Last night I wanted to try to get some new targets, and so rather than just relying on what the Unistellar app suggested, I used the search function. I searched for “nebula,” and then started clicking through the resulting set, looking for something that (a) had a positive elevation (meaning it was “up” in the sky), and (b) was something I didn’t already have a picture of. It worked for the veil nebulae (see my recent post about them), so I figured, sure, let’s give a few a try.
I started with an old favorite to make sure everything was dialed in:
M51 – Whirlpool Galaxy
14m
2025-08-22
Another great photo of a great pair of galaxies.
Great! Let’s try one. Beach Ball Nebula? That sounds interesting.
Beach Ball Nebula (NGC 6058)
28m
2025-08-22
An attempt at a new target, and it proved to be nothing much. If you zoom in on the middle, there’s a small blob of blue, maybe? Looking it up afterwards, it has an angular size of 24 arcseconds – about 1-2% of the width of the circle. Yeah, that’s as good as we’re gonna get.
Ok, zero-for-one, let’s try another one. Maybe… the Cat’s Eye Nebula? The online photos of it look really interesting, let’s go for that one.
Cat’s Eye Nebula (Caldwell 6)
10m
2025-08-22
Another attempt at a new target, which proved to be not a good target for the eVscope 2. The core is the bright spot at the center, and with an angular size of only 25ish arcseconds; that’s tiny. The outer halo is 5 arc minutes, or about 1/6th the width of the circle, but it’s too dim for me to make out from inside a city, especially with such a short dwell time.
Alright, there’s still good night left, let’s try a third.
Elephant’s Trunk Nebula
63m
2025-08-22
If there’s a nebula here, I don’t see it. I even gave it over an hour of dwell time – it can produce great results on nebulae like the Horse’s Head or the Veil. Here? Nothing.
So, the next day (today), I looked around online, and it turns out I did hit the dead center of something called IC 1396A, which is a nebula that’s six times wider than the field of view. This thing is huge! It clocks in at around three degrees – compare that to the moon, which is about 0.5 degrees, and would fill most of the circle. The center of the nebula is the dimmest and most boring part, so I got a whole lot of nuthin’.
The Elephant’s Trunk, which is PART of the larger nebula, is only about 10 arcminutes, which is about 1/3rd the width of the moon. That’d be perfect! Unfortunately, it’s about one or two fields-of-view away from this shot. So it turns out that the Unistellar catalog is wrong: what it thinks is the “Elephant’s Trunk Nebula” is actually targeting HD 206267A, which is the super bright triple star in the center of the larger IC 1396A nebula.
I need to come back here and reposition the camera to point at some of the more interesting locations within the nebula complex, and see what I can catch.
So that the night wasn’t a total waste, I went back to the one that I shot the previous night – but out of focus – to pay homage to a great galaxy for the eVscope. It was nice to end the night on a win, instead of three strikes in a row.
M101 – Pinwheel Galaxy
53m
2025-08-22
Long dwell times definitely help this spectacular galaxy come alive.
So, in the end, I learned that for the most part, the Unistellar suggestion as to good vs. bad nebulae is mostly right, but obviously not fully correct, because the Veil Nebulae were great, and should be included (with accurate targeting info!). And I learned that the catalog is wrong…ish, since for the Elephant’s Trunk Nebula, it’s targeting the wrong thing, but in the right nebula, which is too big for the telescope, but the part you’d want to capture with the eVscope is almost perfect, so they should fix it!
We finally got a clear night last night in Seattle. It’s been cloudy and rainy for a while, which is great for the plants, but bad for telescopes.
I started by trying to catch an old friend, the Pinwheel Galaxy, and only after it’d been running for a while did I notice the focus was slightly off.
This emphasizes something I learned early on with the eVscope 2, but keep forgetting: tiny bumps to the focus knob (e.g. while moving it in or out of the house, or putting it away in your backpack or container) QUICKLY mess up your photos. I often don’t adjust focus before I start, and honestly, I should. These were adjustments that I was able quickly make with the Bahtinov mask, and I only needed a few millimeters of tweaks to get the focus dialed in. That’s how sensitive the telescope is: even though the focus for the images below were only a few millimeters off, … yuk.
