There's also the effects of the 2017 Tax Cut and Jobs Act which kicked in in 2022, changing how R&D can be taxed and triggering mass layoffs in tech.

Yes it does a fine job. Whatever happened with the MSN article, it wasn't a modern LLM.

chat link

They key in on parts of speech on their own, which is pretty cool: https://aclanthology.org/W19-4...

Holy shit - this is like 8 conspiracy theories all wrapped up into one giant sprawling one. And donâ(TM)t miss the comments, wow.

I recently took a look at the terms for Thinkific (an online education platform; I wanted to see if they claimed any ownership over the materials posted), and was pleasantly surprised by the "which means" sidebar breaking down the legalese. https://www.thinkific.com/resources/terms-of-service/

On the other hand, I wonder if slimy companies could word things in such a way as to gloss over slimy sections in the legalese.

We gave NFP a shot; I don't get too many opportunities to share my real experience with it, which is a shame. A little over a year ago my wife was sick of her birth control (subdermal implant), and being from a very Catholic family she wanted to at least learn about and try NFP. I reluctantly agreed (we were stable enough that starting our family a bit earlier than planned wouldn't have been a disaster).

We signed up for a class through SymptoPro through our local church, an organization that provides workshops and materials about it. In summary, I'll the say method is definitely more scientific than old "rhythm" methods, but still leaves a lot to be desired. According to their materials (and duly backed up by studies, as far as I can tell), time between menstruation and ovulation can be quite variable (hence failure of rhythm methods), but time after to the next menstruation is be fairly predictable for a given woman, and further the time of ovulation should be roughly predictable from physical symptoms, and thus also the likely times of fertility, taking into account other factors such as sperm lifetime etc. (the materials are adamant they don't try to predict ovulation per se).

The rules are very complex, but the basic "intro setup" is this: Each day the woman should take her basal body temperature. This must be done in the morning right after waking (ideally still in bed) to avoid noise from other activity. [Work swing shift? Have an over- or under- active immune system? Good luck with that.] Each day she should evaluate her vaginal discharge (character and amount), and for extra accuracy also her cervical opening (yup, you read that right). Each cycle, the last day of these fertile "signs" is taken to be "peak day" (I presume correlated with ovulation, but they don't come out and say it directly). After this peak day is met, she looks for three temperature readings all higher than the previous six; if the third is 0.4 degrees F higher than the highest of the 6, then you're good to sex it up that evening (assuming avoiding pregnancy). There are special cases involved if the third day doesn't reach the 0.4 above level, etc. There are also more advanced, less conservative rules available once 6 to 12 months of individual data have been gathered. It's really the kind of thing that should be handled by software, because it's so difficult to remember and apply. (I could also see it being fertile ground for software bugs. Pun intended.)

Is it scientific, and is it effective? I'll say "kinda" to the first, and "mostly" to the second. There have indeed been studies on pregnancy rates, and results seem comparable to condoms (though condoms have come a long way; high-90s% is about what is claimed by symptopro for perfect use; results vary across studies, YMMV, etc.) What bothers me is that I have yet to identify any science behind the rulesets themselves. Why three days after six, and where does this 0.4 degrees magic number come from? I'd love to be proven wrong, but I think these are essentially someone's hunch.

As to effectiveness, I think it's reasonably effective primarily because it's so conservative. According to the rules, we got on average maybe 6 days per month we *could* have sex and avoid pregnancy (aside from menstruation days); a couple of months we had 0 available days. We ended up using condoms quite a bit anyway. On the plus side, via all this charting we learned that this isn't quite normal: many women get closer to 10 days/cycle of infertile time (even with the basic ruleset), and my wife may have a mild "luteal phase defect." Now that we're actively trying to start our family, we'll be talking to a doctor about it.

In the end, I'd say it's not a terrible program and it's nice to have options. But, more research is needed, and it's far from the easy, one-size-fits-all solution NFP proponents tout.

As someone who will also be staring down 40 in a few years, I wonder what the modern equivalent is. Is an MBA addon to a STEM degree + experience still worth it, or is that path now too polluted to be of value (as this article seems to suggest)?

I've heard of people using git for strange purposes (blogging?!) but for a second reading this headline I had the horrific thought that someone was attempting to create a filesystem backed by git.

Fortunately that's not the case, GVFS (not that other GVFS) provides a git-compatible filesystem abstraction on Windows 10.

Although I'm not a big git fan, I do have to say I'm impressed with Microsoft's recent moves toward open-source and interoperability. If the Linux Subsystem is ever a first-class citizen in the MS ecosystem, I could even see myself using Windows again.

