ARCHITECHT Daily: Open source, in all its flavors

I'm a firm believer that open source is the future (really going out on a limb there, I know), but al
ARCHITECHT Daily: Open source, in all its flavors
By ARCHITECHT • Issue #55
I’m a firm believer that open source is the future (really going out on a limb there, I know), but also that the business models around it will have to change in order to best serve both users and the companies that develop and sell software. The following news items from Thursday don’t answer the question of how open source will evolve, but I do think they highlight some of the promises and perils:

If you enjoy the newsletter, please help spread the word via Twitter, or however else you see fit.
If you’re interested in sponsoring, please drop me a line.
Hear Jay Kreps on the latest podcast
I haven’t yet put this week’s ARCHITECHT Show episode onto the website, but suffice it to say that Jay Kreps—co-creator of Apache Kafka and co-founder of Confluent—has a lot of smart and interesting things to say about the state of data processing, open source and a lot more. You can listen to it in the following places:
Artificial intelligence
I’ve been linking lately to a lot of stories about human-AI interaction in the workplace, which I think is a more fruitful discussion than simply worrying about AI replacing jobs. Here are two more takes on how that might play out:
This is clearly a pretty smart way to train on individual users’ activity without sending everything to the cloud, which presents in terms of both bandwidth and privacy.
The fairest criticism here probably has to do with the GPUs against which Google compared its Tensor Processing Units. They were older, slower models.
If you want your technology to be useful in the short term, this is among the better ways to do it, IMHO. Industry knows what it wants, and academia has the brains to build it.
Some other companies, as well. There’s also a decent discussion here about the competition between AWS, Google and Microsoft for AI in the cloud.
This is promising in the ways that all unsupervised learning is promising, but don’t go crazy thinking the system is sentient. It’s a neuron that unknowingly latched onto sentiment in the course of predicting the next character in text.
This is very cool—essentially an AI system that resembles a workplace hierarchy. A team of agents works for their own rewards, and an aggregator oversees them and makes the ultimate decision when there’s conflict.
This is still one of the most under-appreciated problems we need to solve as algorithms further infiltrate society.  •  Share
Cloud and infrastructure
There’s a lot to digest in this press release, which thankfully includes a lot of numbers. You can track the growth of China and white boxes pretty easily, the latter of which are driving down public cloud spending.  •  Share
… this breakdown of Intel’s Xeon processors could be up your alley. A lot of the specs are above my head, but even I could take away that Intel knows how to keep those profits coming in.
All things data
This probably could have been included in the discussion of open source above—and relates directly to some stuff I discuss with Jay Kreps in the podcast interview. There are many considerations at play when companies decide to build their own tech, and many reasons why one project might catch on more than others.
I wouldn’t touch Yahoo mail with a 10-foot pole, but this is a clever approach to search results. Like a web search, you display the best matches first, but also give users a chronological list so they can feel they’re not missing anything.
Did you enjoy this issue?
The most interesting news, analysis, blog posts and research in cloud computing, artificial intelligence and software engineering. Delivered daily to your inbox. Curated by Derrick Harris. Check out the Architecht site at
Carefully curated by ARCHITECHT with Revue. If you were forwarded this newsletter and you like it, you can subscribe here. If you don't want these updates anymore, please unsubscribe here.