So if you’re new to using an eVscope, or really any telescope, the process of getting set up should include a focus check. If you don’t, you’ll end up with a pictures like these:
M102
17m
2025-08-21
This is what happens if you take a picture when it’s out of focus.
M101 – Pinwheel Galaxy
19m
2025-08-21
Out of focus! Well, if you’re out of focus, it looks like this.
So… yeah. I’m not going to include these in the galaxies page, since they’re so bad. But I include them here as examples and reference, so you know what to do when your stars start looking like squares with crosses in the middle.
Once I fixed the focus, I decided to try for a new target – one that I’d read about, but I never remembered seeing on the eVscope list of available targets: the Veil Nebula. I don’t know why the Veil Nebula doesn’t appear on the list of available objects. If you search for “veil,” though, you’ll get 2 potential targets: the eastern veil nebula, and the western veil nebula. And then, when you pick one, the telescope complains that it’s not in the visible area. The elevation might be in the 30-60 degree range, with a reasonable azimuth, but the “visible during” is listed as “???”. But I encourage you to ignore the complaints, slew to the target, and start the dwell, because you end up with shots like these:
Eastern Veil Nebula
68m
2025-08-21
The Veil Nebula is the visible part of a huge complex called the Cygnus Loop, which is the remnants of a supernova from 10-20 thousand years ago. The Eastern Veil is one of the larger and brighter bits, and it only shows up faintly despite the long dwell time.
Western Veil Nebula
8m
2025-08-21
The Western Veil is another bright part of the Cygnus Loop. This shot got interrupted, but even though it’s only 8 minutes, I really like the fact you can clearly make out both blue and red elements of the cloud.
Western Veil Nebula
57m
2025-08-21
A longer shot of the western veil, and even though this was almost an hour, you can only make out faint elements. The bright star is 52 Cygni, which is what’s called a “foreground star.” The veil nebula is about 1,470 light years away, whereas 52 Cygni is only around 201 light years away.
If you read my article about Amazon’s new pay structure, and you also read my article about how talent is a power curve, the insightful among you may have noticed an interesting correlation between them. Specifically, could Amazon’s talent curve, when combined with the target compensation ranges, actually be a good way of compensating people “fairly?”
Well, let’s start with questioning the word “fairly.” As noted in the paper I cited in the talent post [https://www.hermanaguinis.com/pdf/PPsych2012.pdf], people don’t really like it if someone else is paid three or four times what they’re paid… even though the “someone else” is objectively three or four times more productive than they are. So “fairly” ends up a loaded term.
But I COULD believe an argument that asserts that’s what Amazon’s compensation ranges were trying to accomplish. (I’m not saying I do believe it – I’m just saying it’s possibly true.) Going back to my post, I had an example where the 50th percentile for performance was 41% more productive than the least productive person, whereas the 90th percentile was 216% more productive.
I’m pretty sure the top of Amazon’s compensation ranges are NOT three times higher than the bottom of the range, at least at levels 7 and below (Sr. Manager, Principal Engineer, etc.). Therefore, by at least one measure of “fair,” it seems it can’t be. The fact that talent is a power curve means you WILL have people that are two or three times more productive than another person in the same role. But the compensation ranges mean that you can only compensate them, what, maybe 50% more?
On the other hand, if you’re constrained by the market and by industry averages, and you want to try to pay those superstars closer to “fairly,” then maybe this new compensation model gets you closer to that goal. And by the same argument, if you can’t tell the difference between the 5th and 10th percentile, and the 20th percentile isn’t much better than either of them, then maybe lumping them in at the bottom of the range is close to the best you can do? I don’t know, though – having 65% of the people making 20% or less of the pay range would imply a very steep power curve (i.e. an alpha well above 3). Maybe that’s justified? I don’t have the data to tell.
My gut and experience tells me that software engineers aren’t that polarized – there are some ineffective ones, a lot of somewhat effective ones, and while there are only a few hyper-effective ones, I’m thinking 1 out of ten, not 1 or 2 out of 100. That would suggest a flatter power curve than the one implied by Amazon’s compensation targets.
