Yup, I had an issue filed against an open source project I work on. Was a crazy weird crash.

The reporter actually spent the effort to track it down, turns out it _was_ a Go compiler bug. (https://github.com/golang/go/issues/20427)

Also, it's not always about vulnerabilities directly, but how well / fast things are patched.

Cheap pans. Thin-walled that doesn't distribute heat evenly. So hot spots burn the pasta to the walls.

Secondly, not stirring at all. Boiling will do some circulation, but you have to keep some amount of stirring to prevent a small sticking turn into a burned to the pan problem.

> "none, loyalty goes only one way and trust me when I tell you that no-one is safe"

Yup, I know someone who had been there since 2006 that was laid off. Not even high scores on Perf will save you.

I have no idea what methodology went into it. That's a whole postmortem on its own.

Many "blind" people are not 100% blind. Visual impairment is a sliding scale.

https://www.youtube.com/watch?v=cdPymLgfXSY

Most of the water use is in evaporative cooling. Not really a way to re-use water that's now in the air.

This, in part, is how Google gets an average PUE of 1.1. Which is _really_ good compared to typical.

In Finland, time of day electric use meters have existed for something like 20 years. Someone I know who lived north of the arctic circle had a home heating system that would heat up an energy storage box (I forget if it was an oil tank or simply stone) at night and then cycle the heat out to the house during the daytime.

The big thing I see with a ton of arguments about changing the way we produce and consume energy assume that things will change overnight and the whole system will come crashing down because it's not ready.

Perfect is the enemy of good. Things aren't going to happen overnight, there will be a transition period.

Electric vehicle transition is a good example of this. * OMG, the grid will crash with all the charging. * OMG, You won't be able to charge it <because reasons>. * OMG, it doesn't solve <random gas car use case here>.

The transition to electric vehicles is coming, but slowly. It basically started at 0 in 2010 and a decade later it's still only at 2-5% in the US and something like 5-10% in the EU, China, etc. The ball is just getting rolling.

The grid hasn't crashed, charging is getting deployed where there's demand, and not every <random gas car use case here> needs to be solved now, or even in the mid term future.

Google ended up with a bunch of chat apps because the culture promotes building new over keeping things running.

The first chat was basically an XMPP service. It was decent, it federated outside of Google, it was fully functional. It even supported group channels (XMPP conferences) internally. But I don't think that was ever exposed to the public.

Then Hangouts was created. It was, per usual at Gooogle, a ground-up rewrite. IIRC not the same team. So they spent at least a couple years playing catch up to get feature parity with the XMPP chat. Worse, Hangouts was one of the first services to suffer from strict team-created "Personas" design philosophy. Any time anyone would complain about a feature, or miss-feature, it was flatly ignored because "You're not one of our Personas". It took years of complaints to get them to change their minds.

By the time Hangouts was good enough to fully replace the previous service it was now boring and people left the team for other new projects. Because maintenance won't get you promoted.

The other random chat services were basically experiential toys.

Now we have Meet, which is is likely another case of "Hangouts is unmaintainable tech, we need to re-write it". Years of getting up to feature parity. And miss-features that won't get fixed.

XMPP compatibility was amazing. I remember using one chat client to accept all the different chat services, including google one.

The first sign of trouble to me was when google killed xmpp support. Prior to that everyone just logged into instant messaging on whatever device.

Every decision since then has been pro-ad network technical competition.

Yea, it was great. But it was more lack of maintenance than anything. The original chat devs were passionate about open federation. The new devs were not.

Add to that it was a big source of spam that nobody wanted to deal with.

Apparently the team tried really hard to get more companies on board with opening up federation. But I think the only "major" service that did was what was left of AOL.

This is spot on. The issue was a culture of new over improve.

It had the upside of making some things more integrated. MR and issues and CI work much better together than GitHub or atlassian.

But it had the down side of expanding into feature categories half baked.

I only managed to stay for 4 years before I couldn't stand it. But I'm more an SRE than a feature developers.

I left for a job that has "burn tech debt" as a key goal.

Ugh, speedbumps are so dangerous.

Sure, they make you slow down for them. But drivers now want to "make up" for that slow down. So they drive even faster between them.