead of writing just a few hundred lines of bad code a week I can scale to TENS OF THOUSANDS of lines of bad code a week.
January 26, 2026 at 3:51 PM
Now with copilot instead of writing just a few hundred lines of bad code a week I can scale to TENS OF THOUSANDS of lines of bad code a week.
The amount of logistical things needed to build a startup is so frustrating. I don’t want to build a company I want to ship code that helps people.
Today I:
- created a bank account
- migrated emails
- mailed so compliance documents
Today I:
- created a bank account
- migrated emails
- mailed so compliance documents
January 22, 2026 at 6:46 PM
The amount of logistical things needed to build a startup is so frustrating. I don’t want to build a company I want to ship code that helps people.
Today I:
- created a bank account
- migrated emails
- mailed so compliance documents
Today I:
- created a bank account
- migrated emails
- mailed so compliance documents
I’m pretty sure we’re on like web 7.0 according to a bunch of random crypt bros 😆
January 20, 2026 at 12:30 PM
I’m pretty sure we’re on like web 7.0 according to a bunch of random crypt bros 😆
So it’s funny the bus thread was right above this one in my feed but the “I didn’t get enough likes” I’ve not yet seen… 🔗 👀
January 20, 2026 at 12:28 PM
So it’s funny the bus thread was right above this one in my feed but the “I didn’t get enough likes” I’ve not yet seen… 🔗 👀
Okay, so not only did I do a thing with Tanki I am actually starting a serious company:
relevan.dev
We're building the future of relevancy based search. We're pre-launch (despite me launching this technically here). But we want to talk to small e-commerce businesses about their search problems.
relevan.dev
We're building the future of relevancy based search. We're pre-launch (despite me launching this technically here). But we want to talk to small e-commerce businesses about their search problems.
Relevan · Yet Another Search Site
Relevan — Relevancy as a Service
relevan.dev
January 19, 2026 at 8:58 PM
Okay, so not only did I do a thing with Tanki I am actually starting a serious company:
relevan.dev
We're building the future of relevancy based search. We're pre-launch (despite me launching this technically here). But we want to talk to small e-commerce businesses about their search problems.
relevan.dev
We're building the future of relevancy based search. We're pre-launch (despite me launching this technically here). But we want to talk to small e-commerce businesses about their search problems.
Got bored… made an app this weekend to help me track my aquarium
tanki.dev
In case anyone else with a #reeftank wants a free place to track parameters over time 👀
tanki.dev
In case anyone else with a #reeftank wants a free place to track parameters over time 👀
Tanki
tanki.dev
January 12, 2026 at 10:04 PM
Together though we nailed the problem down in probably 1/10th of the time it would’ve taken before.
January 9, 2026 at 4:12 AM
Together though we nailed the problem down in probably 1/10th of the time it would’ve taken before.
On top of that because we’d elected for a LOT of smaller shards (small relative to GitHubs size mind you) we were ending up with a lot of really small segments which couldn’t be efficiently compacted because of concurrent merge limits.
Copilot was never brining me that far on its own.
Copilot was never brining me that far on its own.
January 9, 2026 at 4:12 AM
On top of that because we’d elected for a LOT of smaller shards (small relative to GitHubs size mind you) we were ending up with a lot of really small segments which couldn’t be efficiently compacted because of concurrent merge limits.
Copilot was never brining me that far on its own.
Copilot was never brining me that far on its own.
After a bit of troubleshooting I realized what was happening. Every time segments merge in ES the vector graph has to be rebuilt. I knew that.
What I hadn't thought through is that because BBQ uses a centroid every segment merge we had to re-quantize all the documents.
What I hadn't thought through is that because BBQ uses a centroid every segment merge we had to re-quantize all the documents.
January 9, 2026 at 4:12 AM
After a bit of troubleshooting I realized what was happening. Every time segments merge in ES the vector graph has to be rebuilt. I knew that.
What I hadn't thought through is that because BBQ uses a centroid every segment merge we had to re-quantize all the documents.
What I hadn't thought through is that because BBQ uses a centroid every segment merge we had to re-quantize all the documents.
We were seeing ingestion backing up exponentially. I had an intuition that it was a bottleneck with segment merges.
I didn't have to remember which endpoints show merge stats. I just had copilot shoot out a profiling script and ran it on an interval.
I didn't have to remember which endpoints show merge stats. I just had copilot shoot out a profiling script and ran it on an interval.
January 9, 2026 at 4:12 AM
We were seeing ingestion backing up exponentially. I had an intuition that it was a bottleneck with segment merges.
I didn't have to remember which endpoints show merge stats. I just had copilot shoot out a profiling script and ran it on an interval.
I didn't have to remember which endpoints show merge stats. I just had copilot shoot out a profiling script and ran it on an interval.