So, as a cis-het white male, born with a trust fund, I got to not only play the game of life on the lowest difficulty setting, I also got “wealth” and “intelligence” as my two best starting stats. Constitution might’ve started a bit on the low side, but again, I was lucky enough to be born into a world with most of the late-game science and medicine achievements already researched, and therefore got to live long enough to go to college, get a job, get married, and reproduce.
If I was a narcissist, or a conservative, I might think that it’s not luck; that I earned my place. That I got what I deserved. That I got where I was because I worked harder and was smarter than other people.
I mean, I did work hard, and I was pretty much always the smartest person in the room. But I also know that I was exceptionally lucky, and that the society I was born into, and the race I happened to be born, stacked the deck in my favor.
How do I know?
Let’s time warp back to April 29th, 1992. (Some of you already know where I’m going.)
In April 1992, I was finishing my sophomore year in college. At the beginning of the year, I had been assigned to live in a suite with four other men – as it turned out, four black men. We each had our own bedrooms, but we shared a single bathroom. There wasn’t a common area, so when we hung out, it was in someone’s bedroom – usually the guy’s that had the little color TV and a VCR.
For the first 8 months of knowing these men, the hardest thing that I personally experienced of “being a black man” was clogged shower drains. Their hair was different than mine, and much more likely to clog the drain.
That changed on the 29th, though. I came back after dinner, and all of the other guys were piled into one of their rooms – the one with the TV.
The Rodney King riots started in Los Angeles on April 29th. A year before, Rodney King had been beaten by four white police officers, and it was captured on video tape. The beating had been broadcast nationally for months. Most of you know George Floyd, or Treyvon Martin, but us Gen-X folks think of Rodney King first.
The four LAPD officers were acquitted on the 29th. At 6pm EST, the verdicts were announced, and the riots started minutes later.
I went into the room, stood in the back, and watched as fires started to burn. I watched as windows were smashed and businesses were looted. I watched helicopter shots of Korean shopkeepers defending their businesses with rifles on the rooftops.
It was terrifying.
After a time, I found some courage, and asked my suitemates to help me understand what was happening. I wanted to understand what could cause this, and why it blew up so violently.
And these men gave me time, and patience, and answers. We talked well past midnight. I asked questions, they answered; they asked me questions, and I answered.
They told me about The Talk. All of them had gotten The Talk.
The Talk is what black mothers tell their sons, about how to behave around police officers. Keep your eyes low. Say yes sir and no sir. Keep your hands out of your pockets. Never try to run. Never fight back, even if they get in your face, or hit you, or push you into a wall, or slam you to the ground.
Their mothers gave them The Talk because they were mothers, and more than anything else in the world, they wanted their sons to come home each night, safe and sound.
On April 29th, I learned that all four of these men, at one point in their lives (most more than once), had been handcuffed and placed in the back of a police car, for the crime of walking while black. They had done nothing wrong, other than being in the wrong place, at the wrong time, with the wrong color skin.
They were in middle school, with a backpack, walking home from the bus stop after wrestling practice.
They were in high school, at a party with some friends, where there wasn’t any alcohol or drugs; when the cops showed up, their white friends went home, while they went to the precinct.
None of them were ever convicted. None ever went to trial. Most never even got charged – just held, maybe in the back of the police car, maybe at the precinct, maybe overnight. But because they’d gotten The Talk, all of them came home to their mothers.
These were four strong, intelligent, family men. These men, through determination, hard work, and exceptional intelligence, had earned their way into an Ivy League school. All of these men had passed their freshman courses and were nearly done with their sophomore year. Today, people might say they were “DEI admissions,” but I know for a fact they were anything but. They’d earned their spot by their talent, grit, intelligence, and drive, and had performed admirably in the academic arena.
So when I say I’m lucky, and when I say “I got to play the game of life on the lowest difficulty setting,” I mean that my mom never had to give me The Talk. I mean that cops don’t look at me like they did at them. I mean that I’ve never been in handcuffs just because I was walking down the street, after dark, with a backpack on.
I mean that I’m not trans, so nobody’s ever called me a pedophile or yelled at me in a bathroom.
I mean that I’m not gay, so I’ve never been beat up because of who I hold hands with.
ICE is never going to give me a second glance.
