GVIrish

Starts with how you evaluate employees for bonuses and promotion. Do you evaluate people on the impact of what was delivered? How fast they delivered feature work? The quality level of what they delivered? How well they worked with others?

The answers to basic questions like that already starts to shape behavior. If you pay zero attention to how people behave, and only look at impact of what was delivered you may promote people who optimize for their own work, but make others miserable. If you don't properly weight quality, especially now with AI code gen, you'll promote people who move fast break more things than is reasonable.

We can easily find examples of suboptimal behavior that arises out of poorly shaped rewards incentives at companies. Empire building is one behavior that is the result of managers getting promoted based on headcount. Stack ranking can and has led to people limiting collaboration with peers because someone has to fail in order for someone else to get a favorable rating. Or people avoid riskier work because failure can put you on the hot seat.

Maybe. Anonymity is where bad actors play.

The problem is when the government changes the definition of 'bad actor'.

Musk intervened at FAA to get a Verizon communications contract cancelled while quietly trying to get FAA to sign to a Starlink contract:

https://www.forbes.com/sites/alisondurkee/2025/03/13/elon-mu...

That hasn't been my experience over 20 years. I've worked in multiple SCIFS that didn't handle SCI at all.

So you're using the word 'compromised'. In this context that would mean malware, unauthorized access, circumvented logging, etc. If someone thought this was happening the answer would be to lock the system down, perform forensic audits, and prosecute anyone who compromised these systems.

If you're talking about fear of leakers, the response to that is to tighten the distribution of information and start a counterintelligence investigation.

In any case the simple risk calculus is, what is the risk of adversaries getting a hold of this information and causing grave and lasting damage to national security and death vs the risk of political rivals leaking something. Pretty simple decision there and one that any cabinet member should get right.

SCIFs are for viewing TS materials, whether or not they are SCI. Even then, SCIFs are often employed for processing things that are only marked Secret or systems only handling Secret. But yes, if we want to be specific, Secret has a lower bar and can be worked on outside of SCIFs but still not in public or at home.

That doesn't really make sense. If they had strong reason to believe that the secure comms systems they were supposed to be using were compromised, using personal phones to communicate outside of SCIFs is very, very far from what any competent person who understands and is briefed on the threat environment would do. Note that none of the people involved are making that argument because it would make them look even more incompetent.

Using Signal in this case is wrong and foolish full stop, and the extremely likely reason they did so is so they could escape standard government record keeping compliance (NARA).

To start with, classified information is ONLY supposed to viewed in a SCIF. Secondly, it should never be loaded onto private devices. The private phones of national security leadership would be prime targets for every hostile intelligence agency in the world. It matters little if the information was encrypted in transit if the host device is compromised.

One would have to be a fool to not trust all of the classified tools and safeguards the US government uses only to then use a commercial app on commercial phones to communicate classified data in public while stateside and abroad. Just the fact that someone could accidentally add an unauthorized person to the chat is but one reason it was crazy for them to do this.

but to attack the core idea as essentially wrong is anti math, science, and rationality

The way Microsoft implemented stack ranking was anti math. You're supposed to measure the data then calculate the level of fit to a distribution, not artificially shoehorn the data into buckets to create the curve. If you analyze the data honestly you may find you have a bimodal distribution, or a heavily skewed distribution, who knows.

Stack ranking just clumsily says, I'm gonna give x% a bad score, y% a middle score, and z% the top score.

That's not the point. The test giver has free discretion to say either answer is correct or incorrect. You could argue that if the intent was to underline "word" that it would have quotes around it, but it doesn't matter because the test is not supposed to be fair or consistent.

Things like this were at the heart of what Jim Crow was in America. Selective and capricious enforcement of the law to disenfranchise and disadvantage black people at best, enable unaccountable violence against them at the worst.

Microsoft pay isn't the best in the industry so if compensation is the only thing that matters to someone, Microsoft shouldn't even be in their top five.

