Showing posts with label education. Show all posts
Showing posts with label education. Show all posts

Friday, September 20, 2024

Perplexity is saving my linguistics classmates

I have a dark past. I asked questions. In class. Lots of questions. Too many questions. I hear things, I get ideas, I notice gaps, I ask questions.

It's a compulsion.

Some of the questions helped classmates. To be honest more were probably confusing or distracting. I likely featured in classmate daydreams -- but not in a good way.

Worse, some of the questions confused the professor. Or exposed what they didn't understand. That could be embarrassing or even humiliating.

Now I'm back in the classroom, doing freshman linguistics.  As a 65yo, I can do classes at Minnesota state colleges and universities for free. We pay a lot in taxes, but there are benefits to living here.

My question compulsion is still there, but LLMs are saving everyone. I set up a linguistics "collection" in Perplexity with appropriate prompts; now I type my questions into my phone (allowed in class). I get the answer with Perplexity and spare my classmates.

Never say AI isn't good for something.

PS. Perplexity is to modern Google as Google was to Alta Vista. A qualitative improvement. It's almost as good as 1990s Google.



Sunday, July 12, 2020

Curbsider CME for non-internists through VCU Health

My favorite CME source, the Curbsider's Podcast, has long offered CME credit for internists (they are, after all, an internal medicine podcast). For family physicians, not so much.

There is now to get free Category One AMA credit [2] for Curbsiders podcasts through "Virginia Commonwealth University's VCUHealth Continuing Education [1] using their Curbsider curriculum.

You have to listen to the episode and complete a post-test. You can and should use the Podcast notes to compete the post-test (how we learn). 

I was able to register with VCUHealth although I have no connection there. After registration I completed my profile. (The web site is ancient and barely works in a modern browser -- don't try it on mobile.)

This is all a bit of a secret. I only know of it from a blurb at the start of recent podcasts. There's a tiny CME link to VCUHealth at the bottom of the summary page for recent episodes.

I've completed one module. There were 3 post-test question, one didn't have a clear answer (practice varies). On answering all 3 "correctly" I received a certificate. I had to answer some annoying 'commitment to change' survey questions that must be part of a (past? forgotten?) VCUHealth initiative. I received a link to an AMA PRA category one certificate that I downloaded, then I entered the CME at the ABFM site (they had entries for VCU).

A happy discovery. Thank you VCU and Curbsiders.

- fn -

[1] Starting from the CME site, it was weirdly hard to figure out what the heck VCU is. The logo is really small and the full name is never used.
[2] To be a AAFP fellow you need AAFP CME, but the American Board of Family Medicine accepts AMA Category one and their own programs. The two organizations don't entirely get along. 

Update 5/1/2023: I once completed a "knowledge feast" CME from the Curbsiders that spanned multiple episodes and was very (maybe too) time efficient. I believe that was a one time thing -- too easy to misuse I fear.

Saturday, December 17, 2016

Piketty's latest work on inequality is wrong about education.

The NYT has a readable summary of Thomas Piketty, Emmanuel Saez and Gabriel Zucman’s US income research. Much of it is familiar, but I was struck by this paragraph:

[since 1979] … Younger adults between 20 and 45 years old have seen their after-tax incomes flatline.

But over the same period, seniors in the bottom half have seen their after-tax incomes grow by over 70 percent. The bulk of that gain represents increased health care spending through Medicare.

Growth rates of a few percent a year do add up; health care is eating everything. Maybe it’s time to reread my old health care post.

Their findings are very important, but one of their recommendations falls flat (emphases mine) …

improving education and job training, equalizing distribution of human and financial capital, and increasing labor bargaining power, combined with a return to steeply progressive taxation

No, education and job training aren’t the answer. Roughly 40-50% of the US population has an IQ of less than 100. People with an IQ of under 100 have many skills, but they are not going to succeed in an academic program. Canada has the world’s highest “college” (includes 2 year vocational programs) graduation rate, and even they top out at around 56% of the population. I’m not sure why economists struggle with this basic arithmetic, my guess is they spend too much time with the cognitive elite.

What is the answer? We need to flip our thinking. We can’t change people to fit the work available in the natural post-industrial economy. We need to change the work to fit the humans. We need to incentivize work that is meaningful and rewarding across the cognitive spectrum. Germany did some of that by biasing their economy towards manufacturing. We can do some of that too (sorry Germany, that’s going to hurt you!), but we’re going to have to think more broadly. We’ll need to provide direct or indirect subsidies for work that’s productive even if it can’t compete with automation. We’ll have to apply work support lessons from the US military (long history of productive work across cognitive spectrum) and from traditional disability work support programs.

 

Wednesday, November 16, 2016

Mass Disability - how did I come up with 40%?

How, a friend asked, did I come up with the 40% number for “mass disability” that I quoted in After Trump: reflections on mass disability in a sleepless night?