I’m not saying I’m better, or thank god I’m not a race or sexual orientation other than what I am. I’m saying that I have not had prejudice and hatred, that others have suffered their entire lives, ever directed at me. And I didn’t really understand that – until that night.
When people say “woke,” it’s an honor, not a stigma. I am lucky, because I woke up in April of 1992. I am lucky, because I got to live with and listen to those men, with the right color skin, in the right place, at the right time. I am lucky, because they were patient and open and honest with me, because I was open and honest with them.
LE and HV1 are 40% of the population, and by the new policy, they get the range minimum. So 40% of people at the company, by definition, will make the bare minimum of the range. Add in HV2, and 65% of people are making 20% or less of the range. With HV3 (50%), 80% of the people are making mid-range or less.
But it gets worse: by policy, when you get promoted, you’re an automatic HV1 at the next review cycle. So when you get promoted, you get range minimum for the next year, guaranteed, regardless of your performance. In year 2, if you get HV2 or HV3, you get LESS than other HV2s or HV3s, because you were (by fiat) HV1 last year.
Therefore, if/when you get promoted, you’re locked into the bottom of the next level’s range for a year, and the bottom half for 1 to 2 MORE years, unless you’re a rock star that can hit TT by year 3 (and even then, you’re capped at 70% of range).
Amazon already makes it hard to get promoted, and now they’ve reduced the incentive to do so. You have to perform “at the next level” for at least a year (or two or three, if you want to get promoted to L7). To them, they get people performing at a full level above their current compensation for as long as possible, but then pay them only AT THEIR LEVEL. And this new policy means that even once they’re promoted, they’re paid range minimum for a couple years, and probably below the range midpoint for 3 or more years. So Amazon gets 4 or 5 (or more) years where they L7 performance out of you for the low-low price of “at or below range minimum.”
One of Amazon’s leadership principles is “Hire and Develop the Best.” In the old days, I was told that one of the ways that we did that was to also pay them the best. I was told that Amazon’s pay range started at the 50th percentile of the industry average, the midpoint was at 75%, and the top was at the top of the industry average.
The new policy makes me believe that what I was told is no longer true, and now Amazon’s midpoint is essentially industry average. But what if it’s still true? That Amazon’s “range” is from 50% to 100%? It means that “the best” people in the industry, who Amazon supposedly “hired and developed,” are mostly getting paid industry average, or maybe a bit better. It means 80% of Amazonians have their target compensation set at or below the 75th percentile of the industry. And you’re trying to get promoted, they’re going to pay you at or below the bare minimum for your performance level for 5 or more years.
Go read the article. All of it. Science is hard! When people say “I’ve done the research,” ask them what papers they’ve read. If they say “I got it off LinkedIn” or “I’ve done my research on Facebook,” laugh at them. But if they say “I got it from this paper,” and provide a link to that paper, and they’ve actually read the paper … well, maybe believe them, maybe not, but please don’t laugh at me.
Ok? Did you read it? Good. But if not, and you really only want the TL;DR, the premise is that a “normal” or “Gaussian” or “bell” curve is not the correct way to model the distribution of talent.
Talent is a power curve. Or, more precisely, a Paretian (power law) distribution. The paper combined 5 studies of different professional fields from acting to singing to NCAA football to Major League Baseball, and showed that for objective measures of talent, the χ-squared fit for a power curve is a billion times better than a Gaussian curve. Not the hyperbolic billion – literally an average χ-squared misfit of 2,092 vs. 2,769,315,505,476 in one of the studies.
So, by now you should be convinced that talent is a power curve, not a bell curve. So what?
Well first, it’s very important to notice that a Paretian distribution is “scale invariant.” [https://en.wikipedia.org/wiki/Scale_invariance] This means that if you cut off the curve at any point, and re-plot the resulting curve, the NEW curve is ALSO a power curve. So if you pick a particular metric, like the number of career home runs, you’ll get essentially the same curve for AAA minor-league baseball that you will for Major League Baseball, even though most every MLB player is probably better than every AAA player. Similarly, if you cut off all the MLB players below the 50th percentile, and re-mapped the ones that are left, the same distribution emerges. This will become relevant later.
Next, let’s discuss human perception.
