Micah Learns
@micahlerner
Followers
4K
Following
11K
Media
136
Statuses
1K
Google Geo SRE 🗺️📍, prev @mapbox, @strava. Following my curiosity and writing about CS/AI/Systems research - https://t.co/aRJWZub62X.
San Francisco, CA
Joined February 2011
RT @googlemaps: #sponsored When @georainbolt began exploring the world on Street View, he would've never guessed that one day he'd be in th….
0
64
0
RT @sundarpichai: 1/ Q2 was a standout quarter with robust growth across the company. As we showed at I/O, we’re leading at the frontier of….
0
196
0
RT @offlinemark: week 66, streaming kernel dev right now. topic: networking. last week we finally got RX working and received network data….
0
19
0
RT @maheshb: [new blog post] On the two types of 10x engineers: I entirely blame @muratdemirbas for encouraging me to write this up. http….
maheshba.bitbucket.io
“It used to be said there were two kinds of chairs to go with two kinds of Ministers: one sort that folds up instantly, the other sort goes round and round in circles.” — Bernard Woolley, Yes...
0
23
0
Excited to announce that I'll be presenting "Product Reliability in Google Maps" at this year's "@ Scale: Systems & Reliability" conference. I will be speaking in person at the Meta campus in Menlo Park. You can join me in person, or attend virtually! . Here's the link to
0
1
9
I love the Monarch paper! Great insights :).
Distributed system design is all about trade-offs. Monarch, Google's planet-scale time-series database, prioritizes high availability and partition tolerance over strict consistency. Because it's used for monitoring and alerting — fresh and timely (even if partial) data is far.
0
0
1
RT @JeffDean: Someone just reminded me of this lecture I gave in 2009 that described the evolution of Google Search from 1999 to 2009. Peo….
0
261
0
RT @offlinemark: week 63, streaming kernel dev right now. topic: networking. last week we finished the transmit side of the networking labs….
0
35
0
Hello new followers and thanks to @eatonphil for the signal boost! . Appreciate the inclusion among an esteemed list of authors :).
2
6
89
RT @himanshustwts: just published! ⚡️. Tired of the Llama 4 hype? .Give this a quick read; won't take much of your time, and you'll see thi….
0
24
0
RT @JeffDean: I'm delighted to have joined my good friend and colleague @NoamShazeer for a 2+hour conversation with @dwarkesh_sp about a wi….
0
223
0
The paper cites one technique that I found interesting, "contextual bandits", which is a lightweight form of reinforcement learning (afaiu), citing this paper:
0
0
3
The next paper I'm going to read and write about is "Autothrottle: A Practical Bi-Level Approach to.Resource Management for SLO-Targeted Microservices". Autothrottle is a practical resource management framework for microservice-based cloud applications that aims to minimize CPU
1
1
13
RT @micahlerner: I just published my paper review of "Resiliency at Scale: Managing Google’s TPUv4 Machine Learning Supercomputer": https:/….
newsletter.micahlerner.com
Resiliency at Scale: Managing Google’s TPUv4 Machine Learning Supercomputer
0
6
0
I just published my paper review of "Resiliency at Scale: Managing Google’s TPUv4 Machine Learning Supercomputer": The paper talks about the next generation of Google's TPU infrastructure. This shift is by an increase in size size and availability,.
newsletter.micahlerner.com
Resiliency at Scale: Managing Google’s TPUv4 Machine Learning Supercomputer
1
6
38
As shifting from reading systems papers to reading AI papers, I found this guide to gaining an intution for the map helpful (I believe the author is @yacineaxya on X :)):
1
0
6
RT @cneuralnetwork: releasing the DeepSeek R1 blog, which explains the whole paper in great detail, not excluding any math, but anyone with….
0
573
0