In the display that TSA agents are given, the image is no longer shown and instead the system shows an outline of a body with the questionable area marked. This was added after the aforementioned outrage from travelers so that every TSA agent didn’t get an x-ray view of your naked body. The scanners are the same, but the agent doesn’t see it anymore.

I absolutely agree, but you’re talking about a situation where we already have 10 different ways and 20 EC2 instances. When you get to that point (or start approaching it), yeah, do the complex thing - no argument at all. The challenge is to wait until the last responsible moment to make that pivot and to not dive deeper into the complexity than you need at the current time and place. I’ve worked with countless small companies and teams in the past that have created whole K8s clusters, Terraform provisioning plans, and the whole kit for a single low volume service because “we’ll need it when things scale out later” and later never arrives.

This is great until

I think that’s the point. Don’t jump to the complex right away. Keep it simple and compose the capabilities you have readily available until you need to become more complex. When the task requires it, yeah, do the complex thing, but keep the simplicity mandate in mind and only add the new complexity that you need. You can get pretty far with the simple, and what about all of the situations where that future pivot or growth never happens?

The philosophy strikes a cord with me - I’m often contending with teams that are building for the future complexities that they think might come up, and we realize later that we did get complexity in the problem later, but not the kind we had planned for, so all of that infrastructure and planning was wasted on an imaginary problem that no only didn’t help us but often actually make our task harder. The trick is to keep the solution set composable and flexible so that if complexity shows up later, we can reconfigure and build the new capabilities that we need rather than having to maneuver a large complicated system that we built on a white board before we really knew what the problem looked like.

It won’t matter if there are ways to side load or circumvent, though. 99.9% of users will not be willing to be bothered with such things and the US market would effectively die for the app.

Either tiktok becomes an American company or leaves… Ah, the free market has spoken

People keep saying this and I’m struggling to understand where this idea is coming from. The bill isn’t saying that they have to sell TikTok to a US company. They don’t have to sell it to the US government, or an owner in the US. Just divorce the company from explicit control by the Chinese government. Currently, the government can request any data they want from TikTok and they are obligated to provided it. Similarly, business laws in China mean that the government can also push changes down into the company, like a tweak to the algorithm to influence foreign perceptions of a topic for example.

The requirements laid out in this bill are meant to break that obligation and influence. It doesn’t say who should own the company - only who shouldn’t.

How would we feel if say, China decided Microsoft/Google/AWS/Oracle had to sell to a Chinese company on the grounds of national security?

But no one is saying that ByteDance has to sell TikTok to a US company. Just divest it to an owner that is not beholden to the Chinese government and obligated to share any and all data upon request. Compared to the legal requirements that China puts on US companies operating in China, this is a pretty tame ask.

It has a secondary interaction interface that’s novel - if you hold your hand at the right position, it projects data or controls into your palm which can then be navigated by tilting your hand and “clicking” with a finger tap gesture. This interface is also more private, and used for entering your pin to unlock the device, but can be used for other interactions like viewing long form responses to voice prompts where you can scroll through the data rather than trying to absorb everything as it’s spoken (or if you don’t want to have a spoken reply).

It’s an interesting concept, but I tend to agree with the user you replied to in that this is a solution in search of a problem.

The Google Reader shutdown hit me hard also. They offered all of the features in a really great app and many of the competitors shut down in their wake, so when they exited the scene, it left a huge hole.

I sort of agree with you, but not in the way I think you meant it.

Vista’s problem was that it’s hardware requirements were too high for it’s time. Operating systems have very long project development lifecycle and at a point early on they did a forward looking estimate of where the PC market would be by the time Vista released, and they overshot. When it was almost ready to release it to the world Microsoft put out the initial minimum and recommended specs and PC sellers (Dell, HP, Gateway) lobbied them to lower the numbers; the cost of a PC that met the recommended specs was just too high for the existing PC market and it would kill their sales numbers if they started selling PCs that met those figures. Microsoft complied and lowered the specs, but didn’t actually change the operating system in any meaningful way - they just changed a few numbers on a piece of paper and added some configurations that let you disable some of the more hardware intensive bits. The result was that most Vista users were running it on hardware that wasn’t actually able to run it properly, which lead to horrible user experiences. Anyone that bought a high end PC or built one themselves and ran Vista on that, however, seemed quite happy with the operating system.