Since talent is a power curve, it’s easy to tell the difference at the top end of the scale. The difference in performance between the 90th and 95th percentiles is much larger than the difference in performance between the 5th and 10th percentiles. So it’s easy to tell who your top performers are, but hard to tell who your bottom performers are.
To put it to practice, let’s say you use lines of productive code as your objective measure (I know it’s not, but go with me, here). And let’s assume the worst coder at the company can write 100 lines of code a week. That means the 5th percentile performer will write 103 lines of code, and the 10th percentile performer will write 105 lines of code. On the other hand, the 90th percentile person will write 316 lines of code, and the 95th will write 447 – over 40% more than the 90th percentile coder, which means the gap between “worst” and “average” is smaller than the gap between 90th and 95th percentiles.
But now you’re a director, and what do you do when you’ve got a group of 10 people in each range, and you had to stack rank them? Take the image above and draw ten lines between the 5th and 10th percentile, without overlapping them. I’ll wait. While you do that, I’ll do it in the spread between 90th and 95th, and we’ll see who finishes first.
(The above chart, and the data I derived from it, is a Pareto distribution, with m=1 and α=2, for anyone that wants to double-check my work. That means the midpoint / 50th percentile is right at √2, or 1.414. So the “average” coder would produce 141 lines of code, which is the green line in the chart above. I concede that coding talent may not be m=1 and α=2; maybe it’s m=3.14 and α=2.718, and maybe it changes from company to company. I assert that my point still holds, and regardless of your function parameters, the relative difference between the 5th and 10th percentile is vastly and obviously smaller than the difference between the 90th and 95th percentile.)
I assume, by now, that you agree that discerning the difference at the bottom of the scale is much harder than the top. So let’s put that aside, just for a minute. Keep it nearby, though, because we’ll come back to it.
But when do you enforce it? If you have 5 people that report to you, how do you apportion 35% of them to HV1? You kinda can’t have 1.75 people ranked HV1 (at least without a chainsaw, and I’d argue you shouldn’t even then). So let’s assume, for the sake of argument, that Amazon strictly enforces The Curve on organizations as soon as they reach 40 people or more.
Ok, seems alright, no? 40 people? That should be large enough to even out, right? Nope.
Turns out there’s a function that nicely describes the likelihood of X occurrences in Y events, when the expected outcome is Z occurrences. It’s called a Poisson distribution. [https://en.wikipedia.org/wiki/Poisson_distribution] You can model this pretty easily in Excel, by generating random numbers for cohorts of 40 people, then mapping how many actually end up in LE, HV1, HV2, and so on, assuming strict application of The Curve. I did it for 100,000 cohorts of 40 people, and my results vs. the predicted Poisson curve matched quite well.
When you run the numbers, if you have a bunch of cohorts of 40 people that you’re trying to stack rank at Amazon, about 13% of your cohorts will ACTUALLY have 0 LE employees, 27% will have 1, 27% will have 2, and the remaining 33% or so will have 3 or more. There’s about a one-in-twenty chance that you’ve got 5 or more in your group!
But… you’re mandated to fit The Curve. So if you do your job “correctly,” and walk into OLR with exactly 2 LE employees in your group of 40 people, there’s a 73% chance that you’re wrong.
Just plain wrong. Not maybe, not sometimes, not “only if you’re a bad manager” wrong. By definition, and by design, you’re wrong. (Probably.)
Furthermore, if you apply The Curve on cohorts of 40, and run a large simulation (or just apply the statistics), you will discover that 53% of people who should be LE… are incorrectly rated HV1. And on the flip side, for each person in your cohort rated LE, there’s a 53% chance that they’re actually HV1.
So even if we assume perfect information, the result of the above policies (i.e. enforcing the curve on orgs when they reach 40 employees) will be that more than half the people that end up LE shouldn’t be there, and that more than half the people that should be LE, aren’t. And ALL of this is assuming that managers are completely unbiased, that there is some objective measure of talent, and managers are able to accurately and precisely measure it.
And that’s where we START. And it’s going to get worse, because it’s time to start questioning those assumptions.
Ok, now let’s go back to the paper I cited earlier. Talent is a power curve, right? And it’s hard to tell the difference between, say, someone at 3% versus someone at 7%, right? Because that’s what managers are asked to decide when they’re putting people in the LE bucket.