I came up with that number thinking about the relationship of college education, IQ curves, and middle class status. The thesis goes like this…

  1. Disability is contextual. In a space ship legs are a bit of a nuisance, but on earth they are quite helpful. The context for disability in the modern world is not climbing trees or lifting weights, it’s being able to earn an income that buys food, shelter, education, health care, recreation and a relatively secure old age. That is the definition of the modern “middle class” and above; a household income from $42,000 ($20/hr) to $126,000. It’s about half of Americans. By definition then half of Americans are not “abled”.
  2. I get a similar percentage if I look at the percentage of Americans who can complete a college degree or comparable advanced skills training. That’s a good proxy for reasonable emotional control and an IQ to at least 105 to 110. That’s about 40% of Americans — but Canada does better. I think the upper limit is probably 50% of people. If you accept that a college-capable brain is necessary for relative economic success in the modern world then 50% of Americans will be disabled.

So I could say that the real number is 50%, but college students mess up the income numbers. The 40% estimate for functionally disabled Americans adjusts for that.

As our non-sentient AI tech and automation gets smarter the “ability” threshold is going to rise. Somewhere the system has to break down. I think it broke on Nov 8, 2016. In a sense democracy worked — our cities aren’t literally on fire. Yet.

Thursday, October 06, 2016

iPad High School: 1 out of 3 students return their iPad agreement

My daughter attends an urban high school. I think it is minority white (she’s not white).

Her school has a mixed reputation. It’s effectively segregated into an intensely academic non-black cohort and a low achievement largely black cohort. Students assaulted teachers at least twice last year.

The school distributes iPads to students. They are supposedly essential but half-way through the first semester they have yet to appear. The digital textbooks [1] the students use are designed for laptop use, they are not optimized for iPad use.

My daughter gets a lot of homework. It’s really too much, but the parents of the elite students are tigerish. Much of her homework requires an internet connection. Even assignments that could, for example, be done on a TI Calculator are better done using Desmos or Wolfram Alpha. Since the iPads haven’t been distributed yet her homework also requires a computer and thus WiFi service. And, of course, that internet connection with working WiFi.

At this school only 1/3 of the student body have bothered to return a signed document required to bring an iPad home. An iPad that, if one lacks WiFi service, is basically an expensive, fragile, and easily stolen doorstop.

This is not going well.

We need free universal urban net access. We can’t make this educational effort work without it. It doesn’t have to be high bandwidth. It doesn’t have to support high resolution video or allow YouTube access. It does have to be universal and free.

For now students would be better served by spending the iPad money on used older edition paper textbooks.

- fn -

[1] Want to give your child a large academic advantage? Order the low cost used paper textbooks from Amazon. Immensely better usability.

Friday, September 30, 2016

How does the world look to Trump's core supporters?

Set aside the neo-Klan-Nazi minority. Set aside the truly despicable - Coulter, Hannity, Falwell and the like. Forget the hell-spawn who think only of their personal wealth.

Think about the white non-college male voter:

… Trump’s fortunes rest on his core supporters, white men who lack a four-year college degree … He leads Clinton among them by 76-17 percent, an enormous 59-point advantage. That’s widened from 40 points early this month; it’s a group Mitt Romney won by 31 points -- half Trump’s current margin -- in 2012.

Whatever happens with this election, that 60% gap is a staggering fact.

How does the world look like to these non-college white men of the 4th quintile?

I have limited exposure to this cohort. A few Facebook friends — but they don’t post much about Trump. A family member with a cognitive disability claims to like Trump. That’s about it.

I need a journalist-anthropologist to falsify my story. I can’t help imagining a story though. It goes like this …

  • I have no hope of a secure economic future with savings, stable employment, good healthcare benefits and a pension.
  • I have limited marriage opportunities. I really miss the patriarchy. I feel that in my bones.
  • I watch Fox. I can understand it. Fox approves of me. Nobody else cares what happens to me.
  • I don’t understand economics, but I’ve lived through the past ten years. I’ve heard a lot of broken promises. Maybe nobody understands economics. Maybe the people who understand economics are lying to me. I definitely don’t understand borrowing from a wealthier future.
  • I don’t like academics.
  • I don’t care about the damned polar bears. I like warm weather. I don’t like bugs. I like motors and pavement. I don’t care about CO emissions.
  • I watch reality cop shows, where every criminal is a black man. I directly encounter crime and it’s always black men. I am afraid of black men. [1]
  • I don’t have a lot to lose.

None of this is going to go away. If we want to keep civilization going we need to give this cohort hope.

See also:

- fn -

[1] As a pedestrian and cyclist I am far more likely to be injured or killed by a white woman on a phone than a black man. FWIW.

Tuesday, April 19, 2016

US programs for physicians reentering clinical practice.

I’ve been out of clinical practice for about 17 years. On the other hand I have a current license and I did well on my board exams a few months ago. I believe I could do above average work, but I’d want a few months of supervision. I can think of a few ways to manage this, including designing a mini-curriculum and following the post-employment path of a fresh physician assistant. 

A formal program for physician reentry would be interesting, but there aren’t any in Minnesota. Which sort of suggests we don’t have a serious shortage of primary care docs. We do have some local fellowships that might be interesting reentry paths, but they currently fill well with post-residency candidates.

