Skip to main content

Searching habits and the value of rare words

One of the interesting aspects of working on search relevance is that it really makes you think about how to be succinct and specific in the expression of a concept. Looking at dissatisfaction reports with search engines gives a little insight into the challenges people face every day in expressing what they're looking for.

For example it's tempting -- and very common -- to describe something in general terms with the expectation it'll be easier to find. It definitely easier to find something that way but the chances of it being what you're looking for are generally worse. It's almost always better to think of the most rare characteristic of what you're looking for and search for that instead.

Of course search engines are trying to reduce the frequency with which this happens by relying more heavily on on context, history, and other things, but it never helps to be more specific whenever possible.

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Boeing 777: Connect any two cities in the world nonstop

From the description of the Boeing 777:

One of the newest members of the 777 family, the 777-200LR Worldliner has the capability to connect virtually any two cities in the world nonstop. It will carry more passengers and more revenue cargo farther than any other jetliner. Also, the 777-200LR can carry a full cargo load on routes where other airplanes are payload limited. This gives airlines the capability to carry the same number of passengers farther and with additional revenue-generating cargo.

Connect any two cities in the world nonstop. An immensely compelling yet succinct statement of ambition.

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Level 5 leadership

From a slide deck on Level 5 leadership, on the subject of unwavering resolve:

They possess inspired standards, cannot stand mediocrity in any form, and utterly intolerant of anyone who accept the idea that good is good enough.
• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Clarke's three laws

Clarke's laws:

  1. When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.
  2. The only way of discovering the limits of the possible is to venture a little way past them into the impossible.
  3. Any sufficiently advanced technology is indistinguishable from magic.
• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −
• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Weekly link roundup

Random things shared on Twitter from the week:

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

On diets: the easier the better?

The claim is simple: Drinking Water Before Meals Aids Weight Loss. By filling the stomach with water before a meal, the sense of hunger is reduced and consumption falls correspondingly.

The 2007 study shows a difference in the effectiveness of this method dependent on age; little effect under age 35 with increased effectiveness in the older population. Does this really just come down to increased risk in the older population or some other hidden driver?

I'd love to see a longer term study over multiple years on this. Could the simplicity of this approach set it up to be easier to maintain as a simple healthy habit as opposed to a forced, non-enjoyable dieting behavior?

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Trends in technology

Data points I saw in a presentation recently with some of my thoughts inline:

As late as the 1940s, the product cycle stretched 30-40 years whereas today it seldom lasts 30 to 40 weeks The product cycle also includes that phase of 'end of life' which is becoming similarly compressed. In hardware and software, the total lifespan rarely exceeds 4 years. I'm not sure how much this can be compressed but suspect that there's a limit at which the cost of familiarization/installation becomes a meaningful portion of the overall lifespan of the product, at which point it'll essentially flatten out.

80% of scientists, engineers and doctors who ever lived are alive today Population growth makes this a tiny bit misleading but it's still a staggering realization.  How many of those are in the US? How many 10 years from now?

All the technical knowledge we work with today will represent only 1% of the knowledge that will be available in 2050 Again this has to be balanced against the creation of an 'exabyte per day' but by then we'd better be surrounded with better research and mining tools than we have today.

The half-life of an engineer's knowledge today is only five years. In electronics, fully half of what a student learns as a freshman is obsolete by his or her senior year Which is why education in these fields should focus on problem solving and the ability to leverage prior research rather than learning by rote.

• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

Weekly link roundup

Random things shared on Twitter from the week:

  • Bing Touch and Decide: Introducing Bing for iPad
  • RT @ash_fontana: OH: You are what you don't automate.
  • RT @noah_weiss: Just found at Color has not a single PM.  That's the best advertisement I've ever seen for why PMs are useful.  Thanks!
  • Moving beyond teachers and bosses
  • Julius Irving: Being a professional is doing the things you love to do, on the days you don’t feel like doing them.
  • The Joy of Tech comic... Everyone's an "innovator".
• −    − •    − • •    − • − −     − − −    • −    • − •    • − • •    •    − • − −

'Gamifying' the system

From an NPR article on 'Gamifying' The System To Create Better Behavior:

Now, imagine that same camera also snaps a photo of your car when you are driving at or under the speed limit. For your safe driving, you are entered into a lottery to win a portion of the money from fines paid by speeders.

That idea was tested in Sweden with great success. It's an example of "gamification," considered the next wave of social engagement and Internet technology.

Instead of being structured around punishment and negativity, he says, the speed-camera lottery is "all about positive reinforcement." If you drive the speed limit, or under it, you may win some money.  "And that positive incentive to create better behavior," he says, "is a core tenet of games."

I think there's a huge opportunity here both for startups wanting to capitalize but also for the consequential cultural changes it can bring. I foresee a significant rise in the prevalence of strategies like this for getting people do the right thing by default.

At the same time it reminds me of another NPR story What A Coin Toss Has To Do With The Housing Market which talks about how humans are wired to be loss averse rather than calculating, probability engines:

He offered me a sweet bet on the flip of a coin. If the coin came up heads, I would win $6. If it came up tails, I would lose $1.

I told him I'd take the bet.

But then he changed the terms — if the coin came up heads, I would win $6. If it came up tails, I would lose $4. That bet I didn't like.

Of course, this is irrational. The bet is still very much in my favor. If I took the bet 1,000 times, I'd almost certainly make a nice profit.


As it turns out, our brains feel losses and gains unevenly: Losing feels worse than winning feels good.

So the real trick is how to leverage this these human traits rather than to work against them.