Whenever I encounter a "data lake" it's just a bunch of random CSV and JSON files on S3, and a bunch of half baked Python scripts to query it. And usually somebody tells me it's "big data", and that's why it is has to be that way. So I check the size and it would easily fit on a laptop from 15 years ago.
Not a data lake expert here, but I think there's a parallel with telemetry/logging for video games. The author talks about mass "ingestion without purpose," which is what I've sometimes flubbed in the past—during playtesting, we'd collect piles of data about where the players won/lost/moved/whatever, and there was so darned much of it that we'd just toss our hands up and not do anything with it.
The article suggests (sensibly!) that the right thing to do is to actually go in and design what's being logged with purpose (i.e., everything should answer a known question). But I don't always know what I'm looking for when I'm testing a thing. At the risk of using LLMs as a hammer that hammers everything, I wonder if they'd help with that problem?
I've pointed Codex at Posthog logs, for instance, and it's been able to say, "yeah, well, look at this pile of residential proxies hitting your site." I wonder if it'd similarly be able to say, "hey, I notice a pattern where everyone who picks the Cleric class stops playing after level 3"?
I'm building a product and exploring these questions in that domain. I'm thinking about it a little differently. I have a known set of hardware sensors that I'm collecting all the data from. Per sensor the data has a strong schema with a timestamp, what is interesting is the insights that can be gleaned across sensors. So what sensors do you have that you want to record, have a strong schema against that sensor, now have something you can use to tie across sensors, now you can use LLMs to query the data and draw interesting insights
The people who told us to push all of our data to the data lake are the ones who are now telling us to feed all of our knowledge to AI. In both cases, the dream was to replace us with computation, either ML or LLM's. We survived the first one relatively unscathed, don't know about the second one yet, at least for my own job.
It’s hard to take an article seriously that uses a snowflake ceo maligning Hadoop as evidence of something. (Note, I am already a believer that data lake efforts usually ended badly.)
Regardless of whether their criticism is factual, it’s absurd to use a direct competitors assessment of the competing tech as a credible source!
If you do that in the opening paragraphs of your paper, it’s wrecked all credibility for everything that follows.
Whenever I encounter a "data lake" it's just a bunch of random CSV and JSON files on S3, and a bunch of half baked Python scripts to query it. And usually somebody tells me it's "big data", and that's why it is has to be that way. So I check the size and it would easily fit on a laptop from 15 years ago.
Not a data lake expert here, but I think there's a parallel with telemetry/logging for video games. The author talks about mass "ingestion without purpose," which is what I've sometimes flubbed in the past—during playtesting, we'd collect piles of data about where the players won/lost/moved/whatever, and there was so darned much of it that we'd just toss our hands up and not do anything with it.
The article suggests (sensibly!) that the right thing to do is to actually go in and design what's being logged with purpose (i.e., everything should answer a known question). But I don't always know what I'm looking for when I'm testing a thing. At the risk of using LLMs as a hammer that hammers everything, I wonder if they'd help with that problem?
I've pointed Codex at Posthog logs, for instance, and it's been able to say, "yeah, well, look at this pile of residential proxies hitting your site." I wonder if it'd similarly be able to say, "hey, I notice a pattern where everyone who picks the Cleric class stops playing after level 3"?
I'm building a product and exploring these questions in that domain. I'm thinking about it a little differently. I have a known set of hardware sensors that I'm collecting all the data from. Per sensor the data has a strong schema with a timestamp, what is interesting is the insights that can be gleaned across sensors. So what sensors do you have that you want to record, have a strong schema against that sensor, now have something you can use to tie across sensors, now you can use LLMs to query the data and draw interesting insights
The people who told us to push all of our data to the data lake are the ones who are now telling us to feed all of our knowledge to AI. In both cases, the dream was to replace us with computation, either ML or LLM's. We survived the first one relatively unscathed, don't know about the second one yet, at least for my own job.
It’s hard to take an article seriously that uses a snowflake ceo maligning Hadoop as evidence of something. (Note, I am already a believer that data lake efforts usually ended badly.)
Regardless of whether their criticism is factual, it’s absurd to use a direct competitors assessment of the competing tech as a credible source!
If you do that in the opening paragraphs of your paper, it’s wrecked all credibility for everything that follows.
I'd say Bob Muglia is pretty credible