Redundancy and diversification is always a good idea in general, but I find the recent clamor for supply chain diversification due to COVID-19 bit of a knee-jerk reaction, especially considering the virus is now global. Where to diversify to is not an easy question to answer, compare to say a few weeks ago, when the virus was mostly just infecting China.
Most supply chain analysis focus on cost/location/scale (this one too), but ignore the manufacturing and engineering skills that are accumulated from experience, which are much harder to replace. Countries level up to make themselves more valuable. And the cost v quality tradeoff is really hard to figure out.
Apple's failing foray into India's supply chain a few years ago is a case in point [0]. I've written about this too in an India and China comparative context when it comes to Apple [1]. Tim Cook said recently on TV that he's not leaving China anytime soon [2].
Not everyone is Apple. How do you think about the tradeoff depends on your business model. But rushing to "rethink" something as complex as global supply chain, because of an unrelated, intervening event, however severe, seems imprudent to me.
Author of the post here. Thank you so much for reading and for your comment! I learned a lot.
100% agree with you that documentation and API stability is a topic worth discussing more, so I might write a separate post on that in the future. I'm personally a bit obsessed with good documentation (most vendors big and small don't invest enough time and resources on documentation). It gets into both the reliability and usability of the cloud.
Since it's Day 2, def appears to be an architectural scalability issue, not some bugs. Quick scan of their engineering blog shows the following stack:
-Zookeeper, Kafka, Spark, Airflow, Faust (Kafka Streams in Python)
-Kubernetes (was/is SALT + Terraform)
-AWS Aurora, PG, ES, Influx
-Presto via AWS Athena, Redshift
-homegrown solutions for a data lake & managing k8s microservices
Fwiw, their team does appear to have a rigorous internal post-mortem process (SEV reviews) [0] drawn from industry best practices.
Let's hope a public one will be made when issues are resolved. (FINRA will be asking for one anyway.)
Not piling on. I genuinely feel for all the infra engineers staying up all day/night fixing this.
IIRC there is at least one Silicon Valley example of this already: Zappos pre Amazon acquisition, debt from Wells Fargo. It was forced upon Zappos b/c VC fundraising dried up, not a deliberate choice, but ended up working out well b/c its GMV was both growing and becoming predictable.
I believe this article is focused on a different type of debt - "recurring revenue securitization" (focused on companies w/ a SaSS business model).
This type of financial instrument is different from a traditional venture / bank debt instrument in a couple ways:
1. it can be more favorable to startups by making payments a % of revenue instead of a fixed amount (ala ISA's)
2. the lender can earn higher interest by "securitizing" (e.g. pooling together) multiple loans and selling them. this allows the lender to move some debt off the balance sheet and in turn deploy the cash for higher yield
3. over time, the lender could provide more favorable terms by creating more accurate risk models by ingesting data from sources like Stripe and Shopify across companies and then building proprietary data sets to manage default risk. I believe Clearbanc does something similar today.
I'll share mine to get things started: my biggest challenge is diagnosing and finding the one thing that needs 10x improvement to generate more adoption: better documentation? more tutorials? produce benchmarks? wider language coverage?
This self advocacy is encouraging and long overdue IMO.
The Chinese colloquial term for developer is "码农". Its literal English translation is "code peasants" -- not the most flattering or respectful way to call software engineers.
I've recently heard horror stories, where 9-9-6 is no longer enough inside one of the Chinese tech giants, and 10-10-7 is expected (10am-10pm, 7 days/week)
I hope this brings about some positive change and won't just get blocked by the authority.
That was my initial thought as well seeing as you can freely and anonymously discuss. But perhaps it's in the government's interest to have the people openly criticize these capitalist practices that have a negative impact on public health?
Most supply chain analysis focus on cost/location/scale (this one too), but ignore the manufacturing and engineering skills that are accumulated from experience, which are much harder to replace. Countries level up to make themselves more valuable. And the cost v quality tradeoff is really hard to figure out.
Apple's failing foray into India's supply chain a few years ago is a case in point [0]. I've written about this too in an India and China comparative context when it comes to Apple [1]. Tim Cook said recently on TV that he's not leaving China anytime soon [2].
Not everyone is Apple. How do you think about the tradeoff depends on your business model. But rushing to "rethink" something as complex as global supply chain, because of an unrelated, intervening event, however severe, seems imprudent to me.
[0] https://www.theinformation.com/articles/inside-apples-search...
[1] https://interconnected.blog/leaving-china-supply-chain-a-pip...
[2] https://www.youtube.com/watch?v=rsDU-QsfkX4&feature=emb_logo