Search This Blog


Tuesday, July 28, 2015

Autonomous Life and Death Decisions...

So the Future of Life (FOL) organization thinks AI-based weapons are bad and that they should not be developed.  According to the linked document above these weapons could include things like "armed quadcopters that can search for and eliminate people meeting certain pre-defined criteria."

The concern, apparently, is a "global AI arms race."

Interestingly, though, cruise missiles are explicitly excluded.  Cruise missiles, at least in the opinion of the FOL folks, don't involve "AI" even though they make their own targeting decisions (or at least not the right kind of "AI").

So Elon Musk, a signer of the FOL document linked above, thinks that one day "non-self-driving" cars will be outlawed...?"  Musk says: "I don't think we have to worry about autonomous cars, because that's sort of like a narrow form of AI ... It's not something that I think is very difficult, actually, to do autonomous driving, to a degree that's much safer than a person, is much easier than people think."

So I guess there are degrees of AI, narrow to what? "wide?"

What about the "ethics" of self driving cars?

Supposed a small, "unseen" child darts out from behind a parked car leaving no time to stop.  In the on-coming lane is a pregnant mom and another infant.  What does the "self driving car" AI decide to do?

If its "narrow" perhaps it simply runs over the child...

Or maybe it instead decides to kill the mom in the on-coming lane.

If its not so narrow maybe some quick facial recognition could be used to decide if any of the potential victims of the accident are "haters", religious folk, or other undesirables who there are too many of in the country.

How would you know if the AI wasn't making these kinds of decisions?

What if there was some sort of "disparate impact" on certain segments of the population relative to the decisions made by the self driving cars?

I don't really think a "global AI arms race" would be the problem...

Wednesday, July 22, 2015

Command Lines of the Future...

Lest anyone thought I was kidding when I wrote this recent post on how command lines are making a comeback, i.e., as you text your bank 'move $100 from checking to savings' to get things done, I now offer CSound.

You can see the details here at Motherboard.

So now, to compose and/or play music live I can write commands (or Lisp functions if you prefer).

Sure glad I took typing in 10th grade...

Sunday, July 12, 2015

Apple: The Bigger They are the Harder They Fall...

I have posted on the slow demise by Apple (see this and this).

Slowly but surely this has progressed.

I have not installed the latest OS X and hope not too.

Here is Forbes saying the same thing.

The bigger they are the harder they fall...

Tuesday, June 30, 2015

Where's the Objective Scientific Evidence for the Dangers of Diacetyl and Acetyl Propionyl?

I came across a blog called  In particular this post stands out:

Basically its a discussion about how a UK lab and company called Cloud9 tested an e-liquid called Five Pawns.

The crux of the article is a table that shows the following data:

vapemestoopid says: "Five Pawns told more than one company in the UK that their liquids were DIACETYL and AP FREE. Not to mention how many in the US? Not to mention numerous email replies to consumers found during a simple google search. Dating back to 2013 from US, Australia and Greece."

vapmestoopid supplies the following three images (copied from their blog - link above):

So here we see Five Pawns claiming no Diacetyl or acetone.  Now acetone is not Acetyl Propionyl (AP) - a miss spelling?  Perhaps, but by how - its all greek to me...

And here we see Five Pawns claiming no Diacetyl - no mention of AP.

And here is some hearsay from GrannyGrump.

So as best I can tell Five Pawns claims not to use Diacetyl - nothing about AP anywhere except in the mind of vapemestoopid...

Not surprisingly Five Pawns claims the Cloud9 results are baloney:

Here is the summary of Five Pawn's latest test numbers:

All but "Perpetual Check" have no detectable diacetyl according to their lab.

However, they do test positive in a number of flavors for 2,3 Pentadedione (AP).

What's interesting is the AP results all seem off by about a factor of between three and five.

So the conclusion of vapemestoopid and Cloud9 is apparently that e-liquids with "high" levels are dangerous because of this study, among others.

Indeed the referenced study cites other well known studies that have appeared elsewhere here and on Facebook and various vaping sites.  However this document also says things like:

"Regarding the combination of flavorings with propylene glycol and glycerin in aerosols, specific concerns exist for lung and cardiovascular function: 1) propylene glycol or glycerin alone may elicit pathophysiological and/or pathological changes in lung function; and 2) interactions between the effects of the individual agents may heighten toxicity."

