Why the Texas gerrymander was overdue

Proportional representation what-ifs, the Mothership vortex, my op-ed on Apple's threat to polling, why Dems hate their party, polling young voters is hard, working class media habits, GPT-5

No. 370 | August 8th, 2025

🇺🇲 2026

The political world is abuzz over redistricting again.

A quick reminder here: when it comes to redistricting, there is no principle. There is only power. And that’s true of both sides.

The New York Times compares how the presidential margin in the Texas GOP’s newly proposed districts stacks up against the old ones.

This is the map that the GOP should have pushed through in 2022. But instead, in the post-Census redistricting, Texas Republicans played it safe, shoring up incumbents with ultra-safe seats that meandered from the suburbs deep into the rurals. That’s because they were spooked by the huge population growth numbers being put up in places like the North Dallas suburbs gradually turning these areas blue. Any suburban seat was shored up with a hefty dose of rural and exurban Republicans to survive a decade of California transplants moving into the Texas suburbs.

This now looks like a miscalculation. Not only did Texas see a huge shift right among Hispanics, but the blue shift in the suburbs looks to have been a one-time Trump-era adjustment that’s already started to be counteracted by a wave of Republican migrants from blue states. That means we won’t see Democratic gains in the suburbs as far as the eye can see, which is what Republicans were looking to guard against with 20+ point buffers for each of their suburban incumbents.

You can see these dynamics at play on the very outer fringes of the Dallas and Houston suburbs. Yes, these were one the very few areas to shift blue. But that’s a function of R+20 subdivisions replacing R+40 farmland. And the subdivisions have way more population, so the GOP’s raw vote advantage grows.

In the Rio Grande Valley, the previous round of redistricting didn’t do very much to maximize the GOP’s growth among Latino voters. Republicans traded a winnable TX-34 (which it briefly won in a special election) for a Lean R TX-15. The new maps improve these districts on the heels of Trump wins in all three seats. But they’ll need more than the normal buffer to overcome the region’s notorious incumbent advantage. Mixing up territory, so that Democrats Henry Cuellar and Vicente Gonzales were running in more unfamiliar territory, might have been the smarter move here, but the region’s geography makes major changes hard without eating into TX-23.

The Texas redistricting won’t create a massive new Republican advantage in the House, writes Nate Cohn.

Crystal Ball looks at how Texas redistricting will affect the median House seat in 2026:

Here’s where Republican and Democratic voters are over- or under-represented compared to a purely proportional map in each state:

Split Ticket tooks at the likely impact of candidate quality in the 2026 Congressional battleground districts:

How Mothership Strategies raised almost $300 million for scam PACs, but less than 3% of that saw its way into Democratic campaign coffers.

🗣️ Public Opinion

I’ve got an op-ed in the The Washington Post on an issue I covered in the newsletter this week: iOS 26’s potentially detrimental effects on political polling:

When Congress first acted to protect consumers from unwanted telemarketing messages — the Telephone Consumer Protection Act of 1991 — it exempted survey research from many of its requirements. This exemption was further upheld with the creation of the National Do Not Call Registry in 2003. Why? Because policymakers recognized that bona fide survey research, by definition unsolicited, was in the public interest in a way that sales and marketing aren’t. And by far, the best way to do this research is to reach out to any citizen through the most widely used messaging platforms available.

Polling texts and calls are a vanishingly small share of all messaging traffic. They do not seek to inflame partisan passions, engaging Democrats and Republicans at what we hope are equal rates. They play an essential role in surfacing what voters actually think, at any level of government.

A new AP poll shows Democrats view their party as “weak” and "ineffective.” (Also, a great Sankey viz.)

🇺🇲 2024

Good discussion on why polling young voters is hard: those very interested in politics (and who respond to polls) look very different politically than the majority who don’t.

Ryan Burge has the latest in his series of postmortems on the 2024 vote among different religious groups. To no one’s surprise, atheists and agnostics are a core Democratic base group. Religious identification continues to be a strong proxy for political attitudes.

📰 Media Habits

Democratic researchers are continuing to dig into why working class voters are moving away from Democrats. New data highlights the media consumption of these voters.

💻 Artificial Intelligence

Noah Smith looks at the potential for the AI data center construction boom to crash the economy if capacity is overbuilt. Tech companies are already turning to private credit to finance construction. Where we really need to start watching out is if they turn to the banks, where a collapse could have echoes of 2008.

GPT-5 was released this week. The always worth-reading Ethan Mollick got early access. What stood out to him was its ability to lots of different things in response to one prompt.

And that extends to the second most common problem with AI use, which is that many people don’t know what AIs can do, or even what tasks they want accomplished. That is especially true of the new agentic AIs, which can take a wide range of actions to accomplish the goals you give it, from searching the web to creating documents. But what should you ask for? A lot of people seem stumped. Again, GPT-5 solves this problem. It is very proactive, always suggesting things to do.

I asked GPT-5 Thinking (I trust the less powerful GPT-5 models much less) “generate 10 startup ideas for a former business school entrepreneurship professor to launch, pick the best according to some rubric, figure out what I need to do to win, do it.” I got the business idea I asked for. I also got a whole bunch of things I did not: drafts of landing pages and LinkedIn copy and simple financials and a lot more. I am a professor who has taught entrepreneurship (and been an entrepreneur) and I can say confidently that, while not perfect, this was a high-quality start that would have taken a team of MBAs a couple hours to work through. From one prompt.

👫 Demographics

Another look at how first names correlate with party registration:

The phones are killing our personalities.