HN user

mlerner

2,073 karma

Software Engineer and computer security explorer.

Posts302
Comments35
View on HN
www.usenix.org 11mo ago

Graham: Synchronizing Clocks by Leveraging Local Clock Properties (2022) [pdf]

mlerner
59pts13
www.usenix.org 1y ago

Autothrottle: Resource Management for SLO-Targeted Microservices

mlerner
22pts1
www.micahlerner.com 1y ago

Resiliency at Scale: Managing Google's TPUv4 Machine Learning Supercomputer

mlerner
1pts0
www.leagueoflegends.com 1y ago

How the anti-cheat is anti-cheating so far

mlerner
3pts0
www.micahlerner.com 2y ago

ServiceRouter: Hyperscale and Minimal Cost Service Mesh at Meta

mlerner
1pts0
newsletter.micahlerner.com 2y ago

A Cloud-Scale Characterization of Remote Procedure Calls

mlerner
2pts0
www.micahlerner.com 2y ago

A Cloud-Scale Characterization of Google's Remote Procedure Calls

mlerner
1pts0
www.micahlerner.com 2y ago

Gemini, Amazon's system for fast failure recovery in distributed model training

mlerner
2pts0
www.micahlerner.com 2y ago

Defcon: Preventing overload with graceful feature degradation (2023)

mlerner
237pts95
www.cidrdb.org 2y ago

MotherDuck: DuckDB in the Cloud and in the Client [pdf]

mlerner
5pts0
www.micahlerner.com 2y ago

Gemini: Fast Failure Recovery in Distributed Training with In-Memory Checkpoints

mlerner
3pts0
newsletter.micahlerner.com 2y ago

Gemini: Fast Failure Recovery in Distributed Training with In-Memory Checkpoints

mlerner
4pts0
newsletter.micahlerner.com 2y ago

XFaaS: Hyperscale and Low Cost Serverless Functions at Meta

mlerner
3pts0
newsletter.micahlerner.com 2y ago

XFaaS: Hyperscale and Low Cost Serverless Functions at Meta

mlerner
3pts1
www.micahlerner.com 2y ago

Blueprint: A Toolchain for Highly-Reconfigurable Microservice Applications

mlerner
2pts0
www.cs.rice.edu 2y ago

Gemini: Fast Failure Recovery in Distributed Training with In-Memory Checkpoints [pdf]

mlerner
50pts13
www.cis.upenn.edu 2y ago

XFaaS: Hyperscale and Low Cost Serverless Functions at Meta [pdf]

mlerner
4pts0
newsletter.micahlerner.com 2y ago

Efficient Memory Management for Large Language Model Serving with PagedAttention

mlerner
3pts0
newsletter.micahlerner.com 2y ago

Efficient Memory Management for Large Language Model Serving with PagedAttention

mlerner
1pts0
www.micahlerner.com 2y ago

Blueprint: A Toolchain for Highly-Reconfigurable Microservice Applications

mlerner
1pts0
www.micahlerner.com 2y ago

Blueprint: A Toolchain for Highly-Reconfigurable Microservice Applications

mlerner
5pts2
www.micahlerner.com 2y ago

Defcon: Preventing Overload with Graceful Feature Degradation

mlerner
4pts1
www.micahlerner.com 2y ago

Defcon: Preventing Overload with Graceful Feature Degradation

mlerner
4pts0
news.ycombinator.com 3y ago

Ask HN: Streaming reading academic CS papers?

mlerner
2pts0
www.micahlerner.com 3y ago

Towards an adaptable systems architecture for memory tiering at warehouse-scale

mlerner
28pts4
www.micahlerner.com 3y ago

Sundial: Fault-Tolerant Clock Synchronization for Datacenters

mlerner
2pts1
www.usenix.org 3y ago

Empowering Azure Storage with RDMA

mlerner
2pts0
www.micahlerner.com 3y ago

Sundial: Fault-Tolerant Clock Synchronization for Datacenters

mlerner
3pts0
www.micahlerner.com 3y ago

Automatic Reliability Testing for Cluster Management Controllers

mlerner
1pts0
www.micahlerner.com 3y ago

TelaMalloc: Efficient On-Chip Memory Allocation for Production ML Accelerators

mlerner
52pts1

https://www.micahlerner.com

I write about new and foundational academic CS research - writing is a way for me to learn and share with others along the way.