I haven’t decided to pursue this direction, but at a recent meeting I asked the Minnesota Medical Association’s president what he knew of. Their policy counsel made up a short list, and as I’ve not seen it online I’ll share it here:

There’s a longer list in a Federation of State Medical Boards PDF, many of the programs deal with “ethics”, “boundaries”, “disruptive behavior”, prescribing controlled substances, and “anger management”, but a few simply focus on “reentry” (though, on visiting the web sites, even they seem mostly to deal with what we call “disciplinary issues”). The Cedars-Sinai program seems closest, but it’s primarily focused on hospital privileges.

I haven’t decided to go down this path, but it’s helpful to know the landscape.

Tuesday, September 01, 2015

Family medicine board examination 2015: One last time.

I’m taking my family medicine board exams one last time. This is not entirely sensible. It’s been 21 years since I did family medicine, and 17 since I last saw a patient. I’m unlikely to practice again. The exam will be difficult; my brain is old and cruddy. (Long ago I did rather well on these, but it does help to actually practice medicine.)

Still, for one reason or another, I’m committed to doing the exam this November. I’ve slogged my way through the ABFM’s intricate preparations, including, for the non-practicing candidate, 6 self-assessment modules (SAMs) and one “alternative” module (which was quite awful and may have been since withdrawn). I even managed to meet the under-documented CME requirements [1]. The expensive Self-Assessment Modules varied from quite good to rather poor; alas the simulations are not worth continued investment [2].

When I’m done I’ll revise this post with what I end up with. By far the best guide I’ve found for someone like me was something written in 2008 (*cough*). I’m basically following my old recommendations (including ignoring audio CME/podcasts). For example:

  • ABFM | Exam Preparation: look for the tiny links at bottom to Study Tips and 2-week checklist. Some of this advice is wrong for me - and probably for most people. Still worth a read.
  • ABFM | Exam Content: this page hard for me to find - maybe my problem. I focus on topics with 5% and above.
  • I have old online medical notes written back when we thought HTML would be a good format for knowledge sharing and documentation. How naive we were! My medical notes started out in pen, moved to Symantec MORE 3.1, then FrontPage/HTML and now they’re back to an outliner (OmniOutliner 3). I have to write to learn. When I’m done I’ll attach a version of my notes here, but they’re really only going to be useful for me.
  • SAM Module Review: The SAM modules were a mixed bag, but the question explanations are superb summaries of current/expected knowledge. I’m mining those for my notes.
  • ABFM in training exam: The ABFM provides 3 years of teams. I’m studying these in depth, identifying any areas of strength, guiding my study, and generally awakening old memories.
  • Online references: this has changed, and not for the better. There’s much less available for “free” online than there was in 2008 [3]. Only Scott Moses’ self-funded hobby/obsession remains - the FP Notebook. So I’m buying selected paper references [4] like the venerable Washington Manual and Sanford Antimicrobial therapy. Some of my old textbooks (EKG interpretation) still work.
  • AAFP Board Review prep: skip over the expensive and inefficient modules and find the free (38 credit!) Board Review Questions. I think this is what the ABFM “exam prep” document was warning against. Needless to say, I’ll be sampling these, though Emily recollects they’re less useful than the ABFM in training exam materials.
  • Monthly Prescribing Reference (print version): still evil (drug money funded), still remarkably useful. Trick is to know what drugs are actually used vs. what are legacy — would be nice to have a version filtered by popularity.

I’m alternating topical work (reviewing Sanford, relearning EKG interpretation) with review designed to rebuild old memories. My medical knowledge network is frayed and fragmented, but there’s a lot of it. Much of my preparation is really resurrection. I've brute memorization ahead - reading, closing eyes, regurgitating. Then exam-guided note review and expansion.

It will be interesting to see how it all goes. Failure is certainly an option.

- fn - 

[1] Dear ABFM: Please note the current cycle progress tracker omits CME requirements but the future cycle includes CME requirements. Could be fixed.

[2] In the late 80s through early 90s we used to get 360K floppy disks each month with a unique DOS based medical simulation. I cannot, just now, remember what medical publisher did them (something Cardinal?). I remember them as quite excellent, I featured them in our residency computer-based training program.  Several clinicians, likely retired now, did some serious work on those. There really is no modern equivalent. Which is a kind of interesting.

[3] UpToDate is by far the dominant online resource for medical information — and it’s very expensive. (Priced for organizations.)

[4] See [2]. Also the movie Groundhog Day.

Update 11/22/2015

I don’t know my exam results yet, but it went more or less as expected. The test environment worked well — though it took me a while to realize I had to select text then click a secondary highlight icon that floated nearby to get highlighted text for review. My foot pulled a power cord out, but when our proctor fixed it everything worked. For most modules and sections there was ample time. We do get markers and writing material, I didn’t see that mentioned in the ABFM exam descriptions.

The AAFP board review questions and the ABFM provided residency training exam questions were a good guide. The AAFP questions generated CME credit, but I liked going back and forth. Whenever the questions exposed an unfamiliar topic I went off and did guided studying. There’s an AAFP board preparation page that provides USPSTF screening guidelines that’s quite useful.

FP Notebook was, and is, outstanding. I wouldn’t make many changes to my studying, but I’d have stared using FPN intensively sooner than I did. It is a perfect way to bring old knowledge on board, to identify obsolete knowledge, and to extend the sample exam question critiques. It’s also a great way to review medication information. Beyond FPN I mostly used American Family Physician articles and a handful of textbooks.

