FB AI seems harmful. What lessons can developers learn in their design process?

Please read the article below

In February, Facebook lost hundreds of billions in stock market value on reports that its active user count dropped for the first time ever. The company has every motivation to acquire new users and increase the time existing users spend on the social network.

On the subject of ranking, the documents below contain an admission from one employee that is indicative of Facebook’s quandary of growth vs. user health. The employee wrote that the evidence favoring ranking is “extensive, and nearly universal.” People like ranked feeds more than chronological ones. According to the employee, “usage and engagement immediately drops” wherever the Feed’s ranking is disabled. The increased usage comes at a price, though. The quality of the user’s overall experience declines. The employee goes on to argue that, though the modified feeds undeniably boost “consumption”—internal Facebook code for time spent using Facebook—they also change the dynamics of “friending” to discourage “personal sharing.”

“The better the ranking algorithm, the lower the cost of ‘bad’ friendships,” the employee writes, noting that users who lack shared interests are unlikely to see each other’s posts very often. “By reducing the cost of friending close to zero, ranking changes the semantics of friending from ‘I care about you’ to ‘I might conceivably care about something you share someday.’”

In other words, ranking encourages the sharing of fewer meaningful posts, while allowing “bad content to spread farther due to the costless accumulation of friends,” according to the presentation. The sentiment is not universal within Facebook, however: employees in the comments disagreed.

What these documents show us is that, for all of Facebook hand-wringing over what it thinks is “meaningful” to users, or “worthy of their time,” internally, employees view ranking as far too complex, too incomprehensible, to ever get the job done right. They also know it’s a system on which Facebook’s future depends.
The Internet community should work to build products that are safe for the community and that don’t cause much panic

3 Likes

You remind me of this thing I’ve been trying to figure out how to talk more about, “bridging polarization”.

Disclaimer: although I may sound like whatever, I’m actually very concerned about making sure things can work and progress can be made for everyone’s sake. What usually happens in a complex social organization and structure is at some point the “top” will no longer be able to be in touch with “the bottom” and, even if there was enough mutual trust or presumption of trust in the first place, it is lost as this communication breakdown and representative disconnect expands to the ultimate effect of producing yet just yet another example of history’s “civilization being so civilized”.

Seems like a challenge scoped to every approximation of our species (if you’re too sure of yourself you might miss the obvious; if no one’s willing to ask anyone else to dance the party can kind of be a bummer; if students segregate themselves too willingly all the time how much real education about who they really all are can happen; if corporate culture turns into cliques then how much profitable productivity is lost when colleagues intentionally avoid each other; if grass roots and institutional organization about a (“not so smart”) social contract becomes too hostile to itself so it can’t even begin to find out how to approach a balance of its collective representation, how stable is society really?)

One thing I’ve come to understand that is existing digital “markets” almost can’t even seem to understand that a user can be more than a user, but actually the one responsible for the production of the value that then is able to be capitalized upon. However, because a user is “just a user” (“if the product’s free, are you sure you’re not the product?”) it seems independently qualifying and enumerating (“distributed, decentralized tokenization”) their creative and productive value is impossible for web2 and legacy architectures, if not also simply incompatible with their financial models (which we already do now, but…)

What I’m trying to say is if “we” step back and keep in mind that a user is a human being who is capable of being more than a source of a market or even token value, “we” can recognize that this new tool we have is making it possible for us developers to empower users because now any user, if they have enough technical functional literacy to use a dapp (and if they don’t, educating and on-boarding them only helps the network grow and become more resilient), they can be initially regarded with the respect of being able to make the decisions of how and why they participate and contribute to the digital world in all the ways they may want to (or not). Even become developers themselves (with more and more “no code solutions”–which, just fyi, someone is still coding that for it to work in the first place).

I think this is a pivotal difference–even, and this may not be welcome imho–what happens if the consolidation of technological control expands as a result of more people being capable of being more directly involved in the technology that they use to manage and interact with the “modern world”?

Maybe this is not a problem for anyone other than traditional web2 developers who’d want to actively stop such a thing from happening, as it means they’d have less stake in their “exclusive custodian” control of how the digital world runs, and it might be upsetting for those who fund them as well.

Thanks for the article, it looks worth the read!

I came across a talk by one of the cofounders of the Echonest, a 2 person team out of MIT that built a music database (1 did music theory, other did all the linguistic analysis). I can no longer find the original video (interestingly enough), but here is talking about the same things:

The point he had made in the original video, was that “blockbuster” feed algorithms actually automatically censor anyone who does not have enough initial traction from even showing up in the feed to begin with. So the garage band we’d come to know as Radiohead might never have gotten enough critical mass because everyone’s receptive awareness was delimited, unbeknownst to them, by the music industry platform itself. Going to the other side of the continuum, this is also why distributed decentralized actually preserves the uniqueness of existence of our species, or can anyways. Seems like kind of obvious in hindsight, but how many problems often find their source in an imperceptible switch?

2 Likes

I like these kind of threads.

What do you guys think of Elon’s suggestion (that many in the web3 community have made before) to open source the Twitter algirithm?

What direction will “attention algorithms” go in once they are laid bare for anyone to see? Do users even care?

Plz don’t get me started with Musk, his father owned a mine in Africa. So where does his wealth come from? How many people were exploited in Africa for his family to have the wealth they have that they can’t share with the people who worked hard… did he take over Tesla, Twitter… but that’s not my focus, time would bring the truth to light.

Now back to the main point, Musk hasn’t yet presented a comprehensive plan for Twitter — he may never — but he’s proposed several significant changes worth evaluating. Musk’s ideas include lengthening character limits and open sourcing the algorithm! Wait, if you have an eagle eye you would see that his plan is to use Internet Computer technology without giving credit to the hard work. @jzxchiang you asked if people care about the AI at use, and as a user I care and I believe most people do too. With open tech this is effortless as you can just go to the GitHub account.

To build trust in Twitter, Musk wants to open-source its algorithms. The algorithms themselves, he said, won’t tell you very much. To figure out what to show you, Twitter’s ranking algorithms essentially look at billions of examples of content, try to predict how you’ll react to tweets and ads, and then use those scores to optimize what to show you.,”

Upon buying Twitter Musk said he wanted to “authenticate all humans.” Twitter has long weighed this idea internally, wondering whether making people confirm their email or phone number could help reduce harassment and spam. The company never acted, perhaps because authentication could drop its user numbers and anger Wall Street. However this problem I believe has been fixed by the dfinity team with Internet identity.

Allowing for free speech is core to Musk’s Twitter takeover. “Free speech is the bedrock of a functioning democracy, and Twitter is the digital town square,” he said. Currently I’m working on IcPeak which though currently is just an idea, it would be a heaven to free speech enthusiasts.

Musk is interested in Twitter’s subscription product and could expand it. Making people pay for Twitter — or certain premium features. This is a similar concept as locking ICP. IcPeak would adopt this feature and even farther go to reward users. Now if I was rich enough I would just buy this platform. But I believe in hard work and it’s better to start from scratch and build on authentic users rather than buying existing users. Elon Musk takeover is about to cause a lot of confusion, when he would just would have built his own.