jaxbot

Doesn't that broadly seem like a problem? Given the preferred pronouns here are "they/them", which is what people ought to be using if they don't know, the problem stems from people inherently assuming "he/him" for any technical work. That's a problem beyond just trans/gnc issues.

Wow, this sounds eerily similar to my situation, though I was only there about 3 years.

I left to join an AV startup and it's amazing how much more I've learned and accomplished in a few months versus the time at Google. Things at Google move slowly, and the amount of work per person is relatively limited. Also, all the complicated infrastructure or codebase decisions were already made for you, or is being handled by someone L+2 at least and outside of your purview.

Edmond Lau's The Effective Engineer talks about this, except to the extreme that he wanted to do and learn everything at Google, and even then he left after ~2 years after he felt his growth was slowing.

I think for many, being at Google for a few years will give you invaluable experience, but then severely diminishing returns on growth and practical experience unless you're one of the lucky ones who gets promo'd every year or two.

For your actual question, though: >Now that I'm thinking of jumping ship to other interesting companies, I'm having serious doubts that I really learned what I should have learned during all those years.

Having Google on your resume always helps get people interested. Having experience working on big teams with big codebases is also something not everyone in this industry has, and there's value to it, even if you'll initially scratch your head at how to build without Blaze.

>Especially since I'm considering companies with a higher hiring bar than Google.

Curious why you're confident in that statement. I've found that Google has a much higher hiring bar to the actual required skill -- they basically seem to hire as though everyone will work on GWS, when in reality many are just copy-pasting CSS and BUILD rules from another project.

On the flip side, many more interesting companies have lower hiring bars relative to the job requirement. It's harder to hire good talent when you're not FAANG with a pipeline right out of the Ivy League.

>How can I keep myself accountable while I'm still at the company to deeply learn the FE/BE technologies to be better prepared for other companies? Should I start by preparing a checklist of technologies and dive into each of them for a month and continue from there?

Well, I would focus LeetCode, tbqh. That's still the standard. But whatever you want to do, focus in on technologies used in that industry/job role. Do some side projects. Maybe take some online classes. I think you'll find the practical experience requirement to be lower than you think. People can generally learn the right technologies, and companies know this. It won't be a big deal unless you're a frontend-only SWE who suddenly wants a ML role or something.

Lastly -- feel free to DM me, I use this handle on twitter and gmail, happy to help, especially if you're curious about where I ended up.

It also wasn't in isolation; he made very similar remarks in 2006 on his blog that he got flak for, but people mostly brushed aside as 'oh, Stallman..'

Comments here seem to mostly equate this situation to a Cancel Culture outcry over an isolated remark. That's not what happened here. rms has had decades of inexcusable behavior for any individual, much less someone affiliated with MIT and heading something as large as FSF. He had to answer for this eventually.

I sincerely appreciate his contributions to this world. But I also sincerely feel that we can't give people free passes for their behavior (see: courtesy cards at conferences) just because they've done well in other respects. We need to end the acceptance of Brilliant Jerks.

If you read through Autonomy by Lawrence Burns, the author states that Levandowski felt the project was losing focus & getting too bloated, and wanted to disrupt it with a Team Macintosh style parallel project that he would head.

It could've just been part of the big ego'd power struggle that occurred on the project and saw the exit of many early stage employees, but I also wonder if he had a point, looking back today at the stages the project went through. It's hard not to feel like they thought they were close to launch, only to hit some roadblocks and need to rebuild major chunks of it. Speculation. But they're a huge company with a product that is at risk of getting usurped at this point.

Read the license on this thing. "open data set" is a stretch here. It's against the license terms to publish a trained model, or its weights, or even use the dataset for models that are run on a physical vehicle (supposedly even if that's just a research vehicle).

Is this normal for 'open datasets'?

If I make project Foo, which leverages #Script, and make Foo MIT-licensed, by your statement #Script is now also MIT licensed in this project. So then Company Blah wants to use Foo, which happens to just be thin wrapper/modification around #Script, and is now permitted to take it and fork it into closed-source repos, yes? That is what this statement implies.

Regardless, this is why getting clever with software licenses, and giving guarantees in the form of comments instead of actual license legalese, is problematic. The intention might be clear, but the governance is only by law, and this appears to be in conflict with your statements.

This is one of our favorite jokes internally, often invoked when new signage is put up to clarify how trashcans work or the like.

Even on a lighter note, I'm getting more and more frustrated at cloud device providers changing features on a whim. For example, I own Nest products that I bought to integrated into my 'smart home'. Now Google is ripping out the API for Nest products and making us use Google Assistant. Google Assistant is repeatedly changing behaviors and breaking my workflow.

I understand I'm a tiny segment of the market, but it's hard not to want to switch over to fully owned, open source alternatives. Any suggestions?

During a snow storm/PABT outage this past winter, I observed a jitney line from 8th ave down halfway to 10th. It's pretty crazy to see the mishmash of alternative transit modes people will piece together when their main route (NJT, PATH, NJ bus, etc) goes offline.

Same. Basically anyone with a newer Veriphone system (which everyone seemed to roll out back when we switched to chip) seems to support it. I use Google Pay, so it's not Apple-specific anyhow. But it's so much faster to process than the chip cards, even if I already had both my phone and card in my hand.

Right. There's no enforcement of visa itineraries so long as your first one has valid flights & hotel. After that, you get 10 years of multiple entry with no required itineraries. I just brought it up as a point, that Xinjiang is not a place you would normally be going. People seem to assume that Xinjiang is just a normal crossing into China that has heightened security. That's not the case at all. You would never accidentally end up there on a trip to China.

I'm not sure and I don't want to misspeak. I know that my company does not do such a thing, and my phone when roaming (Project Fi) bypasses the GFW. However, I do know there are specific different IT infra requirements when connecting from China but none that are gov-monitoring related. It may be a misunderstanding, or it may be a procedure that your friends' IT was following that we aren't.

That being said, foreigners accessing the normal WWW is really not a major concern if you think about it.

If that's the case, it's worth noting (since it's often lost here, on reddit and in the news) that is app is only required at border checkpoints in Xinjiang. Putting Xinjiang on your itinerary would likely result in your visa being refused anyway. If you are traveling to China for business or pleasure, you won't need to worry about this.

Source: Was in Shenzhen within the past month. Normal customs/border control practices. Nobody asked for personal devices. Only surveillance that was obvious was biometric checkpoint at customs, and some sort of face scan (?) at the subway queue.