This Work Chronicles comic rings so true!

(Also, I think Work Chronicles is my favourite comic these days).

“If you’re a marketer, consider sending a single email during an off-period that asks people to update those [holiday-related email] preferences all at once” - this is excellent advice.

I vividly remember scanning 1000+ books into Delicious Library, which was vastly faster than entering them all completely manually, but something like this LLM based book scanning would be so much faster, particularly if it was wrapped into a real tool that did full lookups, etc (like Delicious Library did).

The possibility for hallucinations doesn’t worry me much - quickly verifying the results afterwards (plus the second layer of a lookup) would mostly solve that. I’d be more worried about not noticing something was missed, but good tool UX could probably solve that.

I agree with all of this (except the “dollar store” reference that goes over my non-US head): any time people are paid to work on open-source it’s a win. I saw somewhere the argument that once there is sufficient funding we can start being more critical, which is an appropriate approach.

Interesting post on falsehoods junior devs believe about seniors. I agree with most, but not:

• Having all the answers: the more senior, the more you ought to not necessarily know the answers (although there is a bunch of that) but know how to figure out the answers fairly efficiently
• Time to relax: I suspect this is coming from a US perspective - time to relax should be constant at all levels; it’s basic health and cost effectiveness

Nice intro post to Bloom filters.

Applies equally to work and older schooling: Work Chronicles

This enterprise budgeting cartoon is so true.

Interesting read on random number generator seeding in C++ but broadly applicable. I’m glad that I work in a language where someone else takes care of getting this right!

A very thoughtful look at the hyperscaler clouds and how Europe can and should react. We really need people in a position to take action to be thinking about this in the AoNZ context as well, or maybe a Pacific context (not APAC - Asia is its own thing).

Very detailed post on Sentry’s removal of all cookies and user tracking - helps to have a reasonable familiarity with mar[keting]tech to follow what happened and the advice.

This record type proposal looks nice, but I don’t see much benefit over dataclasses (KW_ONLY is ugly, but also uncommon I think). At least it would be a stdlib addition, not a language change.

Arguments against using @property() in Python - other than to replace an attribute.

I agree this is overused a lot. The other place I see it used where it seems reasonable because there’s no obvious alternative is to mark an attribute as read-only.

Summary of None-aware operators being considered for Python. For example, “obj?.attr” which would be None if obj is None, rather than an AttributeError being raised. I hate the idea of this being added - I hope it never does. Partly for some of the reasons in the article and partly because it’s unnecessary complexity.

Interactive page for figuring out how to undo or fix things in git. The “take a backup first” is the best advice. To be honest, I tend to “give up and go to a known good state” a bunch of the time.

I had never heard of git notes before. It seems like they could indeed be really useful if they were exposed more.

This is a really insightful internal look at Sentry’s pricing and scaling in the early days. I think for most of this period we were using Sentry but doing the self-hosted route (paying Sentry $0), but we probably moved to the hosted product towards the end of this time. I remember a period where every billing cycle there was a new pricing plan.

Great walkthrough of Git terms that people find (and are!) comfusing. Most of the time I don’t need to use the majority of these, but it helps to have at least seen explanationions before.

Pretty nasty supply chain attack when using GitHub self-hosted runners.

I’ve definitely had this happen to me a few times 😂: open.substack.com/pub/workc…

Details of a new test vectoriser from Google. If I was still in this business I would definitely be trying this out already!

I’m a bit hesitant given that it seems to me that Gmail’s filtering has significantly deteriorated over the last year or so, and that’s the period they have used this. Their benchmarks show improvements but internal benchmarking is hard.

Interesting anti FaaS post.

Interesting story on digging into a performance issue that goes super deep.

Interesting analysis on typing use by Python core devs - the methodology is a bit weak (which the author admits) but interesting nonetheless.

Interesting story about learning from mistakes. However, if one person can make a $500M mistake, you have a process problem, not a person one.