My key book references were The Washington Manual, Sanford antimicrobial therapy, MPR Prescribing Monthly, Emily’s ACLS “cards” and my venerable med school EKG primer. That’s similar to 2008, but since then my presbyopia has not improved. This year I found the app versions more readable than the paper versions; the Sanford and Washington Manual print text seemed microscopic. The Sanford app is a $40/year subscription, but Emily uses it too (we share same AppleID for purchases). My paper Washington Manual came with a free code for the electronic version, hosted by inkling.app (iOS). I barely looked at the paper manual, but I used the inkling version often.

The ABFM Maintenance of Certification examinations (I had to do 6!) were not particularly useful preparation for the board exam. I don’t think the maintenance exams a useful guide to real world practice either — they are much too esoteric. I regret the time and money I spent on those, but of course we don’t get a choice.

Even though most physicians would have key references like Sanford, Up To Date and drug information at their fingertips our board exams still rely on brute memorization. I suppose they’d be too easy otherwise. Given the years since I’d done family medicine (1994), and the years all by themselves, memorization was not particularly easy. I had to interact with information. I took handwritten notes on exam questions that I knew I’d never look at — but the writing process was important. I created my own spreadsheets of drug information for antimicrobials, antidepressants, anticoagulants, and, above all, oral diabetes meds (mercifully lipid therapy has gotten much simpler!). In each case I created my own groupings (med classifications), frequently reorganizing them. It helped to organize medications but their modern use, rather than by the sequence in which I encountered them. For the oral diabetes meds I would attempt to recreate my classifications by memory, then see where I got them wrong, then repeat… At about 30-40min of painful memory work a day it took 3 days to learn them.

I moved my old notes into OmniOutliner Pro 3 — itself an old piece of software. This was a ‘back to the future’ moment as my notes started out decades ago in a similar product - Symantec’s MORE 3.1. After a couple of weeks of study I could recognize what was worth keeping, what needed to be rewritten, and what should be deleted (much of lipid therapy!). I updated my old references with FP Notebook on a separate screen; this process helped tie old knowledge to new knowledge. 

The old knowledge was important — during the exam I found myself dredging things up from 25 years ago. I was surprised I could get anything from that far back, but really most of my medical base is from that era. For this exam I was largely refreshing and resurfacing it.

I’ll find out some weeks from now whether I need to do it again in a few months, or years, or perhaps never again.

Update 1/25/2015

I passed the exam by a substantial margin. If the maximal score represents percentile I scored about the 75th. That’s much lower than my score when doing academic practice, but it’s not bad for 20 years without patient care.

I did well in all the areas I studied. I didn’t do as well in obstetrics and gynecology, but I strategically omitted them.

My studying approach worked quite well for me.

Sunday, July 05, 2015

Why don't we have better medical textbook descriptions of disease?

I have a problem with the way medicine describes diseases. It’s not a new problem - it’s bugged me since about my third year of medical school - 1985 that is. I don’t think things have changed much.

To describe my problem, I’ll invent a disease “Y”.  Y has 3 findings - flat feet, pimples, and bad breath. Each occurs in 1/3 of the people suffering from Y.

What’s the probability that someone has flat feet and pimples given that they have Y?

No, it’s not 1/9.

I never said these were independent findings, and I never said they persist. Y starts with flat feet, then is asymptomatic, then patients develop bad breath and pimples. Yvians never have both flat feet and pimples at the same time. The old multiplication rule only works for independent events.

Ok, so you saw this one coming. Obvious, isn’t it? Well, yes, but most textbooks describe diseases as though they were collections of independent events without correlation or sequence or evolution or treatment effects. They provide long lists of symptoms, some of which rarely or ever coexist, followed by lists of findings, tests and so on. They very rarely, almost never, describe the long term sequelae of common acute conditions. (Often that’s because nobody has researched the “natural history” of the disorder. So the section would have to read “we have no idea” most of the time. That would be a start though.)

At the end of reading a classic textbook disease description you might be able to pass your Board exam, but you really have very little idea what the disease looks like, much less how it’s experienced by the patient. Sure, you learn that stuff after 5-10 years of patient care — but, really, that’s nuts.

Another way to learn this is to experience disease first hand - especially if you’re a 50+ physician. My own (one time) 1-2 day episode of vertigo didn’t match the textbook description of acute labyrinthitis or benign positional vertigo — it had features of both. It also left me with a subtle and persistent degradation of my balance - that doesn’t show up in textbooks either. Every 50+ physician can probably tell a similar story - our textbook descriptions of disease are misleading, incomplete, and frustrating.

I don’t think it was always this way. I have a 1930 edition of William Osler’s ‘The Principles and Practice of Medicine’ on my desk, what that book called “symptoms” was more the course of a disease. I wonder if the 19th century editions were even more case based.

We could fix this, but I never see anyone talking about it. We’d have to first admit we had a problem.

Friday, November 28, 2014

Learning new habits and skills - Pinboard, Simplenote, Toodledo and the skr tag

Being new-old means learning to be deliberative about things that were once opportunistic.

