Posts

Early projections for 2024 based on previous Presidential and House returns slighly favor Republicans. These projections are completely unrelated to Biden’s recent polling numbers.

TL/DR Summary

1. BlueSky has a chance to dethrone twitter right now, but that path is narrow. 
2. Its exclusive invite only model means its user base is now small, elite, and homogenous with few bad actors. Almost everyone likes it. But the real test will be when it opens to the public. 
3. It is designed for true account portability and in theory should prevent a single company from owning the entire network as it scales up. 
4. However, it’s unclear if an ecosystem of small companies can do the job of content moderation in the same ways that centralized social networks do. The same is true of running modern feed-ranking and follow-recommendation systems. 
5. There will be growing pressure to make money using ads to cover costs as the network scales up, which will incentivize centralizing key data and resources, undermining the original model.
6. Future possibilities include: (1) BlueSky remains de-facto centralized, “in beta” until it can get composable moderation right, which turns out to be the foreseeable future; (2) big players (Google, Facebook) join the party and dominate the ecosystem; (3) small, unmoderated, ad-free apps proliferate and the network becomes overrun with spam, NSFW, hate, scams and gifts that come with a lack of moderation.

Pretty much everyone at Twitter—and especially Jack Dorsey—has long known that BlueSky could replace Twitter. When I joined Twitter in 2021, I soon learned our CEO was terribly unpopular internally, sporting a job approval rating under 40 percent, by far the lowest of any executive at the company.

Last Friday (2023-03-31) Twitter released what it calls “the algorithm,” which appears to be a highly redacted, incomplete part of code that governs the “for you” home timeline ranking system. And I saw nothing to suggest the parts of the code they put in the GitHub repository wasn’t authentic.

Do you remember the night of Nov 8, 2016? I was glued to election coverage and obsessively checking probabilistic forecasts, wondering whether Clinton might do so well that she’d win in places like my home state of Arizona. Although FiveThirtyEight had Clinton’s chances at beating Trump at around 70%, most other forecasters had her at around 90%.

Regression models are a cornerstone of modern social science. They’re at the heart of efforts to estimate causal relationships between variables in a multivariate environment and are the basic building blocks of many machine learning models. Yet social scientists can run into a lot of situations where regression models break.