Martha van Berkel
speaker
1,330 appearances
7 recordings
1 series
first heard Oct 2024
last heard Jun 2025
Martha van Berkel’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsNo recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.
Appearances
But we're gonna have to just like journey on that.
Like I would say, like we're still pretty early days in that.
I'm excited because we're gonna we're gonna just like make that happen.
I think it's going to be data feeds with context.
And and I I say that very specifically because I don't know, I read like four articles this week where it's like, you lm.txt, like that's not going to be good enough, right?
The reason why all the research says knowledge graphs and large language models work well is because it's data with context.
So it's structured data.
Think that broad concept.
But it has context, which means you're defining relationships between things.
Because it's when those relationships are defined that you can do inferencing.
And the problem we're solving for large language models is the inferencing part.
That's the expensive part, right?
It's the part that hallucinates.
And so I think like historically, if you'd asked me a year ago, I'd be like,
Enterprise are going to be investing in knowledge graphs.
I think like as we have new standards, like that's hard to do and we're seeing people still do that work.
But I think the most important, the reason why knowledge graphs are important, because it's data with context.
I'm buying that and I think they're gonna demand um
Well, I buy that that's gonna be like the trend 'cause it's gonna make it cheaper.
And I think they're gonna run into challenges around sort of making money and making it make sense to do that much compute power.
Showing 61–80 of 1,330 · page 4 of 67
← Previous
Next →