Unfortunately, the ability to accurately assess talent is, itself, a talent. And that would mean that there are managers that are TERRIBLE at assessing talent. And if you think “Amazon only hires managers that are good at assessing talent,” go back and reread what I wrote about scale invariance.
If 5% of individual contributors are LE, that means that 5% of managers are also LE. And in fact, because talent is a power curve, there are actually a LOT of managers that are relatively BAD at assessing talent accurately. Using my example above, half the managers are only 0-to-40% “more accurate” than the WORST manager at the company. Sure, there are some that are three or four times better than anyone else, but that’s only the one-manager-in-20 at the top of the curve.
But something funny happens if you try to add incompetent managers into the simulation. If you go back and run the simulation, this time adding “error” to some of the rows … the incompetence of the managers washes out in the error inherent in enforcing the curve. Since 53% of people that are LE are already (incorrectly) in HV1, then it’s a coin flip as to whether an incompetent manager down-levels someone who SHOULD have been downleveled in the first place (stumbling their way into the right decision) versus downleveling someone who shouldn’t have been. And since 53% of people in LE aren’t actually LE, it’s again a coin flip as to which one the bozo picks to uplevel.
Therefore: even assuming the “normal” levels of incompetent managers, the system itself is so broken that you still end up with a 50-50 shot whether someone’s really LE or not.
But this is only MOSTLY broken.
Where things COMPLETELY break is when biases come into play. Management tries hard to root out unconscious (or even conscious) bias, but you just can’t. Individuals are discussed in OLR for 10-15 minutes at most. If a manager wants you to be LE, and you’re borderline, you’re going to be LE, and they’re going to be able to make a case for it. Remember – we’re trying to compare Mr. 4% to Ms. 6%, so we’re talking about the difference between 103 and 105 lines of useful code per week. Or check-ins per sprint. Or defect frequency during peak. Or whatever semi-objective measure you try to trot out.
Ask yourself, which engineer is better: the one that had twelve more checkins last year? Or the one who caused two fewer sev-2s? Or the one who completed 0.26 more story points per sprint on average? Or the one who blocked the pipeline 3.3% less often? You’ve got 10 minutes to decide which two of these four are LE.
So, apologies that this was so long, and if you actually read the paper, gold star to you for sticking it out. But to sum up:
Talent is a power curve.
Therefore, telling the difference between your lowest performers is extremely difficult.
Therefore, combining top-grading (where you fire your bottom X% performers) with an enforced curve produces coin-flip outcomes by design.
Manager competence does NOT impact outcomes, because there are a lot of semi-competent managers (remember, their competence is ALSO a power curve).
Biases, on the other hand, DO affect outcomes.
And because the difference between the bottom performers is indistinguishably small, bias is all that’s left to decide who gets LE and who doesn’t.
But top performers? They’re easy. They clearly burn the brightest and do the most.
Sorry to anyone that’s been coming for the space pictures, only to be swamped by the slew of leadership- and management-related posts. There should be only 2 more “previous” LinkedIn posts to move over here, and then I plan to have a weekly update focused on leadership & management, with the space posts happening “whenever there’s a clear night sky in Seattle” (so … yeah, who knows?).
But to keep the space people happy, and because I’ve wanted to do it for a while, here’s a post of my current top 10 space pictures, in no specific order. You can find all of these images in the Galaxies, Nebulae, or Solar System Objects pages (none of the star clusters made my top 10, but they’re nice! Go check ’em out!).
Let me know your favorite, or if you have any that you’d like me to revisit.
M33 – Triangulum Galaxy
2025-07-29
54m
The Triangulum is a great galaxy for the eVscope, because it’s large, distinct, and relatively bright. You can really see the different colors in the stars – some are red giants, others are blue hypergiants, and even others just appear white. If you have the time and a good clear night, spend an hour focused on this galaxy, and you’ll be happy with what you can capture.
M81 – Bode’s Galaxy
20m
2025-05-02
Bode’s is a great target, and really benefits from the color-stretching software update. The bright center really pops out, and the wide spirals are somewhat unique (at least for what you can get with an eVscope).