Effectively all flavors and PG and VG might be bad in some way, also nicotine, for you to inhale.

The bottom line here seems to be that, "OMG, vaping certain flavors might be bad!"

So some arbitrary flavors are picked out - diacetyl and a variant of it, AP.

Why these?  They relate to the infamous "pop corn lung" incidents.



Please read these.  The bottom line is no one, not the FDA or the CDC, has any real idea what happened with these incidents.

So I think a couple of things:

1) Running around testing e-liquid and publishing results without documenting the process in specific detail seems stoopid.  For example, the UK lab (according to Cloud9 uses "[the] same independant UK-based lab used by Trading Standards." - whatever that might mean) and the US labs, one would hope, would express how they do their work and how they calibrate and handle their processes, samples and equipment.

In addition one would hope that multiple samples and multiple labs would be used in a double blind study to ensure that both properly calibrated test samples as well as commercial samples are effectively intermixed during testing.

(This kind of testing is what a vaping trade association should do. My suspicion is that many are duped into paying money for bogus processes and results.)

2) The problem, as stated here on this blog as well as in the Five Pawns blog is this: "Further, we feel that efforts to translate industrial exposure limits to vaping exposure limits are flawed. It is clearly not the same.  If this were true, one would expect a population of individuals becoming sick from vaping, but this is not the case (underline mine). There are no known publicly documented cases of anyone having respiratory issues related to vaping AP or diacetyl at the levels currently in e-liquids.  Many websites and blogs discuss this exact issue.  We are confident that studies and future data will show inhalation from vaping e-liquids should not be compared to industrial exposure limits." - basically there has to be a reason vaping and smoking are not killing people due to DA or AP.

Until this issue is scientifically vetted panicking and screaming "Danger Danger" about DA and AP is really scientifically bogus and very misleading for consumers.

Is there a risk? Of course.  There is risk in breathing and in smoking.  Is smoking a greater risk than vaping?  Is it in the best interest of the consumer to continue smoking until, decades from now, all this is resolved?


3) Testing results for DA and PA can and do vary over the life of an e-liquid bottle (again from the Five Pawns blog): "We have funded extensive independent research, including inhalation and heat studies, and have plans to begin in-vitro research as well, looking at the effects of e-liquid vapor on lung tissue.  In our research, we have also discovered that the effects of storage conditions and time on the shelf can also affect the variability of test results (underline mine).  Therefore, we have also initiated long-term stability and degradation testing, which can take up to two years for results."

So testing not only has to be calibrated relative to amounts but it must be applied comparably across testing environments and samples relative to the "age" of the liquid, exactly how the testing is done, and so on.

4) There are many other questionable ingredients like cinnamon which no one seems concerned with yet are demonstrably and objectively "worse" than DA or AP in terms observable of impact on body tissues in humans.  Selecting a small subset of some of these flavors (like DA and AP) but not others is also very misleading for consumers.  Yet I see no hew and cry for their removal from e-liquids.

The bottom line here is really simple:

Cigarettes are known to be significantly more dangerous than virtually all forms of vaping at this time.

DA is already in cigarettes (and vaping for that matter) at comparable levels to the "panic" levels described by vapemestoopid and clould9 yet apparently neither causes OB.

Without an scientific explanation of this fact its really disingenuous to make claims DA or AP are dangerous.

One must also consider the danger of leaving people on cigarettes until the "safety" of vaping is confirmed.

This is like telling children to wait in the burning second floor room until a proper, government approved ladder is available to rescue them.

Panic mongering is not good for anyone...

Monday, June 29, 2015

More Google "Don't Be Evil..."

In May of 2012 I wrote "Google's Waterloo, Patent 6,061,520."

Google's claim was that an API was basically unpatentable.

It looked like Google would be taking Oracle to the cleaners by freely using its Java APIs.

But in a surprise move the US Supreme Court has declined Google's recent appeal on the matter and has thus breathed new life into Oracle's claims (see this WSJ article).

As I wrote before I believe its completely preposterous for Google to claim that an API is not a copyrightable item.

According to the WSJ article: "Google had asked the Supreme Court to hear the case and limit how software makers could use copyright law to assert exclusive rights over computer programs. It argued Oracle shouldn’t be able to claim copyrights on basic software commands."

Really - anyone can just take software written by anyone else just because they invented "basic software commands?"

