Brian Olsen

I'm no longer active on Xitter. Find me elsewhere:

[Brian Zhan @brianzhan1]

1/2 We kept hearing about @LanceDB from companies like MidJourney that were using its open-source Lance table format for AI data. They saw orders of magnitude improvement in latency, storage costs, and more. A quick experiment confirms why: https://t.co/Ojndl1kEvk.

[Brian Olsen @equijoin]

Bergy Bits: Scaling PyIceberg goes Daft https://t.co/QRikCJNeyq

[Brian Olsen @equijoin]

@findepi @mim_djo I don't think it's good or very bad...I would put it in the less-than-ideal bucket.

[Brian Olsen @equijoin]

@mim_djo @dennylee Sure but how is that enforced was my question.

[Brian Olsen @equijoin]

@mim_djo Ah yes, I had son some of the discussions happening around ownership which is something being added to v3 https://t.co/tguHY1OiIh.

@dennylee does delta have an ownership model or when do you check for a table collision?

[Brian Olsen @equijoin]

@tlberglund @rmoff

[Brian Zhan @brianzhan1]

GPT is not always the best model. Mistral is better at legal reasoning, Llama is better at stories.

@LeptonAI : ML engineering = Vercel : web dev. With a couple lines, you can deploy and fine tune any model, w/ metrics, monitoring, autoscaling in place: https://t.co/FNdqR9tFjg

[Brian Olsen @equijoin]

@kjaymiller I was hoping to see a shirt that had a guy in handcuffs and the text saying mmmmmhhmmm

[Brian Olsen @equijoin]

@DashBarkHuss Is this the dude from House of Dragons?

[Dashiell Bark-Huss @DashBarkHuss]

Indiehacker pro tip: You don't have to block these people

[Brian Olsen @equijoin]

@anna__geller I feel like this could be quickly misinterpreted in the cutthroat capitalistic machine we’re in.

“Jessica was taking a snaky approach and I just couldn’t follow.”

“He’s very snaky so I still check his indentation or anything else isn’t off.”

[Brian Olsen @equijoin]

@malware_yml https://t.co/mHPRgO1LRq

Or they just call it English lol

[Brian Olsen @equijoin]

@ananthdurai Nah, you have to relate yourself to the ai boom, like say you integrate with a vector store, or have a function that asks ChatGPT something, and call it AI, or create a new query language on LLMs and call it English

[Brian Olsen @equijoin]

@anna__geller Or compromise: “^(sales|has_people)$”

[Finding Serenity 🌱🏳‍🌈🦄 @mmysticbloom]

This I can definitely resonate with & talk about in great detail!! Who else?

#ActuallyAutistic #AskingAutistics #AUDHD #ADHD #Neurodivergent #MentalHealth

[Brian Olsen @equijoin]

@neelesh_salian And I’m losing my house

[Brian Olsen @equijoin]

If you haven't followed @anna__geller yet, she has an amazing writing style that lends itself to quickly ramping up on a variety of topics. As opposed to my goofy style, she keeps things interesting with story-telling, while keeping the lessons learned so clear.

[Brian Olsen @equijoin]

This blog covers two topics near and dear to my heart.

I'm a huge @trinodb and @ApacheIceberg fan. Anna shows the most common implementations of Trino in the wild, @AWS Athena, querying Iceberg with a #AWSGlue.

Of course, all this can be orchestrated by @kestra_io. 😉

[Anna Geller @anna__geller]

New blog post: Apache Iceberg Crash Course for AWS users with a juicy narrative 🫐

https://t.co/9gjnBceK4H

[Brian Olsen @equijoin]

@Afr030B0TiCiST

[Brian Olsen @equijoin]

I love data vendors, I work for a #datavendor, but also... 🤦‍♂️ data vendors...

#dataengineering