M101 – Pinwheel Galaxy
16m
2023-05-26
This is the one photo that literally made me jump for joy when I saw it. I’ve wanted to witness a supernova during my lifetime, and I got to catch one here! The bright star in the right-most arm of the Pinwheel Galaxy is a supernova that exploded a day or two before this photo. Go to the Galaxies page to see before & after photos, and you can clearly see the bright new star appearing, and then fading away over the ensuing months.
M104 – Sombrero Galaxy
70m
2025-05-04
The Sombrero is nice because it’s bright enough that you can get a reasonably good picture with a much shorter dwell time (5-10 minutes). But as with so many of these deep sky objects, if you can leave it running – like I did here, for 70 minutes – you can get even more striking images. I really like the sharp contrast between the bright core and the thick, light-obscuring dust on the edge of the galaxy.
NGC 0891
19m
2025-01-21
This is definitely in the top 3 for me. It’s so distinct, and so perfectly edge-on. Just like we have problems seeing things outside the Milky Way if they’re in line with (and therefore blocked by) our galactic plane, if there are any intelligent species in this galaxy, they almost certainly will never know that we exist, since we’re positioned perfectly along the plane of THEIR galaxy.
B33 – Horsehead Nebula
55m
2025-01-25
Another in my top 3, and I’ve rotated it 90 degrees to make the horse’s head upright. If you want to catch this one with your own telescope, you need a very clear night, a very long dwell time, and use the enhancement software. I love this photo.
M16 – Eagle Nebula
2025-07-29
22m
Anyone that’s seen the “Pillars of Creation” photo from the Hubble Telescope (or the updated one from the James Webb) can recognize those dark pillar-shaped splotches are in the middle of this nebula. And I got ’em! On my telescope! From my back yard! These are pillars of light-obscuring dust and gasses, and they are light-years tall. I love how you can clearly make out the three-dimensional nature of these massive clouds of gas and dust – the sky isn’t just a flat black sheet with flickering pinpricks of white light! It has depth, and color, and immeasurable beauty.
M20 – Trifid Nebula
16m
2025-05-30
The Trifid Nebula is actually three nebulae: a dense, red, emission nebula, a wispier blue reflection nebula, and a spiky, shadowy dark nebula in front. The VVT software update really makes this an excellent nebula for the eVscope to image. I really want to go back here on a clear night for a hour-plus dwell time, but I’ve not yet had the chance.
M27 – Dumbbell Nebula
25m
2025-07-18
The Dumbbell is a favorite for Unistellar enthusiasts, because of its distinct shape and colors. The red is mostly hydrogen gas, which was sprayed and ejected off of the roiling surface a massive star, as it went through its violent death throes. The blue is more abundant in oxygen, which was created in the star’s core, during last few seconds before it finally ran out of fuel and collapsed under its own incredible weight. It fused hydrogen, carbon, and neon into oxygen, silicon, and iron, and then exploded back out. The huge, expanding sphere of ionized oxygen gas (plus dust and other stuff) chased the hydrogen that was burped out during the previous years or centuries, creating … this!
Comet 12P/Pons-Brooks
2m
2024-03-15
I have captured several comets, but I like this one the best, because of its rare blue-green color. This is before the color stretching software launched; I wonder what it would’ve been like with it turned on. The green is apparently “dicarbon,” which is what it sounds like: two carbons, in a single molecule. It’s rare on Earth because it’s quite reactive, but on this comet, warmed by the sun and in the depths of space, exploding chunks or rock and ice, and outgassing water and dust, it’s apparently unusually abundant.
Another previous post from LinkedIn, archived here. This originally was posted to LinkedIn on July 21, 2025.
To be clear, I’m not saying Amazon’s RTO mandate was a bad decision. I’m saying they never showed anyone why it was a good one.
I’ve often said Amazon is a data-driven, evidence-based company. We collect information objectively, discuss it, test it, verify it, and then the data will tell us the right thing to do. We make two-way door decisions quickly, but are diligent when making one-way door decisions.
RTO was a one-way door. Once you tell people “come back or you’re fired,” you can’t really take it back.
