Mike Robbins

@StedmanBlake @alfongj Heii On-Call heiioncall.com @HeiiOnCall — please let me know what you think!

@alfongj .@compumike! what's your newco website again?

I can't believe Pager Duty is the most used tool for on-call management. It's got horrible UX, and it's extremely overpriced. Is there any alternative that just has the ability to page ppl in their phone number with a predetermined schedule? Easy to use and <$500 a seat?

@pauld_fgc I wrote IdleGC shard github.com/compumike/idle… to combat @CrystalLanguage slow memory growth. Runs garbage collection when not servicing requests. Running in production on totalrealreturns.com Lmk if it works for you.

API & Platform Results: Rails: @Railway wins. Render wakeup boot extremely slow (~1m). Flyio cannot run Rails for free. Crystal: @Render wins. 5-10s wakeup boot time. Railway & Fly keep growing memory infinitely. Node: @flydotio + Railway both win. Very slow boot on Render.

Number one on Hacker News 🎆💥🙂

Today I'm launching totalrealreturns.com - Inflation-adjusted stock charts.

... doing the deep dive today puts new tools in your toolbox, and new tools expand your reach for future projects. The expected payback period is surprisingly short. But this doesn't fit with the conventional sprint/ticket model of software development teams.

My first Kubernetes controller: hairpin-proxy:v0.1.0 github.com/compumike/hair… solves a kubernetes/nginx-ingress/cert-manager problem reported on GitHub issues for years. It's fun to dive one or two layers deeper and figure out how the magic works so I can fix it. More importantly,

@rubiety Thinking carefully & analytically is hard work! Does society appropriately reward those who do that hard work? Does the reward encourage others to think critically too? Do those suffering from a lack of analytical thinking have a viable alternative?

I increasingly believe that a lot of our problems are rooted in vigorous aversion to even *trying* to think carefully & analytically. Congruent with cognitive psychology research, but I still struggle to accept it.

@AaronLCannon Good point! And that means you might be able to relocate offices out of expensive core downtowns as well. (Public transit / bike-to-work fans might be unhappy.) Or use shared space that 5 companies rent out on different days of the week.

@compumike True, but at one day a week, I’d be willing to drive pretty far. So maybe a mixed model extends the radius of living to a city center

I agree with the skepticism. we shouldn’t assume full-remote is our future. But a ton of people (probably not CEOs/execs/founders) would prefer remote. I believe the future is partial-remote. A complex system of in-person and remote schedules, with small offices.

@AaronLCannon Definitely. They probably have a relatively weak revenue-driven motive: increasing Meet adoption probably only slightly reduces G Suite customer churn. Most of the customer lock-in is still elsewhere (documents and emails). But Zoom/Quip/etc threaten that integration.

Im sure a PM at google meet has been trying to increase user adoption for years. It only took a pandemic to convince the Gmail execs to carve out massive real estate to siphon traffic to meet. That PM must be stoked.

"For a good conversation about ideas both sides should share a single motive — finding the truth. I'd like to believe this is the only motive anyone has when discussing ideas but it's not the case." harj.posthaven.com/conversations-…

@spiegart @alexandrosM @eladgil I modeled this as a closed-loop feedback system to quantitatively show the shape of our range of long-term strategy choices, from now until immunity (vaccine or not). circuitlab.com/blog/2020/05/2…

The good news is that behavior seems heterogeneous. Heterogeneous behavior (both geographically, and between more-careful and less-careful households) leads to better overall outcomes than the homogenous models suggest. But those rewards are distributed heterogeneously as well.

A month ago, I published my analysis of the pandemic as a feedback control system twitter.com/compumike/stat…. (I didn't model shorter-term oscillations because I was focused on long-term strategy outcomes.) The bad news is that strategically, the US does seem to be targeting R=1.

Feedback systems pop quiz! What happens when you combine: (1) a process with open-loop exponential growth behavior (2) a feedback signal with lots of delay and noise (3) a not-very-smart controller that's targeting R=1? Answer: oscillations and (unfortunately lethal) overshoot.

@gandibar You can help your customers with the CA root expiration incident status.gandi.net/incidents/026k… by distributing an updated GandiStandardSSLCA2.pem file that replaces the 2nd cert with AAA. I created one explained here: news.ycombinator.com/item?id=233649… Happy to provide the file.

@youyanggu @youyanggu Short-term forecasting is important for risk assessment and policy feedback, but I really like your "what if" scenarios. It would be great to start a public conversation about long-term strategy, and that strategy will probably be driven by health & economic modeling.

@youyanggu Small variations in transmission lead to incredible reductions in human suffering. We're still in the horizontal region of the deaths-vs-duration strategy curve. If we go a bit further we can reduce economic disruption simultaneously. (slide from youtube.com/watch?v=NuNrkG…)

I ran a simulation on what would happen if just 20% of infected individuals reduced their own transmission by 25% (such as by self-quarantining immediately after showing symptoms). Deaths would be 30% lower by today, and up to 50% lower by September. covid19-projections.com/us-self-quaran…

"Surprising COVID-19 Strategy: Reduce Economic Damage and Deaths Simultaneously" youtube.com/watch?v=NuNrkG… I analyze pandemic as a feedback system and show the policy frontier is concave down. Counterintuitive! There is no tradeoff between saving lives & reducing economic damage.

@AaronLCannon @triple_byte bucket_low = 20*floor(actual_percentile_int / 20.0); bucket_high = bucket_low + 20; Ship it! 🚀 (Dear engineering team: I doubt this is actually the issue...)

@compumike @triple_byte Sometimes bugs are good 🐞

"You scored between the 100th and 120th percentile" -- looks like I broke @triple_byte's cool new software engineer skills report! Please don't fix, I'm happy to be in that bracket :)