I host the Applied Science channel on youtube (electronics, mechanics, chemistry, etc):
https://www.youtube.com/user/bkraz333
https://www.linkedin.com/in/ben-krasnow-6796a94
I disabled all wake timers so that the machine will never wake itself from sleep
Then, I set the Windows Update active hours to be early in the morning. Now, when I use the computer I wake it from sleep with the keyboard, use it normally, then put it to sleep when done. This has allowed me to install updates when convenient for me since it effectively stops auto update behavior. Sort of like how it should be.
I've made purchases from Amazon 1-2 times per week for the last few years, and have never once received an item that I suspected as counterfeit. I've ordered everything from food to everyday sundries to pretty esoteric lab equipment. I'm always very surprised by these threads on HN where people say that everything is counterfeit.
Cheap, no-name electronics are plentiful, of course. Not everyone can afford to shop at high-end electronics stores that stock only name brands. Sometimes the knockoffs work better than the original. Often times not. Why do people think the sky is falling?
I made a video in which I tried to replicate this effect. It didn't work. Unrolling tape will definitely create a high potential difference, which will accelerate electrons in a vacuum. For medically relevant levels of x-rays, the potential needs to be 20KV at least, and about 100 microamps at minimum. This is 2 watts. The triboelectric effect is not very efficient, so imagine you have to get 20 watts into the unrolling tape. The adhesive will separate before it will accept this power level. Anyway, I was thinking of revisiting...
I still really like Netflix, but the increased force-feeding of their own content doesn't excite me about the future. I would switch to a competing service if it offered a wider selection of movies, but without any Netflix content. If I found a streaming service that had nearly every movie ever produced, I would pay a much larger subscription fee.
I've had a YouTube channel for years, and talk with other creators on a regular basis. This article is basically correct, but glosses over the point of sponsorships, side deals, and most of all, Patreon. All YouTube creators that I know personally make more money via Patreon than ad revenue. Just a data point. In the old days, Patreon publicly displayed each Creator's pledge amount, and it could be compared to socialblade.com for estimated ad revenue.
The biggest problem with flat management is that it isn't really flat. It only means that there is no formal organizational structure. Think of it like the Wild West: It's flat in the sense that you can get away with any behavior if you have enough influence or force. Those who are powerful end up ruling everyone else, making the system anything but flat. I'm not at all convinced this is better than a structured system that tries to ensure fairness (even if flawed). A lot of folks at Valve had very bad experiences with unfit managers (mostly at Microsoft), and I can understand why they developed a system to avoid the nastiness they experienced in the past. It sometimes works well: The unstructured teams appear to be very effective at product maintenance. The products that are already successful and have an established fan base do well with a small group of folks with an informal leader, who doesn't have to make difficult decisions to keep the fan base happy. For new products, there are lots of difficult decisions that affect the future of employees and customers, and it doesn't appear possible for a group to make these decisions together. I spent many, many hours in meetings discussing the same issues repeatedly because there was no leader to make a final decision and allow everyone to get back to work. Instead we just debated without end, causing everyone to become unfocused and demoralized. It's analysis paralysis in the worst way I've ever seen. I now understand why many social groups from primitive tribes to 1st-world nations have leaders and chiefs. It's fine to vote them out when necessary, but having visionary leadership is really critical to development and I don't see how this can be distributed among a group unless each person shares exactly the same vision, which I suppose is possible, but unlikely.
I think the best solution is to build structures that greatly reward effective management. Having a good manager is a tremendous benefit to both the company and the manager's reports. Of course, this is easier said than done, but the return on investment spent in this area is really quite high.
The work on VR at Valve was the most stressful point of my career, and I wanted a change (ie not to work on VR anymore). When I left, it was unclear if Valve was going to ditch the whole VR thing. They probably would have done so if HTC hadn't gotten a demo of our amazing tech. My new job at Verily (Google Life Sciences) offered better pay, lower stress, and far more job satisfaction, so it ended up being a good move.
I joined Valve because "flat management" sounds a lot better than it actually is. I don't think Valve's implementation is especially bad -- I would never join another company that has "flat management".