I used to get my exercise whenever I could — now I do regular group classes. Similarly, I used to pick up new habits and skills whenever the need arose or training required — but as of today i’m doing something more … deliberative. Something that should work better than my ad hoc approach of the past 30 years, not just because old brains suck but also because my old approach didn’t really work all that well. Even with a (relatively) young brain.

My new approach builds on 3 of my favorite tools: Pinboard, Simplenote and Toodledo — and a tag [1] of ‘skr’. The tag stands for “skills review”; it’s short because I wanted something fast to type.

I created a “habit/skills review” task in Toodledo and scheduled it a couple of weeks out. The task reminds me to review things tagged ‘skr’ in Pinboard and Simplenote. Meanwhile, as I come across things that I want to learn and make a part of my cognitive toolkit, I tag them ‘skr’. In Pinboard I save a Pin with that tag, in Simplenote I create a note with that tag.

Ever two weeks I get to the task and do my review. It only takes a few minutes. If I learn something I can remove the tag, if I’m failing to learn it I can take other measures or decide it’s not worth the investment.

I think this will work…

[1] Rant diversion — why are so many tagging implementations so awful? Why do so many devs exclude tag strings from full text search? Why does Simplenote display tags by data created rather than alpha sort? Why can’t one “merge” tags? Why … why …

Sunday, October 19, 2014

The relativistic version of the linear motion equation: velocity = acceleration*time.

My son is starting to do … physics.

It is fair to say I am more excited about this than he is. There is a reason I have a copy of the 1973 Edition of Misner, Thorne and Wheeler’s Gravitation ($35!) on my bookcase. I’m saving it for retirement, by which time Desmos.com will breeze through all the visualizations. It is true I was a failure (non-physicist) at Caltech, but there’s still time…

Ahem. So when #w was given an assignment to create a poster about linear motion, with equations on time, velocity, constant acceleration and the like, I thought it would be fun to show that the equations of linear motion are merely a low velocity approximation to the relativistic equations of linear motion. I was sure a quick Google search would turn up a simple simulation that would look like:

v=a*t (starting at rest velocity after time t is acceleration times t)

Riiiight. I’m sure that simulation exists, but I was never able to find it. With a bit of thought about the problem I realized I might be better off dropping the tricksy concept of acceleration and looking for a relativistic version of:

v= (F/m)*t (since change in velocity is Force*Time/Mass. Push a trike, push a truck, which moves faster?)

That found an “off-topic” [1] stack overflow article which was just a few parentheses short of the good-enough equation [2]: v = c * tanh(asinh((F*t)/(m*c))).

This is the first time I’ve personally run into hyperbolic trig functions, so I thought that was pretty cool, especially since I’d just read Jon Butterworth’s lovely description of how one hops from simple physics to quantum physics simply by tossing the square root of -1 (i) into a classical wave equation (uses Euler’s Formula, so extra points). That article used the familiar sin/cos functions, so I was getting an extra dose of trig. 

Plugging v = c * tanh(asinh((F*t)/(m*c))) into Wolfram Alpha gave me this cool output (yes, the AIs wll be our death, but for now they’re fun):

Screen Shot 2014 10 19 at 1 20 50 PM

Now that is what I was looking for! It clearly reduces to v = (F/m)*t when the squared stuff is relatively small. [5]

We can now compare my son’s high school textbook linear motion equation (v=F/m*t) to the relativistic equation, using the sneaky physics trick of geometrized units so c=1. Just to make things even nicer I’ll arbitrarily set the applied force to “1” (some unit) and the mass to “1” (some unit). This is what Desmos shows:

Screen Shot 2014 10 19 at 1 56 56 PM 

The last shows how simple this is in geometrized units (v is change in velocity over time t in our funky units):

Screen Shot 2014 10 19 at 2 10 11 PM

and here’s the magic graph that shows what happens as velocity (y axis) approaches “1” (speed of light) in the original linear motion equation and the relativistic version:

Screen Shot 2014 10 19 at 1 58 53 PM

Yeah, hits the speed limit. [5]

[1] The iron law of StackOverflow is that any article of interest will sooner or later by marked off-topic.

[2] I discovered the problem when I plugged it into Wolfram alpha, I had to do some web searches and play around the parens to get the correct expression.

[3] Incidentally, if you’d told me in 1995 that we would still lack easy entry of mathematical notation into web pages in 2014 I’d have assumed some kind of worldwide civilizational collapse.

[4] Once I had this expression, which was provided by the Wolfram AI, I found a discussion thread (scroll down) telling me it can be derived by “substituting F/m for a in the Baez equations(where m is the proper mass) and dividing by c where the tanh/sinh stuff is related to “proper and coordinate time”.

[5] I’ve done great violence here to the principles of general relativity — the meaning of time and distance are entirely dependent on frames of reference and I’ve glossed over all of that — largely because, you know, I’m not a physicist. It is kind of neat, however, what one can kludge together with a handful of web tools. In the general vein of non-physicist at play, it’s fun to compare this to Butterworth’s masterly representation of quantum physics using high school math. I again am left with feeling that physics will be easier to understand once we figure out what “distance” and “time” emerge from.

Sunday, June 29, 2014

Online textbooks are awful. It's time to kill the publishers.