Some of my most popular writing is:

- FoundationDB: A Distributed Unbundled Transactional Key Value Store (https://news.ycombinator.com/item?id=28740497)

- Monarch: Google’s Planet-Scale In-Memory Time Series Database (https://news.ycombinator.com/item?id=31379383)

- Ray: A Distributed Framework for Emerging AI Applications (https://news.ycombinator.com/item?id=27730807)

Author here - thank you for pointing that out! The context in the comment you linked to is also helpful.

In your opinion, would it be correct to remove "uniquely" from the first sentence? Maybe that would help clarify the issue you pointed out. As far as I understand, the CIDs are still unique when referring to _metadata_, and can be used to "unambiguously fetch content from its peers".

This paper is from a conference a few months ago (SOSP), it's possible you read a pre-print (or maybe the CacheLib paper [1]).

Facebook publishes lots of systems research! FWIW I also write a fair bit about the other systems research from Microsoft Research, etc - what you are seeing in my posting history is me sharing my writing a few times (as permitted by HN rules)

[1] https://www.usenix.org/conference/osdi20/presentation/berg

Bitwise Asset Management | San Francisco | Full-time, Onsite | https://jobs.lever.co/bitwiseinvestments

Seed funding from Naval Ravikant, Khosla Ventures, General Catalyst, Avichal Garg (Part-time Partner at Y Combinator)

- Front-End Engineer (React, Node) - https://jobs.lever.co/bitwiseinvestments/f4ccc812-edb6-4099-....

- Full-stack Engineer (Node, Scala, Postgres) - https://jobs.lever.co/bitwiseinvestments/43bb8ae2-2ee9-44e9-....

We are a San Francisco-based cryptocurrency asset manager founded in 2017. Last year we introduced the first cryptocurrency index fund — the Bitwise HOLD 10 Private Index Fund. It holds the 10 largest cryptoassets that cover about 80% of the market.

The firm has a software team with backgrounds across Google, Facebook, Wealthfront, and military software security. Bitwise is backed by individual and institutional investors who backed and built: PayPal, Square, Stripe, Wealthfront, Coinbase, MetaStable, Palantir, and others.

The Bitwise team is a tight knit team. We're growing quickly and looking for people excited about what we're working on: https://jobs.lever.co/bitwiseinvestments

Bitwise Asset Management | San Francisco | Full-time, Onsite | https://jobs.lever.co/bitwiseinvestments

Seed funding from Naval Ravikant, Khosla Ventures, General Catalyst, Avichal Garg (Part-time Partner at Y Combinator)

- Front-End Engineer (React + Node) - https://jobs.lever.co/bitwiseinvestments/f4ccc812-edb6-4099-...

- Full-stack Engineer (Node, Scala, Postgres) - https://jobs.lever.co/bitwiseinvestments/43bb8ae2-2ee9-44e9-...

We are a San Francisco-based cryptocurrency asset manager founded in 2017. Last year we introduced the first cryptocurrency index fund — the Bitwise HOLD 10 Private Index Fund. It holds the 10 largest cryptoassets that cover about 80% of the market.

The firm has a software team with backgrounds across Google, Facebook, Wealthfront, and military software security. Bitwise is backed by individual and institutional investors who backed and built: PayPal, Square, Stripe, Wealthfront, Coinbase, MetaStable, Palantir, and others.

The Bitwise team is a tight knit team. We're growing quickly and looking for people excited about what we're working on: https://jobs.lever.co/bitwiseinvestments

I think rebalancing with Coinbase's selection of coins is pretty easy. When you have a more complex basket (for context, I work at Bitwise Investments, which runs the Bitwise HOLD10 Index), it is more difficult to decide what goes into the basket.

For example, several coins (like Neo and Ripple) have supplies that grow and are centrally controlled, but many coins have planned inflation schedules. We know that the supply of many of the large-cap coins is going to grow over the next couple of years, and that needs to be taken into consideration when valuing them.

To explain why that is important: if people buy a coin at a certain price _knowing_ that a certain amount of inflation is going to happen, that means investors think that the market cap of the coin is actually much more (think of this like Discounted Cash Flow). Restated, if people buy these coins knowing that the supply is actually going up, that means that they think that the value of the coin is actually much higher than the current market cap.

If you want to learn more about indexing methodologies for cryptocurrencies, you should check out our website: https://www.bitwiseinvestments.com/index

The Bitwise HOLD10 (where I work) handles a lot of the hard work around rebalancing and taxes.

You should check out the Bitwise HOLD10 (where I work). We are backed by Naval Ravikant (Cofounder of AngelList and CoinList. Partner at Metastable), Avichal Garg (Part-time Partner at Y Combinator. Previously Director at Facebook), and Keith Rabois (former COO at Square: https://www.bitwiseinvestments.com/about