If I read it right, the 4B ballpark figure is based on total annual per employee cost of 500k * 8000 employees, so the figure is actually 4B/year. 20B over five years.
lots of great stuff, but the plot in the paper is just chart crime.
different shades of gray for references where sometimes you see 4 models and sometimes 3.
this is vibe coded :D. the drop down for Python not working.
I remember when we started working on our teams bot in 2023, Microsoft announced that they will stop supporting Python for the teams sdk, which felt super short sighted. Eventually they silently picked it up again and the old sdk never stopped working.
Kudos for turning it around. I believe it'll last this time as AI agents in communication tools are something a lot of people want. and unlike pre LLM chatbots and agents, they are so much more useful now.
FAANG already did this all the time isn't it? Regardless of their policy. US is no better than China from my point of view. In this case, I see no difference between sending my prompts to US or China companies. At least China models are open source.
I accept that all the providers will do what I would consider unethical with my data and simply don't expose what I don't consider a price of doing the business I want.
The other criticism I see is "ask it what happened in 1989" but as a my use case isn't writing a high school history essay I simply don't care. Or believe one should seek those kind of answers from any AI. (If you're curious it simply cuts off the reply).
I fully appreciate that YMMV and what sits right for others will not align with what's acceptable to me. Anthropic and OpenAI both are in my badbooks as much as Z.ai. pick your poison as they say.
As someone working in the field I have to admit that I'm not familiar with the terms differential storage nor do I really understand what hybrid execution means. Maybe you could describe it both from a simple technical point of view and what benefits it has to me as a user?
Append-only layers with PostgreSQL metadata. DuckDB sees a normal file; OpenDuck persists data as immutable sealed layers addressable from object storage. Snapshots give you consistent reads. One serialized write path, many concurrent readers.
Hybrid (dual) execution
A single query can run partly on your machine and partly on a remote worker. The gateway splits the plan, labels each operator LOCAL or REMOTE, and inserts bridge operators at the boundaries. Only intermediate results cross the wire.
You know some people grow up in untrustworthy environments and auto didact their way to something like first principles thinking and depending on things shaking out you might only believe what you've personally seen with your own eyes. And well, earth looks pretty flat in daily life.
4B over 5 years is 20B, which is significant.
reply