My daughter and I are using “Holt” [1] Mathematics Course 3 for her summer math work. It’s quite a good printed textbook; a used 2007 edition cost about $10 on Amazon.

Her school doesn’t expect parents to buy the used textbook, however. They expect us to use the same material through Holt McDougal Online. Alas, unlike the printed text, the online textbook experience is miserable. Holt is serving up low to medium resolution bitmaps that are barely legible on screen or if printed. Our school district’s acceptance of this awful experience reinforces my fear of their iPad for all learning program. They are not ready for this.

It’s not just the schools that aren’t ready. The big publishers who control school textbooks have had decades to do computer based textbooks — and their products are still lousy.

We need alternatives to the traditional publishers. We need nations, states, provinces and startups to fund new textbooks that are digital from the start. This will kill Holt et al — but we have no choice. They can’t do this. They need to go away.

[1] Publisher names change constantly.

Monday, March 24, 2014

Encyclopedia Britannica lives - at least on iOS

As a kid I wanted to read the encyclopedia - from A to Z. We couldn’t afford an encyclopedia though, so I had to make do with the dictionary [3]. 

So when Britannica, which had more or less skipped the CD era [1], went on the web I was an early customer [2]. They struggled technically though, and by 2006 I dropped my subscription. After 2008 I mostly forgot about them.

It turns out they’re still around, supposedly making money [7] and still charging a $70 subscription for access to most articles (current news topics are often free). That’s a bit steep, especially since a link from our kids school gives me full access to the “High School” version. [5] (Alas, my ancient Britannica Dashboard widget [4] can’t be configured to use that URL.)

On the other hand, the iOS app subscription is only $15/year, and all devices for the purchaser’s Apple ID can use the subscription. On my kid’s parental controlled iPhones the entire content is accessible without authentication needed [8].

So I signed up [6]. For now I’m an EB subscriber/user again; hope they last a few more years.

see also

- fn -

[1] MacKiev World Book, by the way, is $30 on the App Store. I bought the DVD for $40 in 2012 - but I have to confess kids have not used it as much as I’d hoped.

[2] According to a 2005 post of mine, I used to prefer 1990s Britannica’s manually maintained index of web sites to Yahoo’s. This was back in the Alta Vista era. I know I was a paying customer from at least 2001-2006 - at about $70 a year.

[3] Much later I bought a complete set of the 1911 Edition. That is very cool browsing.

[4] Mavericks lets me assign multiple desktops to my secondary vertical display, so that’s where my Dashboard sits. When I want to park a doc in that display I swipe Dashboard away.

[5] I wonder if the school is supposed to post that redirect publicly. I don’t want to get them in trouble, so you’ll have to explore on your own. I can’t see any difference between High School version and public version.

[6] Auto-renews. Note you can manage these subscriptions from your iPhone - go Settings - Store and then Apple ID and tap around.

[7] Given the amount of broken stuff on their web site that can’t be a ton of money.

[8] I found some links that would take me to the web; those opened in Safari and were blocked. So no obvious backdoor.

Sunday, February 02, 2014

21st century market failure: what the rise and fall of Guitar Hero teaches about gamification

My oldest wants to learn to play drums. Learning is difficult for him, and the Smart Music program his school uses is obviously too sophisticated. We need something simpler, something more accessible, more like a game ...

Something like the Guitar Hero music education program I remember from a few years back. Fun, teach the basics, work with our Wii ... perfect!

Ok, I'll just Google that ....

Right.

Guitar Hero is gone. There is nothing like it any more.

Why Guitar Hero died News • News • Eurogamer.net (Feb 2011)

As the dust settles on Activision's decision to put an end to its world-famous peripheral-based music franchise Guitar Hero and the difficult work of sacking those who helped create it begins, one question remains: where did it all go wrong?

Only three years ago Guitar Hero shot through the $1 billion revenue mark – in the US alone.

Now, in what can only be described as a spectacular fall from grace, Guitar Hero is no more. Why? Why did Guitar Hero die?...

... "Guitar Hero was a victim of its success," said Wedbush Securities' Michael Pachter. "The game was incredibly well-conceived, the peripherals were great, and the music offering was deep and broad. All of those factors led to unprecedented success, and each contributed to its demise."

For Pachter, the fact gamers could play new Guitar Hero games with the peripherals they already owned proved to be the killer blow.

"Once people bought the band kit, for example, they didn't feel compelled to upgrade, as the one they bought was high quality and did the job well," he said. "Once people bought a game, they had 60 - 80 songs to master, and few mastered all of the songs offered...

... "There is absolutely nothing Activision nor anyone could have done to save the music genre. We should remember Guitar Hero for what it was, not where it's at now."...

... "It is possible that Guitar Hero will return, but a re-launch would have to be managed on a far smaller scale. Production costs would have to be minimized to enable profits on unit sales in the hundreds of thousands rather than in the millions."

Pachter's conclusion? "The franchise can support sales at the $200 million level annually, so it will still generate profits, but with license fees and manufacturing costs, margins are not that great, and certainly not enough to keep 200 - 250 people employed working on a new version each year."