That's definitely a big part of it, I've reviewed a few different stylus/tablet solutions (except for some of nicer Android solutions I couldn't easily get my hands on like the note) and part of the problem is the smoothness of the ipad screen and the stylus nib; because it's so very smooth many apps do a lot of interpolation of the data to create more natural lines introducing a noticeable lag. This can be adjusted in some, with the tradeoff of an unnatural writing experience. Some (like the bamboo paper app) can be set to use the bluetooth/stylus exclusively for effective palm rejection which is pretty much a must. Sadly, many ipad applications also don't make much use of the pressure levels provided, perhaps the despite the 1024 levels of the bamboo fineline I have it's too noisy to make good use out of.

Another big issue I've found with the ipad stylus solutions is the accuracy simply because of the parallax due to the thickness of the display stack; where the "ink" appears is too far below the surface. At the surface of the glass the accuracy is pretty solid (imo). I did borrow a surface pro 3 during some of my tests, which gets a lot of things right: very little parallax, a resistance-inducing nib for natural writing and low latency.

As someone unfamiliar with Haskell (or much in the way of purely-functional programming at all), would you mind expanding on this a bit? I've dabbled in lisp, but R (insane but it is) is the language that's introduced me to more functional concepts but it's also quite a mixed bag.

For the first time in several years I haven't changed states, jobs, or marital status, so I was excited to do my own taxes*. I was used to getting a hefty refund though, so when I used hrblock.com and it showed a (small) debt I thought I'd go in for service figuring I had missed something. Turns out I am just about even in my witholdings, so I payed someone at H&R a good hundred bucks just to give me the exact same information. Sad day, but a lesson learned for next year I guess.

* Uh, excited relative to previous years I guess.

Agreed - I hate to be "that guy," but I have trouble not seeing how the summary at least couldn't just as easily say "... If they're right, then [solutions to traveling salesman] cannot exist, which explains why we do not (and cannot) observe them in the real world. Voila!" On the other hand, the summary and blog article seem pretty terrible: the paper does address the idea that the complexity of physical processes are related to the complexity of turing-like computation--insofar as we are willing to admit that our current understanding of physics is correct (at least, that's asserted as far as I can tell, top of pg. 7). These ideas have been considered before (Granade, below), but those are pretty strong unknowns.

There is some work on "what the universe can do" with regards to computation and complexity (and quantum theory), for those up for some extremely cool and mind-bending stuff. I can highly recommend Why complexity matters: A brief tour by Christopher Granade. Scott Aaronson is one of my favorites too, with the whimsical NP-Complete problems and physical reality and more philosophical Why philosophers should care about computational complexity.

Anyway, just wanted to provide some "further reading." I'm hoping for some eventual commentary on this from that community. I'm way out of my depth ;)

Oh yeah, speaking of the front page, I'll be honest, I look at three things: the headline, skim the post (awww yeah classic slashdotter here), and I see how many comments have been made. Comment count combined with headline for each and every story is a quick indicator if it's worth checking out the discussion or if I should move on down the page. (An article about a new kernel extension I don't care about it with 40 comments? Boring. An apple article with 854 comments? Probably also boring [unless I'm in the mood for reading some flamage].)

Yup, and here are some suggestions: (sigh, maybe I'll see if I can get this through their suggested email support as well... will that actually help? Editors: what say you? Does this stuff speak more loudly to the higher-ups if it comes through certain channels?)

Keep some space for ads if you want; I don't give a shit and I realize you've got bills to pay. I have the option of turning them off, but I don't because I like the site.

That said, information density is important. If you bump the font size and line spacing or significantly drop the comments column width, we can't read the comments or their surrounding comments' context. There'd better be a lot of lines before I have to "click for more", and I never want to have to "click for more" on the front page. This might mean reducing the size of those terrible banner images.

We need to be able to easily see the information on posts and navigate the discussion. Links to parent posts are absolutely necessary, current score, subject, and at least a preview of the post content if it's collapsed. Other useful information provided that I'd like to see stay prominent includes the username and UID number of poster. It was tough for me to get used to the collapsed/non-collapsed system with the last redesign, but it actually ended up giving a lot of information in a tight space and generally reserved more for better comments.

As it currently stands, the two problems cited above alone will kill the discussion oriented nature of Slashdot, users will desert, and revenue will tank.

Since there's a redesign in the works, this _could_ be a good chance to make some things actually work better! The "full" "collapsed" and "hidden" threshold sliders never seemed to work right for me. Obviously better encoding support would be nice. Maybe someday I won't have to type html to do simple formatting stuff. Since many of us are coders, perhaps some support for inline code could be cool? I won't harp on speed or javascript much, but I'm sure others will.

I will join you. I just burned all 15 of my mod points on other stories voting up "fuck beta" posts, but I won't even visit the site from the 10th to the 17th (or whenever we settle on for a boycott). It will be interesting to see how often I get mod points after this debacle is over one way or another ;)
Slashdot has never been a paragon of beauty OR usability, but this redesign really kills the discussion; I even gave it a shot.