This comment is off-base enough that I've created a throwaway, as I post openly as a Microsoftie on my main, to prevent anyone else from getting the wrong idea.
Microsoft reintroduced stack ranking over the last year or so. It's widely documented.
The individual in question was definitely under pressure. I worked rather closely with them, and this is well documented in the source article as well.
I don't think any of the things I said above (or were insinuated about Microsoft culture by other posters) are in any way insulting to Prateek, regardless of what his individual situation or performance is. If anything, calling attention to it and attempting to address it is a powerful way to show respect to my eyes. The incentive systems at play, the pressures and stressors, will result in these outcomes unless anyone forces a change. End of story.
> Microsoft reintroduced stack ranking over the last year or so
Microsoft has forced differentiation, not stack ranking. They aren't the same and differentiation is much better for employees.
> The individual in question was definitely under pressure. I worked rather closely with them, and this is well documented in the source article as well
I'm sure you're feeling a lot of options due to your proximity, and I'm sorry for what you're going through.
The article says he was under pressure, but it didn't say the source of that pressure. Perhaps it was due to a series of "lower than expected" reviews, or the constant worry about losing their job and their visa. It didn't say that the pressure was caused by internal policies.
I've known many people who put pressure on themselves and burnt out because of it, despite no one expected them to do so. I suffered through that in my own career and nearly quit software engineering because of it.
If you have real details about the situation and that it was internal Microsoft policy or the pressure put in them by their manager which may not have aligned with Microsoft policy, that woukd be very useful information to share with the public.
Microsoft operates like many big companies vs a single company, and some teams go beyond standard Microsoft policy and have unrealistic expectations of their employees. Those departments and managers should be called out and shamed.
> Microsoft has forced differentiation, not stack ranking. They aren't the same and differentiation is much better for employees.
In practice it is exactly the same. EMs are instructed that they must have someone at for instance 80 (below-meets-expectations), or occasionally at 60 (PIP). This has resulted in layoffs. The pseudonyms we're told to use are infantilizing when we can see the results.
> I'm sure you're feeling a lot of options due to your proximity
I'd ask you not make assumptions about my feelings in this, vs taking the statements I'm making at face value.
Not that it's unique to his division, I can assure you I worked organizationally close enough to know that, regardless of other sources of stress, there is(was) relevant work pressure up through the VP level encouraging late hours and long days that goes well beyond healthy, often acknowledged outright as the way to stand out in leadership AMAs. The evidence of this is documented in the article, and your dissembling is somewhat astonishing.
> If you have real details about the situation and that it was internal Microsoft policy or the pressure put in them by their manager which may not have aligned with Microsoft policy, that woukd be very useful information to share with the public.
I'm sharing it, you're just disregarding it. There is a system in place that through basic game theory ensures a prisoner's dilemma without compromise, with the escalation being longer hours and more work. While any given leadership may encourage it more or less, to not see how that pans out naturally when employment is on the line, especially as proven out by multiple recent rounds of 'performance' driven layoffs, seems purposefully obtuse. One doesn't know when the heart attack is coming, but one does know when the PIP is coming; we can't be surprised at the choices people will make.
I didn’t take his comment that way of which GP replied to me. I took it as sarcasm. In that, GP was saying it like “Yea but they found time to make these policies…”
In reality, the outcome doesn't appear to be the result of "pure chaos and randomness" if you ground your tools. Test cases and instructions do a fantastic job of keeping them focused and moving down the right path.
If I see an LLM consistently producing something I don't like, I'll either add the correct behavior to the prompt, or create a tool that will tell it if it messed up or not, and prompt it to call the tool after each major change.
Being able to use natural gas isn't a valuable public good.
I used to live in an area with regular tornadoes, having public weather data is life saving. People shouldn't have to die from tornadoes because they're poor.
I currently live in an area with regular tornadoes. There's a huge gap between a federally subsidized weather prediction program and making sure people in your community can be alerted.
Most even small towns in my area, the kind with one stop light at best, have community storm shelters. Granted those ultimately are partly subsidized by the federal government, but that is still different than a federal agency program.
To that end, if the concern is tornado safety why doesn't the government give communities or individuals storm shelters for free rather than partially subsidizing them for those who are at least well enough off to throw a few thousand at a small unit?
I'm on an M2 and when I have the dog near the stick and it shows the hovering mouse, I left-click and nothing happens other than the do whining. How do I pick up the stick?
> When he shared his thoughts with Ethereum’s cryptographers, he was startled to learn that they were unfamiliar with this work
It would be nice if the article included timelines. Ethereum researchers have been talking about GKR since 2020,so it's hard to imagine the lack of familiarity.
The time is given as "last October," and the work they were unfamiliar with is presumably "contrived proof protocols that are vulnerable to attack, no matter which hash function you use" as stated in the immediately preceding sentence.
That's confusing to me because back in 2020 when they were looking into GKR inside of a Snark, they were worried about these attacks. Following up in 2022, Ethereum researchers were talking about attacking GKR by forging proofs and not having sufficient randomness/collision resistant.
It's hard to align what's being researched on Ethresar.ch and this statement.
I don't believe the "this work" that the article is talking about here is GKR, but work that is referenced earlier in the article:
> In the early 2000s, computer scientists showed how to do just that, contriving interactive proof protocols that were specifically designed to fail when they underwent Fiat-Shamir
Indeed, the artcile points out that targeting GKR was the idea of the Ethereum Foundation researcher.
> Soukhanov had the idea to target a Fiat-Shamir proof system based on something called the GKR protocol
It's easier to relate to people on a trip than in a war, too many people died during the great war so being concerned is overwhelming, the Titanic had the perfect headline as the unsinkable ship, rich people were on the Titanic and we care about them more than poors dying in a war, etc...
> The whole movie felt like "reels" stitched together, with no scene lasting longer than 5 minutes.
I felt the same way about the first hour of the new mission impossible movie. It was so annoying to me that I left after that, so maybe it got better in hour two.
We don't have enough information to support that.
Bi-annual performance reviews themselves aren't a bad thing that force overwork.
If he had a history of good performance reviews (100% or higher on average), the risk of getting a PIP would be very low.
Microsoft stopped stack ranking years ago.
I don't think we should speculate on people's behavior or how they aligned with company policies, because we might accidentally be insulting this man.