We’ve been using a new system in the Lab for a few months now, and it has really captured our imagination. The system, called ‘Curriculum,’ is a real-time stream of topics from Lab members’ web browsing activity. So for example, right now, the latest topics in Curriculum are:
- ‘DHT humidity/temperature sensors’
- ‘3.3V i2c interface’
- ‘PIR thermometer device’
…these topics are generated by a semantic analyzer that reads the content of the page and infers the topics that the page is about (in this case, I was just researching some sensors for another project). Of course, there’s a healthy amount of weird noise as well; here are some less-intelligible topics that were also browsed recently:
- ‘deepest thoughts’
- ‘G2108 G2110 G2111 G2112 G2113 G2116 G2124 G2125 G212C’
- ‘insane flow’
Even with some noise, the feed is fascinating enough that it has quickly become a habit for all five Curriculum users to check the feed quite regularly. Checking the feed is rewarding because it is always at least a little bit funny (the imperfections of semantic analysis make for some great robot poetry), and it often affords a deeper or more intimate view into what my colleagues are working on.
Click through for a more in-depth look at Curriculum and how it was designed.
We often try to live in the future here at the Lab, and Curriculum feels like a particularly authentic fragment from the near future: it lacks the one-sidedness of other technologies whose promise is relentlessly positive, purporting to ‘optimize’ or perfect an activity by listening at a higher resolution or over a longer timeframe. Rather, Curriculum tries hard to be ‘flat’ in its delivery, treating the superficial and the substantial with equal care; the aim is to augment our intelligence with an additional stream of information, not to over-process or summarize that information. (For example: filter, prioritize, etc. We’ll still use our brains for that.)
This doesn’t mean that the Curriculum feed is, or should be, the end of the line: Curriculum is designed as a resource, rather than a product, so we can hang other experiments off of the feed (more on this very soon!).
Origins of Curriculum
Curriculum began as an experiment last year that applied the ‘Quantified Self’ philosophy to online habits: Jer and I made a browser extension (called ‘Semex,’ for ‘Semantic extractor’) that presents your web history as a series of sessions and topics rather than URLs and timestamps. For example, one session might be a fifteen minute period this morning where I was researching the topic ‘humidity sensors.’
Semex was useful when it helped me remember where I was in a problem that I hadn’t worked on in awhile: seeing the sequence of topics I browsed made much more sense than a list of page titles. At the same time, Semex felt anticlimactic in the way that a lot of Quantified Self projects do: there was a sense of OK, we recorded all this; now what?
Then, an idea: if Semex was most useful to me as a way to record my cognitive context, the state in which I left a problem, maybe I could share that state with other people who might need to know it. Sharing topics from my browsing history with a close group of colleagues can afford us insight into one another’s processes, yet is abstracted enough (and constrained to a trusted group) to not feel too invasive.
Channels / Bandwidth
Curriculum, then, is essentially Semex for groups. Each user in a group has a Chrome extension that submits pageviews to a server to perform semantic analysis and publish a private, authenticated feed. (I should note here that the extension ignores any pages using HTTPS, to avoid analyzing emails, bank statements, and other secure pages.) Curriculum is carefully designed to be anonymous; that is, no topic in the feed can be traced back to any one particular user. The anonymity isn’t perfect, of course: because there are only five people using it, and because we five are in very close communication with each other, it is usually not too difficult to figure out who might be researching a particular topic.
Making a tool that looks at a group’s browsing activity is interesting because it adds a new channel of information available to all of us. The context we all share as a result of the Curriculum feed allows us to tap into each other’s information diet with little overhead: we can quickly get a detailed sense of exactly what problems we’re Googling, what topics we’re researching, and who is shopping for Christmas presents. Augmenting ourselves with this extra dimension expands the bandwidth of our interactions.
Privacy / Intimacy
When we think of ‘Future of Work’ scenarios, it’s easy to wax hyperbolic about the potential for employers to intrude on one’s privacy: biometrics, firewalls, CCTV and the like make for easy jokes about the nightmares of tomorrow’s workplace. However, the flip side to privacy is intimacy: how can close-knit working groups like mine do even better, how can we build on top of the high-bandwidth interactions we already have with each other and add valuable new dimensions and superpowers to ourselves?
In this light, Curriculum is kind of like a Fitbit for context, an effortless way to record what’s on our minds throughout the day and make it available to the people who need it most: the people we work with. The function Curriculum performs, that of semantic listening, is fantastically useful when people need to share their contexts (what they were working on, what approaches they were investigating, what problems they’re facing) with each other.
Reactions / Future
The Curriculum feed is truly a new channel of input for us, a stream of information of a different character than we’ve encountered before. Having access to the residue of our collective web travels has led to many questions, conversations, and jokes that wouldn’t have happened without it. Part of Curriculum’s appeal is the “flatness” that I mentioned before: there is no attempt to extract “useful” topics, or throw away “noisy” results.
Stay tuned for more on this–I’ll be writing up the first few things we built off of Curriculum in a few days.