Nathan Danielsen
banner
ndanielsen.bsky.social
Nathan Danielsen
@ndanielsen.bsky.social
Living the dream. @pydatala, @peacecorps, @scikit_yb also slow triathlete but getting faster. Deep into automotive data
That moment before you begin writing the code for an exciting POC after brainstorming for a day. You know that you won't stop until it's a game changer....
January 24, 2025 at 7:21 PM
Heartbreakingly overheard in a playground in San Diego. Young kids talking about living in Airbnbs, cars being packed up with entire house, and origins of the Los Angeles fires.
January 13, 2025 at 7:48 PM
Fly over at the #roseparade #pasadena
January 1, 2025 at 4:09 PM
Made the fatal mistake of messing with my pytorch environment in n my Mac studio.
December 31, 2024 at 7:53 PM
Anyone have any good tools using llms to generate docs from a code base?
December 31, 2024 at 5:36 PM
Pattern that I'm encouraging at work: prototype on the frontier LLMs. In 6 months, there will be an equivalent open weight model to build production on.
December 23, 2024 at 5:38 AM
Peak millennial parenting is feeding your kids avocado toast
December 22, 2024 at 8:05 PM
Saturday morning dad time, rewatching Zach Snyders Justice League.
December 14, 2024 at 3:02 PM
Muppets Christmas Carol hits a little different as a parent.
December 14, 2024 at 3:09 AM
Always a bummer to see new and exciting llm models released under a weird private only use model. Ie athene-v2
December 8, 2024 at 7:59 PM
My 4 year old showed me 5 yoga poses. I like what the preschool is teaching
December 8, 2024 at 1:27 AM
Reposted by Nathan Danielsen
This is such an interesting angle, because the number of large organizations that have this level of discipline to evaluating new models is likely to be tiny

Running a bunch of test prompts specific to what your company does through a new model feels like it should be pretty low hanging fruit
A test of how seriously your firm is taking AI: when o-1 (& the new Gemini model) came out this week, were there assigned folks who immediately ran the model through your internal, validated, firm-specific benchmarks to see how useful it as? Did you update any plans or goals as a result?
December 7, 2024 at 5:00 PM
My algo is better here than on that other Xocial app. More software development, ml, AI here and much less screaming politics
November 26, 2024 at 1:51 AM