I’m engineering data in warehouses you’ve never heard of.
Kansas to Colorado
Joined January 2011
- Tweets16K
- Following1.8K
- Followers2.3K
- Likes256K
Michael Gagliano retweeted
We should be stealing a lot more from data engineering when we build AI systems.
DAGs - does this really need to be one giant agent loop?
Lineage - which sources, artifacts, prompts, and model versions produced this?
Change detection - if a source changes, which generated artifacts are now outdated?
Incremental processing - what can we keep instead of rerunning the whole agent?
ETL - what context do we pull in, what does the model do with it, and where do we save the result?
Data engineers have been solving these problems for decades.
Michael Gagliano retweeted
There was once this wonderful place on Twitter before I was involved in AI where I could find content from other people in the data space who were funny without being cruel, educating without marketing, and just enjoyable all around. I’ve made long lasting friendships with these people and learned so much from them. If you never experienced that side of Twitter you’d have a hard time understanding what has since been lost. I rarely see them post anymore, and instead most of the comments I see on posts are so full of mean hatred that it’s just not fun here anymore. I still login out of some nostalgic pride that I have for old twitter but the monetization of engagement has really driven the last bit of joy out of this platform. I really hope someday it returns, there was nothing like it.
This quoted post is unavailable.
Michael Gagliano retweeted
seems like this should be way bigger news🤯
s3 files effectively means that you no longer need to spin up a sandbox vm to give agents access to POSIX tools
you can now point arbitrarily large amounts of compute at s3 to run massively parallel agent swarms on the same filesystem
i'm beyond excited to play with this
Announcing Amazon S3 Files.
The first and only cloud object store with fully-featured, high-performance file system access.
Learn more here. go.aws/4tw17Zg
Michael Gagliano retweeted
Large scale image embedding has gotten so good over the past few years. Example: I was able to accomplish something that @MikeDoesData and I did over the course of months in 2019 in a couple of hours this morning.
Michael Gagliano retweeted
Cannot wait for teams that build their custom JIRA/Workday replacement/custom CRM to one day turn around and ask:
“Why do we have so much internal software that is buggy / has poor UX + we need to maitain?”
Seen this movie well before AI, when Uber built uChat (custom Slack)
Our cracked team just used Software Factory to rebuild and replace Jira in a little more than a month.
We first spent 3.5 weeks planning. This is Software Factory’s superpower.
It allowed our lead PM, Designer and Architect to thoughtfully describe and detail exactly what they wanted. Software Factory then did the heavy lifting in filling in the blanks and allowing our senior tech folks to sharpen the direction of what they wanted.
Then in 2.5 weeks 2.5 junior devs built a replacement.
This will launch as an updated Planner module inside of Software Factory on Tuesday.
It’s beautiful, clean and super useful.
Try it here: 8090.ai
Michael Gagliano retweeted
Why does this guy always talk like a World of Warcraft npc
Michael Gagliano retweeted
I literally did all of these things and took ten AP courses in high school. Sounds like a skill issue.
AP classes should be banned.
Instead of filling kids with useless knowledge we should let them date, play sports, work jobs, and learn about the real world.
The answers to life aren’t in textbooks!
Homework for 2 hours a day is a joke!
Readers added context they thought people might want to know
AP classes are elective & take the place of classes the student would already be taken. There is no guarantee of increased homework, merely more difficult subject matter to count for a college credit. AP students are able to do all the things the post suggests they do instead.
efacademy.org/en/blog/articl…
Michael Gagliano retweeted
not being able to stream json is a skill issue
JSON not having streaming/comments support will be the death of it in the age of AI engineering
@mojombo's toml > yaml because it's sane and streamable, but for whitespace dependency. idk about TOON, not in the data enough, but someone should create a TOONbench and just make labs hillclimb it
Michael Gagliano retweeted
AI model driven evals where you have the model do the evals sounds like some shit from the token man to sell more tokens
Michael Gagliano retweeted
"EBT should only pay for gruel that comes in a metal pail" is the kind of argument that must feel soooooo good to make if you're evil and also have an enormous hole in your frontal lobe