For example, we just came across an issue with indexing vectors into Elasticsearch. ChatGPT/Copilot/Claude were all stumped.
Thankfully I (mostly) know what I was doing and was able to guide them through the issue.
Deep expertise is still needed but is going to be harder and harder to come by.
Thankfully I (mostly) know what I was doing and was able to guide them through the issue.
Deep expertise is still needed but is going to be harder and harder to come by.
January 9, 2026 at 4:02 AM
For example, we just came across an issue with indexing vectors into Elasticsearch. ChatGPT/Copilot/Claude were all stumped.
Thankfully I (mostly) know what I was doing and was able to guide them through the issue.
Deep expertise is still needed but is going to be harder and harder to come by.
Thankfully I (mostly) know what I was doing and was able to guide them through the issue.
Deep expertise is still needed but is going to be harder and harder to come by.
Not going to lie so much engineering work is getting replaced by AI.
It’s shocking but l’m starting to see a world where engineers write less code and do a lot more product + architecture work.
You very much still need someone with expertise at the wheel.
It’s shocking but l’m starting to see a world where engineers write less code and do a lot more product + architecture work.
You very much still need someone with expertise at the wheel.
January 9, 2026 at 4:02 AM
Not going to lie so much engineering work is getting replaced by AI.
It’s shocking but l’m starting to see a world where engineers write less code and do a lot more product + architecture work.
You very much still need someone with expertise at the wheel.
It’s shocking but l’m starting to see a world where engineers write less code and do a lot more product + architecture work.
You very much still need someone with expertise at the wheel.
three things people mean when they say they’re working on a startup:
- side project that you’re “working on” and “has potential”
- you’ve spent all your time building the perfect c-corp, landing page, and name…
- and then there’s actually building a startup
- side project that you’re “working on” and “has potential”
- you’ve spent all your time building the perfect c-corp, landing page, and name…
- and then there’s actually building a startup
January 5, 2026 at 12:40 PM
three things people mean when they say they’re working on a startup:
- side project that you’re “working on” and “has potential”
- you’ve spent all your time building the perfect c-corp, landing page, and name…
- and then there’s actually building a startup
- side project that you’re “working on” and “has potential”
- you’ve spent all your time building the perfect c-corp, landing page, and name…
- and then there’s actually building a startup
I just had ChatGPT respond to my question with an OpenSearch diagram that I was the author of.. Is this what it means to be full circle?
December 30, 2025 at 8:06 PM
I just had ChatGPT respond to my question with an OpenSearch diagram that I was the author of.. Is this what it means to be full circle?
Finally, wrapping up a big PR. Everything's working time to refactor a few bits and bobs anddddd... broken... why broken 😞
December 19, 2025 at 2:36 PM
Finally, wrapping up a big PR. Everything's working time to refactor a few bits and bobs anddddd... broken... why broken 😞
They can’t figure out how to build product that people will pay for so they charge people for things they’ve already given for free.
December 17, 2025 at 9:38 PM
They can’t figure out how to build product that people will pay for so they charge people for things they’ve already given for free.
Nothing like provisioning a whole ~100 node cluster just to realized the attached disks aren't large enough 😮💨
This is what I get for not checking my SKU's more closely 😭
This is what I get for not checking my SKU's more closely 😭
December 10, 2025 at 3:47 PM
Nothing like provisioning a whole ~100 node cluster just to realized the attached disks aren't large enough 😮💨
This is what I get for not checking my SKU's more closely 😭
This is what I get for not checking my SKU's more closely 😭
IMO you make the difference. There are a lot of people that click content simply because Jeff Geerling has produced it and the first two don’t surface that as quickly.
December 9, 2025 at 4:43 PM
IMO you make the difference. There are a lot of people that click content simply because Jeff Geerling has produced it and the first two don’t surface that as quickly.
Perfect! I can creat yet another PowerBI report that my executives will love and not look at even once.
December 7, 2025 at 6:00 PM
Perfect! I can creat yet another PowerBI report that my executives will love and not look at even once.
Also, there’s the concept of “food desserts” where the higher quality food is not readily available in stores near them further driving them to choose worse options.
I’m not going to shout into the void over there. No point fighting with someone who couldn’t be convinced they’re wrong.
I’m not going to shout into the void over there. No point fighting with someone who couldn’t be convinced they’re wrong.
December 2, 2025 at 3:55 PM
Also, there’s the concept of “food desserts” where the higher quality food is not readily available in stores near them further driving them to choose worse options.
I’m not going to shout into the void over there. No point fighting with someone who couldn’t be convinced they’re wrong.
I’m not going to shout into the void over there. No point fighting with someone who couldn’t be convinced they’re wrong.