People refuse to acknowledge that their conservation efforts for a year are undone by some guy in Texas in five minutes

I don't think that's a good way to look at things. Some guy in Texas is polluting a lot more than you, ok, but would it be better if that guy keeps polluting and you pollute just as much? We can't get hung up on, 'well some person/company somewhere else is undoing my savings'. That kind of gets into tragedy of the commons thinking.

It is disheartening to see parts of the country going in the opposite direction than we should be going for sure. And well-meaning but not very useful policies can be a pain. But I try not to be disheartened at backward thinking in other locales, I try to look at the places making advances (for example, India is ahead of schedule in the shift to renewables) to be find some optimistic amidst the bleakness.

Typically you're not asking for a price(salary) upfront. Ideally you've done a bit of homework to figure out what positions can pay in the ballpark of what you want. Then you get their offer at the end of the interview process and negotiate as appropriate.

Yeah you may go through hours of interviews and not end up with an offer in some cases but think of it this way: you could potentially earn tens of thousands to over one hundred thousand extra dollar per year off that time investment in interviewing.

Attempting to line up multiple interviews and balancing offer timelines is hard, but the payoff can be huge. Considering that people spend 4 years or more working in college to get into their career, making a time investment of a few weeks to get a potentially large raise is nothing.

I think the author is taking a bit of the wrong lesson from his experience.

One of the challenges of coordinating a group of people is getting everyone to buy into the same vision. Fact is that other people see the world differently and may have different goals. Here the author is attributing that to narcissism and maliciousness when most of the time it isn't that.

So yes, as you add more people it gets more challenging to get everyone rowing in the same direction. This is why setting a clear direction and clear communication is key, but the increase in communication overhead as the team grows is always going to be difficult. In this case, as others have said, he could've just open-sourced his code so the people who had a different idea were free to run with it.

Reputation may not affect Google's ad based revenue, but it absolutely affects their ability to profit in lines of business outside of ads.

Google might have made more inroads with enterprises with G suite and GCP if they didn't have that reputation. The gaming industry is a 200 billion/year market that Google could've captured a decent size of if potential customers trusted that they wouldn't quickly give up. All of that represents billions of dollars in lost opportunities.

Since 9/11 there's been a shift towards not siloing information as aggressively as before in order to make intelligence failures less likely. Don't know if that was at play in this case though.

At the very least, access control systems should've flagged unusual access to more information that this person would've had a need to know. But as big as the US intelligence and defense apparatus is, not every agency, program, and office is gonna have rigorous enough controls to catch people like this. Seems like the lesson should've been learned after Manning and Snowden.

I would argue that the red tape is not so bureaucrats can make more money, it's that Americans have a very, very strong aversion to seeing people benefit from social programs that they think are undeserving. So we put all of these hurdles up to make things 'fair'.

Just look at the difference in outrage between when there's some story of someone on public assistance buying something like smartphone, vs the reaction when we found out about all of the businesses abusing the PPP loan system. Or the fact that many of the PPP loans were forgiven.

It's like as a country we're fine with wealthy people abusing the system. But then we turn around and would rather let 100 deserving people struggle just so maybe one person can't get a free ride.

This is why police reform is so hard to achieve. When crime spikes, people are far less committed to police reform.

Even when crime isn't spiking, police can just sit on their hands whenever reforms are pushed and instantly change the narrative. This happened in SF with the AG recall and has happened in numerous municipalities. Local politicians can't afford to anger the police union because all the police has to do is stop enforcing the law aggressively and a local politician is done for.

So New Yorkers may want police reforms but the NYPD union has a lot of leverage to resist it. And voters generally dont' have the stomach to oppose police union backlash.

The federal government certainly is not adopting remote work as much as they could, but this OpEd is a bit off base.

For starters, there are a lot more federal agencies embracing remote work since the pandemic. From the people in my network working as government employees or contractors, I'd say more than half have more remote work options that they had before, with some going fully remote. There are still stubborn leaders and organizations but remote work has made major inroads in the last few years.