Sure sounds like a money grab on Google's part to me....

Saturday, June 27, 2015

The Future (Past) of UI Design is Auto Correct Madness!

I found this article at Wired.

Basically the idea is that all these apps and HTML5 web page user interfaces are tedious and difficult to use relative to "texting."

Why not, the article postulates, use texting to interface to, for example, interact with your bank?

After all, everyone constantly texts and its a familiar interface.

So why not just select "My Bank" on the text destination menu and say "move $20 from account A to account B?"

After all, according to the article, 97% of people use texting on their phones - more than any other app...

I am reminded of the quote by George Santayana "Those who cannot remember the past are condemned to repeat it."  Or, perhaps more accurately, my corollary "Those who are repeating the past are condemned not to notice."

Apparently the future of smartphones is the command line.

After all, what does texting your bank do but create the notion of a command line interface.

I find this extremely interesting.

Forty years ago while first becoming introduced to "computer science" the command line was all there was.

And it was considered the largest impediment to non-computer-literate users using a computer.

You heard about "arcane command structures" and so one.

No normal person could or would use this...

Thus the notion of a "User Interface" or UI was born.

Why not invent something like a "mouse" or a "touch screen" to make things easier.

And so, for forty years, that's what we do...  Spend probably trillions of dollars on graphics, faster chips, touch screens (don't forget the lost "stylus years"), and on and on.

Until we arrive again at the notion of "command line..."

Of course, if my iPhone is any indication of what Apple can do for texting look no further than this site:

There was a lot of research in the ancient computer past of what a good command line might look like.

As far back as TOPS-20 there was the TENEX command line interpreter (see this).

It worked hard to remove ambiguity and confusion from users typing commands.

The problem with texting, of course, is that the domain space is unlimited where as with a command line interpreter the commands you can conceive of and use are bound by the set of commands the computer actually understands.

To me its like the attempts to replace something like "Emacs" as an editor (emacs descended from TECO).  Minimal command interface.  Once you learned it well the commands would spring from your hands without thought.  And much, much more efficient that having to type, stop typing and grab a mouse, then let go of the mouse and type, etc.

I was always amazed in the late 80's after PC's became popular how much less efficient a kid that had grown up on PC's was at editing than a good emacs user.

The bottom line here, unfortunately for the new "command line" texting model, is that for command lines to work efficiently you have to have a domain of knowledge where the next portion of the command can be inferred from context.

For example, many modern OS's today let you type something like:

  copy a

where what is inferred after you type the letter 'a' is a choice of all files starting with 'a' - then, of course, there's some mechanism to pick which file you are interested very quickly.

Texting there really isn't much to infer from.

So if I text Bob "let's meet at a" is 'a' a set of restaurants, hotels, car dealers, etc.???

Now maybe we could try and guess but most likely in that case we would end up in the land of Apple auto-correct madness.

Now I don't know what the best UI idea is, but apparently I've already seen it because we are starting at the beginning again.

(I doubt we'll have everyone carrying around punch cards for the smart phones...)

Sunday, June 21, 2015

Ten Years of Contrails...

Almost ten years ago I wrote one of my first blog entries (see

The point of the article was that apparently no one had considered or measured the effects of jet contrails on surface temperatures (think global warming science).

Now I did not discover this, it merely reported the results originally discussed on the BBC program Horizon.  Researchers had figured contrails matter but were unable to prove this until 9/11 here in the USA when all commercial jets were grounded for several days.

So, merely ten years later, climate researchers are now looking into this issue (see this article Science Codex).  From Science Codex:

The researchers report that the "diurnal temperature range was statistically significantly reduced at outbreak stations versus non-outbreak stations." In the South, this amounted to about a 6 degree Fahrenheit reduction in daily temperature range, while in the Midwest, there was about a 5 degree Fahrenheit reduction. Temperatures the days before and after the outbreaks did not show this effect, indicating that the lower temperatures were due to the contrail outbreaks.

"Weather forecasting of daytime highs and lows do not include contrails," said Carleton. "If they were included in areas of contrail outbreaks, they would improve the temperature forecasts."

The underlined part (my underline) is quite interesting.

I my experience daily weather forecasting is, at best, abysmally unreliable.  One has to believe that 6 degrees is quite a bit to miss out on (see this and this as well).

I wonder what impact this has had on the various temperature stations used, at least in the US?