I’ve started to post raw notes from events I attend. For more information, see the Notes Policy.
From O’Reilly’s Web 2.0 Expo.
Clay Shirky: “It’s Not Information Overload. It’s Filter Failure.”
- “How fast is the amount of info growing?” Exponential curve. We love this chart.
- For 15 yrs we’ve been hearing abt “info overload”
- “Info Overload” is the normal case. Why do we treat it as extraordinary?
- Guternberg introduced info overload. It was shortly after the printing press that there was more info available that a person could read in a lifetime.
- Printing (high up-front costs) press also introduced risk: what if books don’t sell?
- There’s no necessary reason to have the same person specialize in 1) publishing books (a trade), and 2) selecting books to publish (literary curation). But there is an economic reason.
- Solution: publishers become the filter
- All mass media (sound recording, television) shared this economic model: high up-front costs to be recouped over time.
- The internet introduced post-gutenberg economics: no more high upfront costs.
- the filter for quality is now way downstream from the site of production
- This is filter failure, not information overload
- Every so often there’s a bunch of spam that gets through, in higher numbers than usual. Clay measured his periodic spam increases over time; it wasn’t an increase in spam volume, it was filter failure.
- Spam is a good example: multiple filters (human & machine), its never over (arms race), the volume always grows
- This is not bad design. Its a (social) system failure
- An ITP colleague broke off her engagement; it came time to change her FB relationship status, but she wanted to proceed with discretion. She realized her Facebook status change would be a billboard. She tried to use the privacy controls, but the “I’m now single” message still went out to everyone. If ITP kids don’t get the interface, who will? FB tried their best. Who’s at fault? The problem is that managing privacy prefs is unnatural for humans.
- “privacy” controls information flow
- Our notion of privacy is not moving from one engineered system to another; it’s moving from evolved systems to engineered systems.
- “personal life” isn’t a phrase we use anymore
- privacy used to be built on (in)convenience. That was a feature, not a bug
- how can we design the filters so privacy works the way we want them to?
- new story: 18 yr old creates study group on FB, gets charged w cheating.
- FB doesn’t have another media metaphor; FB is FB
- small groups defend against free riders, large groups don’t. They’re free-rider tolerant, not free rider resistant
- this isn’t a fight over info, its a fight over flows
- we don’t have obvious tools to fix filter failure; that’s why “info overload” is recurrent alarm
- if you have the same problem for a long time, maybe its not a problem. Maybe its a fact.
- old filters have broken for structural reasons.
- The fix will have to do with rethinking social norms