So to recap - about 5-6 years ago we had a mini-cultural phenom -- a low cost high fun solution for music education. The Wikipedia article on gamification is written in 2010, around the peak of the Guitar Hero story. A few years later and it's all gone - the game, the console, the hardware, everything. In 2014 some replacements may slowly emerge on the iPad, but we're basically starting over again.

What's going on here -- besides our 21st century penchant for rapid cycles of creation, destruction, and recreation?

Maybe the root problem with gamification is that education doesn't have the economics, or the life cycle, of entertainment. Entertainment has visciously short lifecycles with massive floods of money. That can bring great products out quickly, but this amphetamine fueled growth has a cost. The entertainment products wipe out the weaker educational market -- and when Guitar Hero burns out there's nothing left to replace it. The education market has to be slowly grow back -- only to be wiped out again by the next cycle of the entertainment market.

Ultimately, the entertainment bubble is destructive, and the end result is a peculiar form of market failure.

PS. Garage Band is an interesting exception. It was clearly driven by Steve Jobs passion rather than any kind of business logic. It endures as a monument to Jobs, and because Apple doesn't have to put much money into it. It works, it's done, and the Mac platform is far more stable than entertainment-oriented consoles.

See also:

Monday, October 07, 2013

How to practice your French (or Chinese, English, etc) pronunciation on your iOS device. For free.

Do you have a lousy French accent?

Never fear, it's better than mine. What we need is a victim, a native speaker, who will type whatever we say. Someone who never tires and never complains. Someone who is poor at interpreting foreign accents.

Someone like your iPhone.

Here's the trick for an English speaker:

  1. In Settings, General, International, Keyboards enable French (or Chinese, English, etc) keyboard.
  2. Go to an app like Notes. Tap world icon to switch to French keyboard.
  3. Tap microphone icon and dictate your French phrases. Watch the iPhone get them completely wrong. Keep practicing until they come out right.
It's even easier than doing the same thing on OS X. Handy for kids learning foreign languages. (I'm certain Android does this as well.)
 
(Credit to Talking to Siri by Sadun & Sande for discussing this in terms of Siri control.)

Saturday, April 20, 2013

Learning programming for middle school - Python

We survived our monster spring break trip to Florida (example), so now it is time to think about how to make our children miserable this summer. For #3 it is math, for #1 I'm still thinking, and for #2 it is learning to program.

Program in what?

Python of course [1], as as discussed on app.net and almost universally identified as the best learning language [2]. It helps that I know the basics of it, and would enjoy learning more.

There are several options we can explore to help with this project:

We started out registering for the Python Coursera course from Rice University. I enjoyed it, but it's probably better for later high school or someone like me. The use of CodeSkulptor is interesting.
 
I suspect we'll go with Python for Kids and perhaps some Khan Academy and/or CodeAcademy supplements.

[1] Specifically Python 2.7.4 for Mac. The latest version of Python is 3.x, but when I did my Google App Engine tutorial at Strata we were told to use 2.7,  Coursera and most texts also prefer 2.7. On Mountain Lion you can install 32bit or 64bit versions, but the 64bit requires a TCL upgrade to run the handy integrated dev tool (IDLE) so I just went with 32bit. OS X ships with a version of Python, but it's worth getting the IDLE version. It's exasperating that the standard Mac Python distribution doesn't include an uninstaller; I wrote up some directions here after I foolishly installed Python 3.

[2] I suspect TurboPascal was the best ever, but it's no longer practical. Other contenders on the Mac environment are JavaScript and (yech) AppleScript. 

Update 4/21/2013:

We ended up starting with the free  Python 2 edition of Snack Wrangling for Kids. Not because it's free, but because it uses Python 2 (which imho is the best current version), and there's a PPC version of Python 2.7.4. The PPC version is desirable because we use an old G5 iMac as a "Learning Workstation"; unlike our other workstations there's no limit or authentication required for use of that machine. It's a good place to host the Python IDLE link and the PDF.

Although the language of SWFK is more for 8-10yo than our 14yo he doesn't mind it and the exercises build nicely.

See also29 common beginner Python errors on one page | Python for biologists

Math education if you don't care for everyday math and religious purity is not required

Medicine is barely evidence based, so it's not surprising that education is not.  Research is expensive, and kids are so variable that even evidence-based conclusions wouldn't work for everyone. In any case, there's no money to do true randomized well designed trials on math curricula.

So instead of evidence we get opinion-based fads. In the late Clinton era we got everyday math, which swept into Minnesota in the 00s. Now it's receding elsewhere, but we're still stuck with it. Obviously Everyday Math must work for some, but it hasn't done well with our three.

So, for this summer, we decided to do something different with #3 before she enters middle school. I started with the homeschool site curriculum reviews, but of course many of them are concerned with religious purity as much as education. I didn't see anything there I really liked. I wanted a well done textbook, but Saxon seemed too rigid and dull and Singapore Math too strenuous.

Next I tried places that I thought might do a good job with Math teaching, such as California and Ontario. I decided I liked Ontario's approach to 5th and 6th grade math best. They use the Addison Wesley's "Math Makes Sense" series. (I enjoyed the exercise animation that used hockey puck weights.)

I found used copies of the 5th and 6th grade books on Amazon for a few dollars each, so for about $12 total we have the material for our daughter's summer work.

