Polymath Engineer Weekly #47
More and more to learn
Links of the week
What people are missing in the analysis is that LLMs, and the products like ChatGPT built on top of them, aren’t a disruptive innovation – they don’t underperform on things that matter to mainstream computers and make it up by serving a small new or low-end niche well – but a superior product across nearly all dimensions.
So Google isn’t facing the Innovator’s Dilemma; it’s just so deeply Positioned in search that it’s been a sitting duck for the first superior technology or business model strong enough to take it on.
You probably will not find this word in any service level agreement. Cloud providers promise to make "commercially reasonable efforts" to make their services available with a specified uptime percentage. If they don't meet this in a specific month, they agree (typically upon your explicit request) to deduct a pro-rated amount from your bill, corresponding to the duration of the outage or a fixed percentage based on the actual range of uptime they delivered. The cloud providers do not reimburse you for your actual losses
Further, we decided that the right long-term goal for Mojo is to provide a superset of Python (i.e. be compatible with existing programs) and to embrace the CPython immediately for long-tail ecosystem enablement. To a Python programmer, we expect and hope that Mojo will be immediately familiar, while also providing new tools for developing systems-level code that enable you to do things that Python falls back to C and C++ for. We aren’t trying to convince the world that “static is good” or “dynamic is good” - our belief is that both are good when used for the right applications, and that the language should enable the programmer to make the call.
As cost control becomes paramount, some may question the philosophy that has been at the heart of the modern approach to data management since the Hadoop days – keep all your data, dump it all somewhere (a data lake, lakehouse or warehouse) and figure out what to do with it later. This approach led to the rise of data warehouses, the centerpiece of the MDS, but it has turned out to be expensive, and not always that useful (read this good piece: “Big Data is Dead”).
Administrators are sometimes hard pressed to defend security best practices without examples that demonstrate the security implications of risky configurations. I hope the examples laid out in this post and the manifests contained in the Bad Pods repository help you enforce the principle of least privilege when it comes to Kubernetes pod creation in your organization.
Science training doesn't teach us where hypotheses come from. It assumes they're already there. We spend forever talking about how to run experiments in a valid way and to not bias our observations and to make our results repeatable, but we spend almost no time talking about why we test the things we test in the first place. That's because the answer is embarrassing: nobody knows. The best testable hypotheses come to you in the shower or in a dream or when your grad students are drunk at the bar commiserating with their friends about the tedious lab experiments you assigned because they are so straightforward they don't warrant your attention.
Book of the Week
Do you have any more links our community should read? Feel free to post them on the comments.
Have a nice week. 😉