My newsletter on building cool data stuff: https://buttondown.com/datadash
Personal website: mcmullarkey.github.io
Opinions are my own and don't reflect the opinions of my employer
If you're not subscribed yet you can check out my latest Data Dash newsletter on how to do that here
buttondown.com/datadash/arc...
If you're not subscribed yet you can check out my latest Data Dash newsletter on how to do that here
buttondown.com/datadash/arc...
If you aren't subscribed yet you can learn how to have more fun writing a README here
buttondown.com/datadash/arc...
If you aren't subscribed yet you can learn how to have more fun writing a README here
buttondown.com/datadash/arc...
I'm good enough at coding to not break prod and good enough at humans that I used to be a therapist
Reach out via the email on the resume I've linked below or DM with leads, thank you!!
I'm good enough at coding to not break prod and good enough at humans that I used to be a therapist
Reach out via the email on the resume I've linked below or DM with leads, thank you!!
The first two principal components for embeddings of podcast episode descriptions line up really well with episode type labels
The first two principal components for embeddings of podcast episode descriptions line up really well with episode type labels
I try to walk what I talk, so I labeled all the episodes over the past year+ for one of my favorite podcasts
I used a Shiny for Python app I built, and that made the process go much faster vs. a spreadsheet
I try to walk what I talk, so I labeled all the episodes over the past year+ for one of my favorite podcasts
I used a Shiny for Python app I built, and that made the process go much faster vs. a spreadsheet
I reduced those embeddings to two dimensions + created this interactive plot
I reduced those embeddings to two dimensions + created this interactive plot
I used Qwen2.5:0.5b and got ~8 tokens per second
dev.to/koolkamalkis...
I used Qwen2.5:0.5b and got ~8 tokens per second
dev.to/koolkamalkis...
1. Take the (helpful!) approach @randyau.com describes in this screenshot
2. Spend an inordinate amount of time engaging in polite fictions
Engaging in some polite fictions is required in any job, but ideally doesn't constitute the team's entire role
1. Take the (helpful!) approach @randyau.com describes in this screenshot
2. Spend an inordinate amount of time engaging in polite fictions
Engaging in some polite fictions is required in any job, but ideally doesn't constitute the team's entire role
www.aisnakeoil.com/p/faq-about-...
www.aisnakeoil.com/p/faq-about-...
Even if you think you know what happened to Twitter this book is a wild ride worth taking! For example, Elon appeared to make multiple critical decisions based on the advice of his biographer(??)
Even if you think you know what happened to Twitter this book is a wild ride worth taking! For example, Elon appeared to make multiple critical decisions based on the advice of his biographer(??)
(This very much includes code I wrote while in academia. When we have too many jobs it's hard to do any of them as well as we'd like)
(This very much includes code I wrote while in academia. When we have too many jobs it's hard to do any of them as well as we'd like)
I followed this blog post by @simonwillison.net up until downloading the model
simonwillison.net/2023/Aug/1/l...
Downloaded the model using the instructions in this Github issue
github.com/simonw/llm-l...
I followed this blog post by @simonwillison.net up until downloading the model
simonwillison.net/2023/Aug/1/l...
Downloaded the model using the instructions in this Github issue
github.com/simonw/llm-l...
rachaelmeager.substack.com/p/finish-you...
rachaelmeager.substack.com/p/finish-you...