Paco Xander Nathan1/
Beyond my obvious flippancy, OAI is a problem. Deets: https://medium.com/p/5358c2769936
IMO, most likely, over time, the MSFT/OAi thing resembles Intel/Cloudera.
Let's review a brief history of Big Data...
2/ circa 2009:
Cloudera: We dominate the Big Data future! All must obey, purchase our platform. Devs who've touched Hadoop source and are worthy of being added to the collective's perfection must assimilate ASAP, while all others become our sworn enemies! Resistance is futile.
Databricks: Hey, check out this Mesos demo, it's called Spark and it runs super fast. We're adding it to the Apache Software Foundation.
3/ circa 2011:
Hortonworks: LOL Cloudera, your Borg imitation was, like, hella funny as performance art, or something. Hey, can you spell "governance" using only eight different letters?
4/ circa 2013:
Apple: Hey Cloudera, you have 24 hours to return with Spark in the product offering. Or don't come back at all. You're welcome.
Cloudera: Yeah, um, so we're Cloudera, and we don't do Spark.
Apple: Yeah, um, so we're Apple. Wanna talk with VPs in Engineering who sign deals, or with our Legal who'll, like, get medieval on your hiney? Your pick.
Greenplum: Greetings, fellow ... [ audio cuts ]
5/ circa 2013 (cont'd):
Cloudera: (grudgingly) Yeah, um, so, Databricks, we need to cut a deal. Like *YESTERDAY*, for one year, *MAX*.
Databricks. Sure thing. Although you might not be quite as relevant by then, but meanwhile we're glad to help.
6/ circa 2014:
Google: Srsly, STFU! We *invented* map/reduce. Get a g-d life, would you?
Intel: We're investing insane amounts of capital into Cloudera, since they will inevitably dominate the future of all technology.
Wall St: LOL Intel. Also, you misspelled Hortonworks.
7/ circa 2016:
Industry: Yeah, um, so most of our use cases aren't actually all that "Big" and it's simpler just to use Spark.
8/ circa 2018:
Snowflake: Greetings, fellow Big Data platform providers.
Industry: I'll take "What is machine learning?" for $200
NVIDIA: Hey, we like machine learning too!
Wall St: LOL Cloudera.
9/ circa 2021:
Wall St: Some mix of "Guy Looking At Other Girl" meme involving Databricks, Snowflake, and NVIDIA, where the roles get interchanged frequently.
Get the point?
Henning Holgersen@pacoid nice. So basically, first-mover advantage isn’t much of an advantage?
I was still an economist in 2009, but as far as I can tell Cloudera didn’t have many (any?) real competitors back then? Now, there are already plenty of people in the AI/LLM space plus an infamous “there is no moat” paper.
That's a good point.
Although yes, Cloudera had multiple competitors within the first few years:
* Hortonworks
* MapR
* Greenplum
* Databricks (where I worked)
* Concurrent (where I worked)
@pacoid oh wow, I had completely forgotten about mapR…