Tuesday, October 23, 2012

Minnesota's Coursera eviction will be reversed

A few weeks ago Minnesota got some app.net attention. Alas, it was because we looked a bit silly ...

Minnesota Gives Coursera the Boot, Citing a Decades-Old Law - Wired Campus - The Chronicle of Higher Education

... The state’s Office of Higher Education has informed the popular provider of massive open online courses, or MOOC’s, that Coursera is unwelcome in the state because it never got permission to operate there...

... Tricia Grimes, a policy analyst for the state’s Office of Higher Education, said letters had been sent to all postsecondary institutions known to be offering courses in Minnesota. She said she did not know specifically whether letters had been sent to other MOOC providers like edX and Udacity, and officials there did not immediately respond to questions from The Chronicle.

But Ms. Grimes said the law the letters refer to isn’t new. “This has been a longtime requirement in Minnesota (at least 20 years) and applies to online and brick-and-mortar postsecondary institutions that offer instruction to Minnesota residents as part of our overall responsibility to provide consumer protection for students,” she wrote in an e-mail....

I asked my illustrious state representative, Michael Paymar, about this. He responded by paper letter (that's the way it works). Briefly the Office of Higher Education will work with legislators to change the law. The Director Larry Pogemiller said that nobody should bother registering and Coursera was fine in Minnesota.

Unsurprisingly this was a well intended law designed to protect students from old-style educational fraud, but it's obsolete now.

Sunday, October 07, 2012

Baumol's cost disease: medicine, education and post-AI disruption

William Baumol was born in 1922. In 2012, 90 years later, he's listed as first author on a new bookThe Cost Disease: Why Computers Get Cheaper and Health Care Doesn't.

Damn. It's one thing to win the brain lottery, but winning the longevity lottery is really piling on. Even if all he did is read the page drafts he's doing pretty well.

That's not the most irritating thing about Baumol though. The most irritating thing is that I keep forgetting about his fundamental insight, one that I first blogged about 8 years ago...

... The disparity between rapid productivity growth in mechanized sectors and slow productivity growth in human-service jobs produces Baumol's disease—named after the economist William J. Baumol. According to Baumol, in a technological economy falling prices for manufactured goods and automated services eventually increase the relative cost of labor-intensive services such as nursing and teaching. Baumol has predicted that the share of gross domestic product spent on health care will rise from 11.6 percent in 1990 to 35 percent in 2040, while the share spent on education will rise from 6.7 percent to 29 percent.

The shifting of relative costs need not in itself be a problem. If Americans in 2050 or 2100 pay far more (as a percentage of their spending) for health care and education than they did in 1900, they may still be better off—if they pay correspondingly less for other goods and services. The problem is that as the relative cost of services like education and health care rises, more and more Americans will find themselves in service-sector jobs that, unlike the professions, have historically been low-wage...

Today Education and Health Care are famously afflicted by Baumol's disease. Law used to be, but then full-text search decimated legal employment (and yet, legal costs have not fallen ....).

Baumol argues that even if these professions remain labor intensive, and even if health care comes therefore to claim 50% of our GDP, that we'll be able to afford it nonetheless.

His argument is persuasive, but is that likely to happen? College education today is experiencing widespread disruption including iTunes Ucoursera (Caltech, University of Toronto and many more), edX (MIT, Harvard, Berkeley), California open-source eTexts, Stanford Online, Khan Academy and numerous for-profit ventures. Education is deep in whitewater times.

Health care, particularly medical care, isn't changing as quickly. The fundamental tasks of sorting out what's going on with a particular patient, and how best to manage that problem in their personal context, and then how to manage the patient's psyche and health -- those haven't changed much [1] over the past century. 

We're accumulating more health care data though -- for better and for worse [3]. "Analytics" is the "hot" area in health care IT now, including running Google/Facebook style algorithms against large clinical and financial data sets [2].

That doesn't necessarily sound disruptive, unless you know that the techniques used in extracting meaning from large data sets are the same technologies that power our post-AI world. (Yeah, I used the forbidden acronym.) If you know that, then you know "Analytics" can be thought of as the current pseudonym for "Medical AI". Whether it's disruptive or not remains to be seen, but I suspect that we'll get to health care cost disruption well before health care hits 50% of a much larger future GDP.

 [1] It's interesting to read articles written in the 1970s during the early days of diagnostic lab testing. They imagined patients walking into a series of lab test queues staffed with low wage workers, then emerging with a set of diagnoses and plans. Similar plans arose during the last period of genomic enthusiasm. They will come again ... 
[2] The base stats is generally pretty simple stuff, if only because more complex algorithms don't scale well to terabyte data sets. The trick is that simple stats on large data sets enabled by cheap computation can produce surprisingly useful answers. This is best described in the terrific Halevy, Norvig and Pereira paper: The Unreasonable Effectiveness of Data.
[3] In 1996 I was part of a theater-style presentation called "Dark Visions: 1996-2010" that included a fanciful and intentionally dramatic timeline of dystopic data sharing. By 2005 India was the world center of clinical AI, and by 2006 elite health care providers had moved to more private paper records. Maybe we were a bit hasty :-).

See also: