AI Food Fights in the Enterprise with Databricks' Ali Ghodsi

Ali Ghodsi, CEO and cofounder of Databricks, and Ben Horowitz, cofounder of a16z, explain the data wars happening inside and outside enterprises and how they could impact the evolution of LLMs.
[0:00] Why is it so hard for enterprises to adopt AI?
[3:08] Data wars
[4:28] Big vs. small LLMs
[8:13] Finetuning
[13:52] Open source AI
[17:51] Benchmarks
[19:30] Why Ali isn't afraid of AI
This conversation is part of our AI Revolution series, recorded August 2023 at a live event in San Francisco. The series features some of the most impactful builders in the field of AI discussing and debating where we are, where we’re going, and the big open questions in AI. Find more content from our AI Revolution series on www.a16z.com/AIRevolution.

[@SuperKillaki]

I really like how Ben asks questions. Simple and uses analogies that are easy to understand.

[@simonseverino]

Best AI conversation I heard this year.

[@bilalobeidat5772]

Great questions and great AI insights from Ali especially when he talks about the internet and what many people thought of the key players,...etc

[@videowatching9576]

Great content on ai! Particularly the point about so many databricks customers want mosaic that it’s massively gpu constrained.
Also was really fascinating to hear the Databricks founder say that Llama open source has substantially advanced the opportunities in AI, compared to if all dependent on closed models. Makes me wonder then about if open source models get to GPT-4 or future GPT-5 and beyond levels, combined with advances made by people in open source, then how this can accelerate. Including the example given of if training was fast, then doing fast training that finds ways to improve on itself then could drive ongoing further improvements.
Small suggestion for databricks founder: the charisma could be higher, and at places comes off as needlessly harsh in subtle ways, that could easily be improved. I think ChatGPT could identify places to be improved from the transcript.

[@balapillai]

Ben & Ali, I sense you two and this discussion can be the basis for prioritising.
For storytelling ease, say prioritising AI for the #SpiceTradeAsia region.
Aim: what mechanism(s) can we use so that we can continue where we left off on #SpiceTradeAsia region’s gradient ascent before it fell in circa 1450.What can we do to simplify this storytelling complexity?
Background: For a mental rejog/time relapse see esp Min 20 of
ruclips.net/video/NuZujx-LMfg/видео.html