the_data_department (@tddcomics) header
the_data_department (@tddcomics) profile picture

The ordinary adventures of John and his team in the department of cloud-based data-driven AI-enabled magic https://xitter.gq/t.co/CmIV6ltPXZ…

Joined December 2020
Photo by the_data_department (@tddcomics)Photo by the_data_department (@tddcomics)Photo by the_data_department (@tddcomics)Photo by the_data_department (@tddcomics)Photo by the_data_department (@tddcomics)Photo by the_data_department (@tddcomics)Photo by the_data_department (@tddcomics)Photo by the_data_department (@tddcomics)Photo by the_data_department (@tddcomics)Photo by the_data_department (@tddcomics)
the_data_department retweeted
How many users does Twitter really have❓ Depending on where you look, Twitter has 300-400 million users. @elonmusk says that Twitter claims >95% of these are "real, unique humans"... 🤖 ...but Elon doesn't believe this. 🤷‍♂️ Here is my challenge/offer to @twitter and @paraga:
2
23
2
44
the_data_department retweeted
There are now 100 (one hundred!) code examples at keras.io/examples, covering essential workflows across computer vision, NLP, generative learning, RL, etc. Created by the Keras community. Focused on readability/clarity, reusability, and best practices.
1
82
6
451
the_data_department retweeted
🧘‍♂️ MoveNet is now on mobile! The state-of-the-art pose estimation model finally comes to mobile via TensorFlow Lite with CPU acceleration enabled. Now you can run pose estimation in realtime on modern smartphones. Learn more 👉 goo.gle/3m494pn
12
146
13
572
the_data_department retweeted
It’s certainly enlightening to see a cartoon dedicated to #AI and #datascience , Make sure to follow @tddcomics for more #comics #cartoons Thanks Adrian @lowrank_adrian for suggesting this gem. #66daysofdata #100daysofcode
The menu card for martinis
1
5
11
the_data_department retweeted
I have been "spuddling" with our latest episode which we are SO excited to share!
Word of the day is ‘spuddle’ (17th century): to work ineffectively; to be extremely busy whilst achieving absolutely nothing.
2
3
the_data_department retweeted
"The best time to rob a bank was 20 years ago. The second best time is today." ~ Ancient Proverb
1
1
the_data_department retweeted
CHALLENGE - THE BLUEPRINTS: Can you find a route which takes no more than 217 steps which goes: Entrance --> Reception Cabinet --> Employee Records --> Foreign Exchange --> Business Analysts --> Vault
1
1
the_data_department retweeted
So the big question is whether the deep learning or the my little pony community is bigger..
1
2
the_data_department retweeted
Shaking while I read this. The past 4 years at Google brought trauma and near-constant drama for me. Still, @timnitGebru and I cranked out papers and made a new org. Still, that wasn't good enough to avoid being thrown out and internationally disparaged. bloomberg.com/news/articles/…
6
98
9
421
the_data_department retweeted
Design patterns in machine learning by Ágoston Török towardsdatascience.com/desig…
1
1
the_data_department retweeted
Design patterns in machine learning by Ágoston Török buff.ly/3ehneOa
8
1
16
the_data_department retweeted
Businesses making revenue projections (ab)using mathematics
1
2
the_data_department retweeted
NEW EPISODE "How 2 create a dating app". How do dating apps like @Tinder and @hinge work? We discuss here (available literally everywhere): ⚓ anchor.fm/how2robabank/episo…
3
6
the_data_department retweeted
A new self-supervised learning framework for model selection and hyperparameter tuning for large-scale forecasting provides accurate forecasts with less computational time and resources, and without sacrificing accuracy. Learn more: ai.facebook.com/blog/large-s…
7
160
7
683
the_data_department retweeted
Dad joke lol
12
417
20
1,434
the_data_department retweeted
Here's a piece of information that will send a chill down the spine of anyone who's ever designed a database schema: Our new house that we just moved into... has two zip codes!
230
631
276
4,885
the_data_department retweeted
We're introducing an optimizer for deep learning, MADGRAD. This method matches or exceeds the performance of the Adam optimizer across a varied set of realistic large-scale deep learning training problems. github.com/facebookresearch/…
23
482
66
2,014