Overall, I think Valve is a good place to work, and I learned a lot from all of the incredibly smart people there. The main reason that I left was the difficulty in merging hardware development with the company's exceptionally successful business model. The hardware team was pressured to give away lots of IP that could have been licensed, with the explanation that hardware is just so worthless anyway compared to online software sales, there was no other choice. It's possible that this was a good faith gamble, however it still doesn't preclude the use of business contracts that would have protected our investment. It also isn't so great for morale to hear everyday that your years of work are going to be given away to another company, and then watch that company get acquired for $2B. This is especially the case since many employees strongly voiced concerns about just such a scenario.
Growth at Valve is a little funny. It's like an oil-rich nation. Gaining more citizens is not really desirable since the wealth will just get spread more thinly, and there is no way that new employees will be able to make more money than Steam already does with a very limited number of employees. I know the company is trying new things, and making honest investments in new areas, but it's hard to shake the bottom line.
It fits a pattern. I was a hardware engineer at Valve during the early VR days, working mostly on Lighthouse and the internal dev headset. There were a few employees who insisted that the Valve VR group give away both hardware and software to Oculus with the hope that they would work together with Valve on VR. The tech was literally given away -- no contract, no license. After the facebook acquisition, these folks presumably received large financial incentives to join facebook, which they did. It was the most questionable thing I've seen in my whole career, and was partially caused by Valve's flat management structure and general lack of oversight. I left shortly after.
I used lots of surplus parts from eBay, collected over 5-6 years. The actual build process took a few months of pretty focused night/weekend work (10-20 hours per week). Total cost was around $1500. Details here: https://www.youtube.com/watch?v=L6HxTk9tfQk
Most commercial SEMs use electromagnetic lenses because they provide higher power (more drastically change the path of electrons) in a smaller space. The problem is that they require custom iron pole pieces. It's easier to build electroststatic lenses, which are just metal conductors and the geometry is not extremely critical. Search for Einzel lens for details. The condenser and objective lens in my scope are Einzel lenses made from copper pipe from the hardware store. I like the idea of looking at lead type with the SEM!
The angle of the incoming electrons toward the specimen is what creates the viewing angle in the image. The angle of the secondary electron detector is what creates the apparent illumination direction. It's true that setting the specimen to split the beam and detector angles will provide a good image, similar to having a good key light in photography.
Thank you! Yes, I'm still making new videos, just more slowly than in the past. My "new" job at Verily (Google Life Sciences) is very compelling and rewarding, so I've been focusing on that.
The "f number" of the SEM is about 200, so it's depth of field is much larger than an optical microscope or camera. Also, the SEM signal of bright and dark areas is made of electrons, which can travel in curved paths and around corners. This gives the SEM image an appearance that is distinct from light images, and is helpful when viewing 3d objects for ease of feature recognition as well as aesthetic quality. Even at low magnification, SEM is often more useful. Thanks!
No. Even when I started the homebuilt SEM, I considered it a learning opportunity and wanted the thrill and satisfaction of having built a working machine. I didn't have any SEM research to do, and suspected correctly that the image quality is not better than a light microscope.
No. The resolution of this (primitive) device is 10's of microns, which is too large to be disturbed by normal vibrations. Most SEM vibrations in simple machines are caused by insufficient damping from its own vacuum pumps.
Yes, I have often wondered the same thing, but I think there is a rational explanation. Over evolutionary timescales, the human groups who put more social pressure on having lots of kids and raising them carefully were ultimately more prolific. So, the genes that encouraged these social norms were selected and later, the notion of "love" became heavily intertwined with the amount of care someone gave to their spouse and family. Additionally, having kids is one of the very few major life decisions that cannot be undone and requires a huge amount of effort from the parents. Folks who honestly regret having kids experience a lot of cognitive dissonance since they necessarily committed a large part of their lives to the endeavor, and to admit it was a mistake would be admitting a huge personal error. It's emotionally easier to insist that loving a family is the most important thing a person can do in his/her life, because realizing it's just evolutionary mechanics causes even more discomfort. Questioning the importance of love leads a pathway to cognitive dissonance, so is avoided because of the consequences of that realization. I always assume that the people who talk about the importance of family are mostly trying to convince themselves, as with much rhetoric. But of course, I'm now arguing against, so what does that say about me?
