HNHacker News
TopNewBestAskShowJobs

stuartaxelowen

701 karma · joined March 26, 2015

stuart at sao dot dev
submissionscomments
stuartaxelowen··on New effort to study astronomical ‘wow’ signal after almost 40 years
That's because "interestingness" takes a _lot_ of context.

Humans interpret the low dimensional natural world into much higher dimensional space by turning things into models (cars, cats, science, etc), and often "wow" comes from a combination of many of those high level interpretations.

stuartaxelowen··on Job hopping helped Silicon Valley thrive. Why do other states restrict it?
Seriously, this thread is full of crazy people.
stuartaxelowen··on Facebook quietly deprecated their Python/Ruby libraries for Instagram API
Enterprise clients (the ones they care about) are likely making requests from multiple machines, meaning you have to track API usage per token in a database somewhere.
stuartaxelowen··on Facebook quietly deprecated their Python/Ruby libraries for Instagram API
Because they are shallow layers over their REST API? There's nothing useful in there anyway.
stuartaxelowen··on Intel Xeon E5 v4 Review: Testing Broadwell-EP With Demanding Server Workloads
GPUs are useful for deep learning and a few other easily parallelizable algos, but the majority of open source ML software is still stuck in the CPU.
stuartaxelowen··on On GitHub’s Programming Languages
"relegated"?
stuartaxelowen··on Intel Xeon E5 v4 Review: Testing Broadwell-EP With Demanding Server Workloads
That is definitely false - single core performance still matters a lot in many ML applications, where utilizing many (>8) cores is still difficult.
stuartaxelowen··on Intel Xeon E5 v4 Review: Testing Broadwell-EP With Demanding Server Workloads
Compared to the 5820K i7, these don't seem much better... Is there something I'm missing?

http://ark.intel.com/products/82932/Intel-Core-i7-5820K-Proc...

stuartaxelowen··on Comprehensions in Python the Jedi way
The composition of features is more important than just the individual features - CoffeeScript is a great example of when feature composition goes wrong.
stuartaxelowen··on Python, Machine Learning, and Language Wars (2015)
Knowing other languages is great, but there is very real overhead in learning them. Python is mature enough that you have mature libraries available for pretty much everything.
stuartaxelowen··on Postgresql 9.6: Support parallel aggregation
Because they are building an open source industry leading database. What's not exciting about that!
stuartaxelowen··on Giving Up on TDD

    actual engineering

    prima donnas, hipsters

    have a spine
And then...

    The prima donnas will attack you ad hominum
I'm not for or against TDD, but you've gotta see the irony here.
stuartaxelowen··on Type hints – a mediocre programmer's reaction (2015)
First reply is a very practical response:

> At this point, you may want to just stop caring about the exact type.

> Part of the point of gradual typing is that you can short-cut a lot of

> this. And quite frankly, this isn't really helping anything. Just skip

> it and say that it's Union[Mapping, Iterable, None].

stuartaxelowen··on After fifteen years of downtime, the MetaFilter gopher server is back
Core problem is that creators of value want to be paid for it. The delivery medium doesn't change that.
stuartaxelowen··on Ask HN: How to handle 50GB of transaction data each day? (200GB during peak)
Or... Arrow?! :O

JK it's too early :(

stuartaxelowen··on Ask HN: How to handle 50GB of transaction data each day? (200GB during peak)
IMO investing in a hadoop system constrains you - make your data accessible via S3, Kafka, and most importantly, a good data model, and you can attach the right data processing tool when a problem that requires it comes along.
stuartaxelowen··on Ask HN: How to handle 50GB of transaction data each day? (200GB during peak)
Check out LinkedIn's posts about log processing [0] and Apache Kafka. Handling data as streams of events lets you avoid spikey query-based processing, and helps you scale out horizontally. Partitioning lets you do joins, and you can still add databases as "materialized views" for query-ability. Add Secor to automatically write logs to S3 so you can feel secure in the face of data loss, and use replication of at least 3 in your Kafka topics. Also, start with instrumentation from DataDog or NewRelic from the start - it will show you the performance bottlenecks.

0: https://engineering.linkedin.com/distributed-systems/log-wha...

stuartaxelowen··on Ask HN: How to handle 50GB of transaction data each day? (200GB during peak)
Can confirm, don't use mongo.

We're stuck with it now, and I really wish we had used Postgres for biz data and Cassandra for the high volume non-relational data. They're just materialized views on-top of your event stream anyway.

stuartaxelowen··on Kuhn's Paradox
I'd think this is caused by efficient markets and competition between software companies.

After all, developer productivity is greatly based on tools, and the best tools are rarely free.

stuartaxelowen··on The Elastic Stack: Future of ELK Platform
Does anyone have a good description of the ES query DSL? In spite of the time I've spent with it, I'm consistently unable to write a reasonable query without going to google first, even for basics.
stuartaxelowen··on Records: Python library for making raw SQL queries to Postgres databases
They're essentially the same footprint IIRC.
stuartaxelowen··on Idea Debt
Yes! He was building a single model to describe reality.
stuartaxelowen··on The Rails Doctrine
Good tools don't do this to you.
stuartaxelowen··on Bayes's Theorem: What's the Big Deal?
It's stuck in the reality you define - you can add the last hand, all cards seen so far, and the color of people's jackets to the model if you so please.
stuartaxelowen··on Bayes's Theorem: What's the Big Deal?
Read through most of the article just for "people can abuse priors"? Come on. Anything, used wrongly, can promote superstition and pseudoscience.
stuartaxelowen··on Permission to Fail
From Andre 3000: "Baby boy you're only funky as your last cut, you focus on the past your ass'll be a has what"
stuartaxelowen··on Mark Zuckerberg’s Philanthropy Uses L.L.C. For More Control
Well, 50%? 25%? 5%?
stuartaxelowen··on DspGuru
DSP is the reason I could transition from electrical engineering to machine learning - it's funny how fundamentals pay off.
stuartaxelowen··on 18F's Micro-Purchase Experiment: Why I Bid $1
Soup kitchens and restaurants are not competitors.
stuartaxelowen··on Spinning Up a Spark Cluster on Spot Instances: Step by Step
It only specifies pricing per CPU - how is affected by memory per node? Is that configurable?
← PreviousPage 6 of 7Next →