@Nightshadessx@Yuchenj_UW I usually see:
“Astra is amazing, Claude sucks.”
“Claude sucks, Astra is out of this world.”
“Please give us a reset ______”
“Frontier models are cooked, go open source!”
“Bring back 4o”… which they should.
As much as I wanted to get some early rest tonight, I've seen many posts about @typesafeai's Jev from practitioner folks that I respect (aka, not AI hypefluencers), so, checked it out.
As a sucker for cost savings and efficiency, I love the vision and approach. Old-school feel to dev work in an agentic area.
@elonmusk Honestly @elonmusk, it was a rocky start, but as someone who spends most of his time on another social media platform, X is getting better and better.
The one issue I have is that unique content is hard to find. One good unique post for every 5 copycat type posts.
Ali Ghodsi explains how Databricks’ token consumption strategy has evolved:
“Around Q4 last year was when the models got really, really good, and we started noticing that it’s actually starting to give much better productivity.”
“It’s called Unity Gateway, this gateway used to provide token capacity. So you can get OpenAI, Anthropic, Gemini, Grok capacity.”
“Any customer can come to us, and we’ll just provide them that capacity because we have a relationship with those, and any open source model.”
“We started putting budget constraints in place and giving people warnings. Like, ‘You have this much of your budget left. You’re getting close to your ceiling.’ So we started doing that per person and per group, and then we started doing great analytics, so we could predict exactly where the costs were going.”
“We were able to actually bend the curve. The tokens have continued to go up, but the costs have been stagnant.”
Amazing chat @alighodsi@martin_casado
@mcuban I have nothing against Andrew, and I am somewhere down the middle on this whole debate, but his choice of words is quite interesting to dance around facts.
Had an interview for a data role.
Candidate opened Snowflake.
My CFO ended the interview.
Then interviewed someone else.
Candidate opened Databricks.
We ran out of time waiting for the cluster to start.
Another person showed up.
Candidate opened Fabric.
Told the candidate to come back in 3 years.
Last interview of the day came in.
Candidate opened Excel.
They are now my boss.
#satire
Interviewed a new candidate today. First thing he asked is if it was okay to use AI
I said "Of course, use any tool you want, I just want to better understand how you think and problem solve"
He starts sharing his screen and opens Claude. So I hung up and ended the interview
@gdb Personally, Astra has felt like the first model that truly completes whatever task I throw at it. And when rework is needed, gets it right the second time.
@pwendell Been using Astra on High and Extra High. Been able to tackle challenges that I haven't been able to solve well with other models. Very happy with it.