rossjudson

- Liability - What can you really fit in HW4? - No right answers in many situations. - But mostly liability. It's pretty good.

You're highly critical. What would you be doing differently? So far Tridge has elected to:

- generally decide to fix security issues over preserving compatibility - rewritten an aging test suite in what appears to be a highly responsible way - brought on additional qualified developers to help with the workload

Not bad for a guy who's retired.

You care enough to complain on HN. You could be a part of the solution.

What were you going to do differently, specifically?

Yeah, we definitely need to make sure that we take the considerations of the mob into account.

The person owning the project is using the master branch in the way he sees fit.

Incidentally, there is no amount of communicating "correctly" that quells a mob. There's a Venn diagram of concerns, and those with concerns not being met will generate (now infinite) outrage.

I recommend "How to measure anything in cybersecurity risk". Really interesting read about putting actual value on security.

I'm truly hopeful that AI will open a new of prototyping. Back in the day, prototyping was how you figured out what to build, you'd very deliberately toss the entire first (or second!) version, and you'd plan to do that.

High quality ensued. Usually ;)

I think that subscribing to another person's life prevents you from living your own. Also, "Everything is Lies, I Guess".

For me it's been useful as an idea categorizer: "oh well, that turned out to be a crap idea."

It's allowed me to clear out some long-standing brush on the forest floor. And burn it down once or twice.

Quoting:

"Because the effort was removed, so was the commitment, and with the commitment the focus, and with the focus any meaningful product at all."

This is the truth. Otherwise known as "easy come, easy go".

This is an excellent pattern; do as much as you can in the database.

External Spanner provides changes streams. Internal spanner is different, mostly because of the extreme scaling requirements in some cases (and a healthy dose of "because it already works" mixed with "arbitrary change streams are scary").

Internal Spanner allows any transaction to write queue entries, where queues are (more or less) tables with some special time awareness. You can schedule delivery. Entries get pushed from queues to a handler which can also do writes to the DB within the dequeue transaction. And all of the same scaling is there.

Yes, you are reading too much into things. The ultrawealthy are supporting the current MAGA nonsense because they wish to permanently lock up the massive wealth transfer they've engineered over the past two decades, and the only way to do that is through a combination of nationalism, populism, and fascism.

Every part of the MAGA platform is a smokescreen of outrage, intended as cover for policies that favor the ultrawealthy.

An aware and motivated population legislates and taxes their way out of the establishment and perpetuation of dynasties; this has been done in the past.

If you live in a Red State, it is highly like that my Blue State money pays for your health care, highways, narcan, and a myriad of other transfers.

Like, I feel for you and your situation, but I just don't think it's sustainable for Blue States to keep being patsies by letting the Red States control what happens with Blue State money.

I have a habit of upvoting attempts at civilized argument, so I upvote once again.

For the "people who understand supply/demand", why use "want a limit" language? What you actually mean is "want a lower limit, from Y to X".

It's flat-out amazing to me that you blame immigrants for the problems of the American medical system -- which are entirely political in cause and financial in nature.

Stop with your logic, please. Obviously they must complete American Nationalism training, readily available in whatever country they come from, which they can learn from Voice of America.

I cannot understand why people downvote otherwise civilized posts they disagree with, so I'll upvote.

That said, you are impressively wrong. If someone doesn't agree with me because they choose to believe obviously false or made-up data, that is being irrational.

Is it rational to suppress large-scale studies of vaccination? If someone says "I am against vaccination because there are no large-scale studies", is that rational?

Well, the first thing the "new" antigravity did was f^&* up all of my existing projects by replicating them (with one replica per conversation within each project). That's really bad.

I now have 45+ projects pointing at a the same 5-7 folders (the actual projects). Can I delete those extra projects? The warnings are sure telling me not to.

Like aphyr?

Gun deaths in the US are just the cost of doing business, and business is booming (for some).

Sad.

One person who spent 15 years learning and building in the domain. He very much knows what he's doing, what questions to ask, and what machines do.

Every child reads a book about solving problems, assumes they can now solve problems, and is disappointed when that is not true.

If you write your own paper (mostly) and choose your own references (because you've actually read the papers) you won't have a problem.

This puts the burden to make sure it's right on the submitter, where it should be. Verification can come at any time after that; the submitter understands the consequences of hallucinated references. Verification can be crowd-sourced (and likely will be).

Nothing stops someone from putting a PDF on the internet. I'm fine with ArXiv holding a high standard.

Not sure where the nefarious is in that. That's how PKI works.

100% up-to-date VSCode is still pretty trashy, IMO. It's a mixed bag of plugins without cohesion, no awareness of code other than what that mixed bag attempts to provide (poorly). It is and always has been little more than a progressively more complicated mobius loop of autocompletion-oriented UI experimentation.

Ah, I feel so much better now. ;)

VSCode never made it past the first 10% of what Eclipse did (does). VSCode did succeed at being something for everybody, available everywhere.

It's not about forcing everyone to use the same IDE. It's about making sure that there's at least one IDE that everyone can use.

Over time, engineers realize that Code Search is more important than their IDE.

Some of the most productive engineers I know at Google are proud (and adaptable) VIM users, always have been, and nobody is going to tell them they should use anything else. They're also just fine with AI tooling, and fit it right into their VIM workflows.

Stadia was pretty technically awesome, and also a rounding error on Google's overall engineering budget.

"Ambitious" engineering means something very different inside of Google. Example: Spanner. Infra Spanner is correctly described as a "generational achievement". Very few people outside of Google have any idea that it exists, or what it does, and that's fine.

I came from storage, so the monkey knife fight there was between PARMs. Very entertaining. For storage engineering could basically say "Well, figure it out, because if you don't find XX capacity, Google will stop working. Like, all of it."

Oh so it's YOUR fault. ;)

Don't worry -- I came to love Cider for the simplicity. I tolerate Cider V, but its "anything" nature means it's not good at anything in particular. These days, I mostly use it to peek into what (Antigravity's internal equivalent) does.

I was in the Eclipse camp, prior to the IntelliJ reversal. At the time there were at least double the number of active daily users of Eclipse, Google had hired some original Eclipse devs who did an awesome job making Eclipse work at Google scale, and basically I was back to where I had been (in productivity) before joining Google.

The decision was made to go with Eclipse. Then it magically went into some sort of internal box/decision process, and came out IntelliJ instead. I've always thought this was because of a sufficiently highly placed Android person with a personal preference, but I could be 100% wrong.

This made me sad. I escalated internally, compiled all of the usage numbers, did feature comparisons on what actually worked in each IDE, to no avail. Near the end, Eclipse's C++ support and refactoring actually worked reasonably well on Blobstore, which was NOT a small thing.

IMO IntelliJ never worked very well in google3, and certainly didn't have anywhere near the level of fluidity and speed that Eclipse had (all the way back from its VisualAge Smalltalk roots -- something even most users of Eclipse never really understood or got into). That said, Eclipse just had the wrong architecture for a massive monorepo. It could be made to work (and it was), but it was never a good fit...and getting the upstream changes needed was apparently problematic.

Plain simple Cider was better (in my mind) than IntelliJ's broken functionality that worked in the outside world, but not in google3 (at least not on the code bases that I worked in).

Plain old Cider just kept adding smart features that solved problems and made it nicer. By the time Cider V was coming, it had big shoes to fill.

The Acquired podcasts on Google are a solid background.

https://www.acquired.fm/episodes/google

My standard joke here:

Q: Say, what does this Python code do?

A: Nobody f&%^ing knows.

Nefariously how?