The fact remains that some jobs can't go full remote as a function of security clearance requirements. It's laughable to say, "...so long as the employee has access to a sensitive compartmented information facility (SCIF) in their location," as if SCIF's are readily available around the country, and that somehow it's easy to get a seat in one. If you work for one three letter agency, you can't just sign up for a hotdesk at another three letter agency's SCIF.

The market is shifting in the DC Metro where even cleared jobs are allowing some remote work, but it'll still take some time for organizations to change.

Twitter actually had some profitable years lately and in the last 5 years before Musk bought it they recorded a net profit. Revenue had been grown 60% over that time.

But now with Twitter reportedly losing 40% of its advertisers and taking on $1 billion of annual debt servicing costs, it seems unlikely they'll be able to get back to profitable if things keep going this way.

If you fire people hastily but then are forced to ask them to come back because you discovered they were actually key personnel, that speaks to incompetence. The speed of Twitter's layoffs were completely under the control of Musk. There was no crisis or external factor that made them cut people so quickly.

Mistakes certainly can happen, but they're much more likely to happy when things are done chaotically without proper planning.

The fact that Twitter had to ask some of the people it laid off to come back shortly after having fired them says it all about how incompetent and self-defeating those layoffs were.

Sure we won't run out of water, the challenge is gonna come when weather patterns shift in a way that cause droughts on crop-bearing land or abnormal temperatures in key parts of the growing season. Some parts of the world may be in for some considerable pain in those scenarios.

Bingo. This has made me spend a lot less time on Youtube because recommended content and search are worse for me. If I'm searching for a marine biology video I don't want to see recommendations for clickbait influencers 3 results in. I used to go down random rabbit holes when I found related interesting content, but now that just doesn't happen very much anymore.

The problem is that Google makes enough money from ad business that they don't care that they may be losing out on more revenue/profits by having better support. This fundamental belief is one reason Google has a hard time making inroads in the enterprise market. In that market, support absolutely matters and the perception that Google doesn't care about it makes it easier to exclude Google altogether when considering a cloud platform or office productivity suite.

Stack ranking seems to misunderstand how statistics work.

This is the fundamental flaw with stack ranking that really irks me. Stack ranking force fits people to a normal distribution but that's backwards. With normal distributions you're supposed to measure the phenomena then use statistical formulas to see if the data fits a normal distribution. Maybe it does. Maybe it's bi-modal, or tri-modal. Maybe the data is evenly distributed.

But no, companies butcher the concept of a bell curve and end up doing serious damage to the culture of the company as everyone starts optimizing their behavior to succeed in an utterly irrational system.

To be fair, in the internal email they described it as 'discretionary pto'.

My guess is that Musk wins the battle of the wills here; he no longer needs to maximize advertiser income, just get enough to make the wheels turn

Twitter now has to pay $1 billion a year in debt servicing costs. Twitter was not wildly profitable before, so it seems very unlikely Musk can make enough money to break even, much less profit, while losing a large portion of advertising revenue.

That's why he's trying cut costs so aggressively, and rushing to gain new subscription revenue.

You didn't see Twitter Spaces, Fleets, Birdwatch, Tweet Editing, and Twitter Blue? Some of those features have been successful for them, some not, but they've been releasing features.

The thing I always wonder is, in habitats where birds would normally have to deal with feline predators, how is the domestic cat worse than that? Any birds in environments that have feline predators would already be adapted to avoidng cats, unlike say, small Australian fauna.

I can understand the damage inflicted by cats in environments that never had a very efficient bird predator like cats (e.g. New Zealand and other islands). But in North America there's the Bobcat, Lynx, Ocelot, Margay, and Jaguarundi. Even without the domestic cat, birds would still get eaten in sizable numbers by cats. Sure there are a lot more domestic cats out there, but without civilization there would be a lot more wildcats out there eating birds.

Maybe the impact is simply that outside house cats exist in numbers that are not supportable for a normal wild predator due to the fact that house cats have meals provided by humans to sustain them. So house cat numbers don't reach an equilibrium with their prey the way a wild cat could. Either way I suspect that the impact on bird populations by domestic cats is often overstated.