They can, though the employees would be able to claim unemployment if the job was remote and then changed to on-site but if the job was on-site with a temporary remote policy the employee wouldn’t have a leg to stand on there and could be dismissed for cause.

In the US, what you can and cannot fire someone for is complicated and counter intuitive.

A low performer that is part of a protected class is hard to fire because you need to have copious documentation that they were dismissed due to poor performance and were not targeted for their protected class status. This is a good thing and prevents unscrupulous bosses from firing a woman for getting pregnant, targeting people of a particular race, religion, or gender, or any number of other awful things. Those things will only come up if the former employee sues, and many will not, so some bad bosses or companies get away with this while others end up in court because someone that needed to be fired is crying discrimination.

On the flip side, if it falls outside of those protected classes, you can fire someone for any other reason or no reason at all. “I woke up in a bad mood and picked a name out of a hat to fire” is legal. You may get a fight if the person you picked claims discrimination on one of the protected classes and you have to explain to a judge that you’re actually just a bad human and not discriminating, but it’s allowed.

In the US, there is rarely, if ever, a contract. Unless you can show that you were let go for a legally protected cause (your age, race, religion, gender, and some other things), employers can fire you without any reason at all.

The only caveat here is the differentiation between for cause and without cause, as it impacts your ability to collect unemployment insurance payments. Employers pay those insurance premiums to the government and they are based on how often people let go from that company claim the insurance payments, so a company that lets go of a lot of employees is going to pay more than one that manages to find a way to fire them for cause or get them to quit.

I think this is very likely, though it’s also prolonging this whole exercise by avoiding the dramatic conclusion and spreading the pain out over a longer time.

If every manager at Amazon woke up tomorrow and said “screw it, we’re enforcing this policy”, that would result in a mass firing event of quality talent, and Amazon would feel the pain of their policy decisions and either have to swallow that and try to move on or beat a hasty retreat and call this whole thing off. It would be a quick and decisive end to this whole debate, but instead we have month after month of employees stressed and angry while looking rebellious and unmanageable, managers stressed and frustrated while looking ineffective, and the senior leadership frustrated and looking impotent.

Someone’s going to win this fight eventually, but everyone trying to find middle ground and skirt the policy just takes what would be one big fight and turns it into many months of slow unease and turmoil that’s bad for everyone. I want the remote people to win this, but sometimes the way to win is the lose on purpose. Let the dog catch the car so he can realize what an idiot he was being.

The bit that kills me is that “make Google Chat not suck” doesn’t seem to be in the list of options for addressing this problem at all. I work for a company that uses GSuite and chat is universally loathed with a bunch of Slack instances running around the company, both sanctioned and unsanctioned. If they spent time working to improve chat, the momentum of being a GSuite company would carry the rest of the weight here. It doesn’t have to be better than Slack, just closer.

China coming in with the latest in new tech for people with too much money to accidentally kill themselves. The finest innovation in the field since the Cesna.

AI resume screeners are very much at risk of bias. There have been stories about exactly this in years past. The ML models need to be trained, so they get fed resumes of candidates that were hired and not hired so the model can learn to differentiate the two and make decisions on new resumes in the future. That training, though, takes any bias that went into previous decisions and brings it forward.

From the Amazon I linked above, the model was prioritizing white men over women and people of color. When you think back to how these models were trained, though, that’s exactly what you’d expect to happen. No one was intentionally introducing bias to the AI process, but software teams have historically been very male and white, and when referrals and references come into play, those demographics were further emphasized. And then let’s not pretend that none of those recruiters or hiring managers were bringing their own bias to the table.

If you feed that into your model as it’s training data, of course the model is going to continue to favor white men, not because it’s actually looking for men, but because resumes that men typically submit are the kinds that get hired. Then they found that resumes that mention a professional women’s organization or historically black or women only colleges were typically not hired. The model isn’t “thinking” about why that is - it just knows that when certain traits exist, the resume is ranked lower, so it replicates that.

Building a truly unbiased AI system is actually incredibly difficult, not the least due to the fact that the demographics of the data scientists working on these systems are themselves predominantly male and white themselves. We’ve also seen this issue in the past with other AI systems, including facial recognition systems, where these systems built by teams of white men can’t seem to make reliable determinations when looking at a picture of a black woman (with accuracy rates 20-30% lower for black woman compared to white men).