It's really sad to see no mention of Valve anywhere in the article. The work that was done at Valve in 2012-2013 on VR hardware and software was critical to the success of Oculus. Valve created a complete VR demo that was shown to Andreessen-Horowitz near the end of 2013, and was a deciding factor in their decision to invest, as well as Facebook's decision to buy. At that time, the best hardware created by Oculus did not have 6DOF, high-resolution tracking. It did not have low-persistance, globally-illuminated displays. It did not have proper lens distortion correction. It did not have presence. Basically, all of the hardware factors that have made modern VR exciting were discovered and developed by the team at Valve. Unfortunately, Valve was not able to capitalize or commercialize all of this work, and basically just gave it away through a very awkward interaction. Source: I worked there on VR. Of course, they are now working with HTC on the Vive, which is really great, and is going to give Oculus major competition.
In some cases, a really unhelpful or ineffective person will cause enough problems among peers that the situation will be escalated to more senior folks via informal channels. The senior folks compare the badness of the person with other past problematic people at Valve, and ultimately make the decision whether the person should be fired or tolerated because they provide some sort of special value to the company. In many cases, this all makes sense, and works fine. In other cases, the firing does not originate with peers, but with the senior folks. I worked with a guy who was providing solid help with a project. He was fired along with others in the "great cleansing" of 2013, slowing our progress and creating unnecessary turmoil. We were told by the senior folks that this person was bad in the past. He may not be bad now, but his past badness was too much, and he had to go (?!). Who knows if this made sense or not. Information is not freely distributed within the company, so it's often hard to tell why certain things operate the way they do. With proper perspective, it may make sense, but you have to work at Valve for 10+ years to get that perspective.
It's very easy to verify that a company provides free lunches. It could take months or years of employment to determine that a company's "flat management" structure actually exists in a form that agrees with employees expectations of the designation. They are not the same kind of perk, at all. Flat management may not even be a perk! Good management is a perk.
I worked at Valve for several years, and my experience was quite different from what the popular handbook described. It wasn't an especially bad experience, but it wasn't nearly as good as the lore makes it out to be. As you might expect, without written rules or hierarchy, the scene is dominated by social forces. If you put a bunch of people in a room, tell them that there is no hierarchy, then compel them to complete a difficult task, natural human interaction will create a pecking order among other very ugly control systems. Without some rational thought (ie written and mutually-agreed rules and laws), things can turn into Lord of the Flies pretty quickly. It's true that no one tells anyone else what to do at Valve. It's also true that people get fired without warning. Think of it more like the wild west than anything else. Sure, you can be very successful there, but the chaos level might be too high for some folks (me). Check out glassdoor for other anecdotes.
I suspect the dark area was a piece of lightweight dirt (a tiny lint ball, or sawdust) that began accumulating charge as it was imaged by the electron beam. After 20 minutes or so (20 frames in realtime), the dirt developed enough repulsion force to be launched off the surface of the record. It's sort of like a charged comb or other object repelling styrofoam particles.
Thanks! I really appreciate it. My source image is 1000x1500, which redbubble.com says is not large enough for a poster. I admit, even at 150dpi, it wouldn't be much of a poster. As I upgrade my SEM image acquisition hardware, I'll be able to make higher res images. If you'd like the raw PNG, check out my blog: http://benkrasnow.blogspot.com/
It's true that a SEM wasn't strictly necessary, but there are a few interesting features that make SEM images visually different, and perhaps more appealing, than light microscope images. The electron signal emitted by the surface of the objects in the chamber doesn't behave like light bouncing off an object's surface. Normally shadows are cast by photons traveling in straight lines, and our minds have become accustomed to this behavior. The electrons in the SEM can travel in curved paths, so the illumination in the SEM image can effectively go around corners, leading to surreal shadows and highlights.
The aperture size in the SEM is much smaller than that used in a camera or microscope, leading to much higher depth of field for a given geometry and working distance. A camera might be f/16, while this SEM is effectively f/200
The brightness of objects in the image is partially influenced by the material properties and local geometry, so an exposed protrusion or edge will tend to be very bright -- something that doesn't happen with macro photography.
Hey, it's fun to play with scanning electron microscopes.