Therefore, I expected Amazon to back up their decision with evidence. But when AWS CEO Matt Garman said, “nine out of 10 Amazon employees he’d spoken with were ‘actually quite excited by this change,’” or “when we want to really, really innovate on interesting products, I have not seen an ability for us to do that when we’re not in person,” he didn’t provide any data to back it up. (Unless “nine out of ten people he’d spoken with” has a p-value under .05? Probably not.) And then Andy Jassy, in his RTO announcement, said “the advantages of being together in the office are significant,” without providing any actual evidence. These showed that the decision was more important to them than the data.
Wait. “[N]o significant changes in […] firm values after RTO mandates.” Firm values, like, Leadership Principles? So there’s data saying we’re not better at invent & simplify if we’re in the office?
So on the one hand, we have Amazon executives insisting that “our culture” is about working in an office, and that we’re not innovating if we’re not in the office, and that there are “significant advantages” to being in the office, because …REASONS. And on the other hand, there are research papers with observable, measurable, verifiable data that says … not that?
Again, I’m not saying RTO is a bad decision. I’m just saying that Amazon never showed any evidence of evidence-based decision making while making it.
The first post I made on LinkedIn was my goodbye post, announcing that I was leaving Amazon, with a few comments as to why. I was surprised when it got nearly 250,000 impressions, over 1,500 reactions, and 150 comments. It’s the closest I’ve ever had to anything going viral.
After 17 years and 9 months, today is my last day at Amazon.
The belt hook is from the Retail Pricing DICE (Diversity and Inclusion Career Event), which I proudly planned, organized, and ran back in 2019. Before that, my belt hook had the AWE logo – Amazon Women in Engineering. I’ve always been an advocate for diversity, equity, and inclusion. So when Amazon publicly announced they were abandoning their DEI programs in a craven capitulation to the Orange Taco, the question changed from “if” to “when.”
Unfortunately, Amazon’s become a Day 2 company. Senior leadership has ditched developing the best and customer obsession for profits and politics. And they never lived up to their promise to become one of Earth’s best employers.
What’s next? Well, I’ve made a LOT of money during may career by making other people a TON of money. My goal for the future is to make enough money doing something that makes the world a better place. I’m not sure yet what that means, but I know I can’t do that at Amazon. For now, I’m going to spend the summer with my family and exploring my options. You can see the results of one of my hobbies on http://www.arneknudson.com, where I’ve uploaded all the photos I’ve taken using my digital telescope.
The people that I got to work with, mentor, and learn from at Amazon remain some of the smartest and best people I’ve known. I’m proud of my accomplishments and will always treasure the great relationships I got to make. And I got to live in Europe for 4 years! I have no regrets, and I’m grateful for the opportunities, but I am sad about what Amazon’s become.
As mentioned in the about page, I’m not only an amateur / hobbyist astronomer with a high-tech telescope, I’m also a software professional. I graduated with a computer science degree back in the mid-90’s, and pretty quickly got into management. I’ve been a software manager / leader for more than 25 years, and many of those were with Amazon (almost 18).
I’m no longer at Amazon, but I’ve started writing about my thoughts and experiences. To date, I’ve mostly been posting on LinkedIn, but I’ve decided to start archiving them here, as well. That means, about once a week, there’s going to be a post about leadership, management, my experiences (and displeasure) with Amazon, and what I think others could learn from those experiences. I’ll still update this site each time I have new and interesting photos from the telescope, and I’ll keep categorizing those as “Space on the Back Porch” – you can select just the pretty space pictures by finding the “Categories” section on the right side of the page, and choosing that header. And the individual pages for Galaxies, Nebulae, and so on will also stay.
It means there’s going to be a spate of posts – basically, the first 5 that I’ve already posted on LinkedIn. But after that, you should expect about one a week. I’ll probably put all five of them up over the weekend.
The leadership posts will have a different category title. For now, I’ve called it “Being a Good Leader,” because that’s pretty much what I want to write about. I could try to stick with the theme and call it something like “Speaking From the Front Porch,” but that’s kind of corny. I do like the idea that “space on the back porch” also implies making room for people and welcoming them in, like at a barbecue. Sharing thoughts over a beer, chatting about whatever, and enjoying the company. So I’m going to keep the overall title for the site (at least for now).
Let me know if you have thoughts or ideas. Everyone is welcome here, and I’m glad you came! Pull up a chair, grab a beverage of your choice, and let’s chat.
