Right? The good news is companies are getting there; the bad news is it will still take some time to get to us. If I were Queen of the World, I would have made more liberal use of the governmental ability to centralize efforts towards testing and PPE production ... last January.
Guilty as charged. And though I haven’t had a chance to read the Ed Yong article yet, I agree with the general stance that it’s surprising we understand anything at all about immunology, and what we do understand is incredibly surface level. As someone who has worked with both software/tech and science, I encourage you all to learn more about immunology— it’s a giant, multi-agent network that is incredibly complex, but the limitations of tooling mean we study it one parameter at a time, mostly using rudimentary classification frameworks (eg, cell types as determined by particular cell markers), rather than large data-driven approaches.
While many companies still use Python 2.X, they are sufficiently similar that I would recommend learning 3 and retrofitting your knowledge set if necessary later on.
Can we all agree that choosing a photo of Kevin Spacey when talking about bad bosses is in really bad taste at this point in time? I get the House of Cards reference, but...
Sorry if I seemed too negative initially-- I was just typing on a phone :) For the sake of public record, I'll add a few notes here.
"12. Don't be afraid to ask for the job after your interview..." Only, only do this if you can pull off suave confidence well, and if you are fairly certain that you killed the interview(s). It is very rare to have someone ace all interview segments in the modern tech interview (for better or for worse), and if someone were on the fence, but seemed to be so totally unaware as to ask for the job on the spot, I would be concerned about self-awareness. Not to mention that many companies have a review process anyhow, so it can seem naive to be over-confident. That said, perfectly fine to exhibit confidence about how the interview went-- I feel like this went well, I'm really excited, what are the next steps, etc.
"20... You can type something like this: Project Manager - Actively seeking a new opportunity..." To each his/her own, but I would go one step further, and just avoid having "actively seeking" in the title anywhere. There is a lot of noise in LinkedIn messages, and I tend to ignore those that are too needy at the first outreach.
"24. Write your resume in the same font. Write it in the same colour and sized font." If you don't know anything about type design, this is probably safe. But after a dozen or more resumes fly by, I don't mind a little pop of color or finesse :)
"27. Leave photos or graphs out of your resume. These can confuse an employer's tracking system...." While I totally agree with the tip, I think the reason is wrong here. Many tracking systems are fine with photos. The real problem is that it's very jarring to see a photo on a resume, and feels too personal. Or heaven forfend, your marital status, which I have also seen.
"30. Apply for jobs where you only have to upload your resume...." It's annoying, but some of the best employers in the tech industry still do this, so, probably a little premature to use this as a hard filter.
I disagree with many minor points here, but one strongly enough to comment: do not send your resume as a Word doc. Any modern tracking system parses PDFs, and Word docs make it look like you don’t know how to use a computer.
"The more constraints one imposes, the more one frees one's self of the chains that shackle the spirit." - Stravinsky, who was apparently gender-neutral before it was cool. Use it as a challenge to come up with creative, human comments that don't rely on gender to be charming :)
As a female software engineer, I haven't historically found "he" to be a huge barrier-- the bigger barrier is too few comments to begin with :) I would generally prefer any comments, gendered or not, but, all else being equal, it seems like we may as well un-gender them. I think here of Geena Davis's comments about the movie industry (loosely paraphrased): When people complain about the lack of realism in showing big company boards as full of women and people of color, she said, why shouldn't we? Movies are the one place we can say, "Poof! Everything is magically evenly distributed!" and maybe in doing so encourage more women and people of color to dream of being board members and the like. Similarly, in comments, we can magically make things more equitable with almost no investment, and if that helps traditional outsiders even a little, then, great!
Note that it is similarly without any substantive details. Not to be a naysayer, buuuutttt... knowing a fair amount about type 1 diabetes, I can suggest any number of possible explanations that are far short of "miracle." The most obvious one is that he had over-treated type 2 diabetes and increased physical activity is leading to an increase in insulin sensitivity that looks like a miraculous cure. Alternatively, especially given the late age of diagnosis, he might have Latent Autoimmune Diabetes of Adults (LADA), which is kind of like really slowly developing type 1 diabetes. The body naturally goes through ebbs and flows in relative efficiency and autoimmunity, so if he still had some functioning beta cells, a temporary period of reduced autoimmunity could allow those cells to resurge and produce enough insulin to support his body without the addition of exogenous insulin. However, that's unlikely to last if that's the case.
Again, not to pooh-pooh all theoretically good news, but... I see lots of miracle cures for diabetes in headlines, and none on the pharmacy shelves.
It is worth noting that while there is evidence that murine beta cells can regenerate even late into life, there is very little indication that human beta cells can do the same. Many beta cell regeneration studies die in between mouse and human studies. Some more on the questions around regeneration here: http://diabetes.diabetesjournals.org/content/59/10/2340 with discussion of human versus rodent toward the end.
I generally seed with one word, but I could in theory use as many as there are memory-steps in the LSTM. For the shorter poems, I run with 50 steps worth of memory, so could reasonably seed with up to 50 words, and the machine would carry on from there.
As to multiple meanings, I do not force an interpretation, and the model is simple enough that it does not explicitly disambiguate. Part of the fun for me as an artist is letting the model go where it may, often surprising me when it does.
As a mother with children, the idea of a sick room is terrifying. It's bad enough to bring a sick kid (or even post-sick kid) into an office full of otherwise healthy people, but the idea of my kid then being incubated in a room with everyone else's sick-kid germs, touching things and sleeping in the same sleeping bag... ick.
As a manager, my policy is that the worse thing you can do is infect us all. If you are sick, please, quarantine yourself! Work from home. You will not be dinged for working at home, but if I end up sick, I will be grumpy indeed ;)
It might be the lack of detail in the piece, but it's unclear to me why this isn't a hammer to kill a fly-- that is, why wouldn't a much simpler peak-finding algorithm be appropriate here? What is the NN doing that's more than just peak finding over many molecules? Is there some interdependency that I am missing, or is this just signal-processing over millions of independent traces?