Veterans for Political Innovation is a 501(c)(3) nonprofit with a mission to mobilize veterans and supporters to advocate for election innovations to make our political system less toxic and more competitive.
Site Navigation
Search
Latest Stories
Join a growing community committed to civic renewal.
Subscribe to The Fulcrum and be part of the conversation.
Top Stories
Latest news
Read More

Texas Democratic Senate candidate James Talarico speaks at an American Federation of Labor and Congress of Industrial Organizations rally at The Texas Gulf Coast Labor Federation on October 21, 2026 in Houston, Texas. As a former union member, Talarico spoke of the fight between billionaires and workers.
(Photo by Danielle Villasana/Getty Images)
Are Independent Voters About to Flip Texas?
Oct 23, 2026
Texas is in the middle of one of its most contentious and heated elections in modern history. The question that lingers in the national press is, will the state flip to Democrats in the major statewide contests for US Senate and governor?
But the voters who will decide this question aren’t Republicans. They aren’t Democrats. As previously predicted on IVN, it will most likely be independent voters—if they turn out.
- YouTube youtu.be
According to a new poll from the University of Houston Hobby School, 60% of independent voters support Democratic US Senate nominee James Talarico, while 32% support his Republican opponent, state Attorney General Ken Paxton.
Overall, Talarico narrowly leads Paxton, 48% to 46%.
In the governor’s race, Democratic nominee Gina Hinojosa does not poll as high among independents as Talarico, but they still prefer her to Republican Gov. Greg Abbott, 55% to 36%.
Overall, Abbott leads Hinojosa 49% to 46%. Both are tight races.
“High gas and diesel prices, inflation at the grocery store and an unpopular war all have created headwinds for Republicans. More than half of Texans say they have an unfavorable opinion of President Trump, just two years after he carried the state by more than 13 points,” said Renée Cross, executive director of the Hobby School.
“But it has been more than 30 years since a Democrat was elected to statewide office in Texas, and many voters remain loyal to the Republican Party.”
Here’s why this matters.
According to VoterCount.ai, which pulls from the L2 Data voter file, there are more than 2.7 million registered voters in Texas (just over 15% of the electorate) that identify as independent, compared to 48.6% that are Democrat and 36.2% that are Republican.
When IVN first reported that L2's figures showed more Democrats in Texas than Republicans in August 2025, the comments and emails stacked up from people accusing L2 and the Independent Voter Project, a co-publisher of IVN, of making up the numbers.
They pointed out—correctly—that Texas does not register voters by party. But L2 explained its methodology (which has set an industry standard for data modeling) and noted that a modeled electorate doesn’t always represent the electorate that shows up to vote.
While people may argue over these numbers, a bigger question was raised in IVN’s original reporting: Has Texas been a purple state this whole time? Or, at least in the last few election cycles.
Look at how close statewide contests have been recently, especially in US Senate contests. US Sen. Ted Cruz won re-election with 53% of the vote in 2024. US Sen. John Cornyn won re-election with the same percentage in 2020.
That is not an overwhelming majority. These races are getting tighter and tighter.
Now, look at the polling from 2024. According to CBS Austin, the same pollster that shows Talarico leading among independents this year found that Ted Cruz had more support from independent voters compared to his opponent in the last cycle, 48% to 42%.
Republican candidates had strong support from independent and Latino voters in 2024. What if it wasn’t Republicans who ensured Cruz kept his seat, but independents?
Now, these voting blocs—independents and Latinos—favor the statewide Democratic candidates, especially Talarico, who already showed what impact courting independents can have in the March primaries.
Talarico received more than 1.2 million votes in the Democratic primary. He beat US Rep. Jasmine Crockett, who received over a million votes. It is the first time since the 90s that turnout in a non-presidential statewide Democratic primary exceeded Republican turnout.
Turnout in the 2020 Democratic presidential primary was higher than the Republican primary, but President Donald Trump still won the state.
In the IVN analysis following the primary, it was noted that the way voters turned out had a few possible implications:
- It could be an indication that there are in fact more Democrats in Texas than the common narrative would have people believe;
- It could be a preview of what may happen in November;
- Independent voters may have driven up the numbers for Democrats; and/or
- Some Republicans may have crossed over to vote in the Democratic primaries
It could also be a combination of some or all of the above.
It is important to note that Talarico didn’t just campaign for Democrats. He reached out to independents as well, which seems to have paid off. The last public survey before the primary election showed his support among these voters at 62% compared to Crockett’s 35%.
Independent voters are allowed to participate in the state’s public primary elections. They can choose between a Republican and Democratic ballot. And Talarico won.
Now the question is, will his continued favorability among independent voters turn into votes in November? Could observers see the same in the governor’s race? Only time will tell. However, one thing is clear:
Independent voters are now impossible to ignore in the Lone Star State.
Are Independent Voters About to Flip Texas? is republished as part of The Fulcrum’s partnership with IVN and the Latino News Network to expand voter‑first journalism and strengthen public‑centered coverage of democracy.
Keep ReadingShow less
Recommended

AI chatbots are always guessing, even when they sound sure. Experts and Claude itself explain the AI trust gap, bias and why to check sources.
Oscar Wong/Getty Images
AI’s Information Crisis: Why Trust Is Harder to Sustain Than Ever
Oct 23, 2026
In all the debates about human and machine intelligence, one thing has become clear: We humans love to talk about AI. We’ve speculated whether AI will create or eliminate jobs, revolutionize or destroy science, comfort or torment us, entertain us brilliantly or rot our brains, or some combination of all these things. But amid the noise, we may have paid too little attention to a subtler impact: how AI is disrupting our relationship to information and making it harder to know what to believe.
The problem isn’t just deepfakes, the malicious manipulation of images, video, and audio. Even benign, well-intentioned AI applications can be profoundly misleading. The nature of the large language models that power chatbots means that they are always guessing, however confident they seem. How can we use AI to produce reliable information that improves public discourse rather than distorting it?
To explore this dilemma, I moderated a webinar on Trust in Information in the Age of AI, co-sponsored by NAPA, CODE, and the Bridge Alliance, with four leaders in the field: Richard Coffin of USAFacts, Professor Renée Cummings of the University of Virginia, Meghan Herbst of WIRED, and Dr. Alan Shark of NAPA and George Mason University. I also interviewed Claude (Opus 5.5 Extra), which assessed its own limitations and offered some useful advice.
The human experts and the machine intelligence agreed on several points.
A significant AI trust gap exists. Several surveys show that more Americans use AI, but few trust its results. However, many of us still have no firm opinion on AI, suggesting that trust in AI will depend on how it performs in the future.
Our skepticism is justified: AI often gets things wrong. Large language models don’t actually “know” anything – they answer queries based on patterns they deduce from the data they’ve been trained on. While accuracy rates depend on the task and how you measure success, rough estimates hold that chatbots make errors about half the time. The general consensus is that these AI hallucinations can be reduced but may be impossible to eliminate.
AI is often biased in ways that can be hard to detect. Algorithmic bias has been a concern for years, long before today’s chatbots were developed. Because AI is trained on historical data, it can reinforce old biases about gender, race, or ethnicity in hiring, criminal justice, and other areas. This can be a serious issue in any AI application used for public policy.
AI doesn’t quite know how to use government data. Chatbots often rely on secondary sources even when more reliable primary government data is available. The problem stems from the fact that much government data isn't “AI-ready,” and AI models aren't designed to go to government sources first, even when they are fit for use.
AI’s impact on information needs regulation. The debate over AI regulation has focused on the big picture: the need for guardrails to ensure AI doesn't turn on or even destroy humanity. While that priority is hard to argue with, we also need regulations to prevent AI from misleading or misinforming us. The European Union now requires a watermark on images or text produced by AI. In the U.S., several states are moving to prevent bias in public uses of the technology.
AI may offer real benefits for all kinds of human endeavors, but its flaws remind us to treat the technology with caution and care. This is especially true for using AI in public policy and decision-making, where poorly informed plans can have major repercussions.
For now, the best advice is to use AI with an awareness of its shortcomings, don’t put human judgment aside, and ask for and review your chatbot’s sources. As Claude said about our interview at the end: “If your readers take away one habit from it, I hope it’s checking the sources I’ve cited rather than taking my word for any of this.”
You can watch the full video of the webinar “Trust in Information in the Age of AI” on YouTube at: https://napawash.org/event/napa-technology-leaders...
Joel Gurin is the president and founder of the Center for Open Data Enterprise.
Keep ReadingShow less

Universal basic income would cost trillions. Wage insurance, a bigger child tax credit, and a tax break for work could help workers left behind by AI at a fraction of the price.
Andriy Onufriyenko / Getty Images
AI’s Costs: The Human Element & the Response
Oct 23, 2026
The age of AI will be a revolutionary age. In medicine, engineering, and more, there is the possibility of great benefit to humanity. Consider some Wall Street Journal headlines: “The Quest to Use AI to Help Find New Drugs,” and “Fusion Startup Gets an AI Boost.” Already, AI is a substantial portion of GDP growth.
Yet a rising tide will not lift the boats with holes in their hulls, as Jack Kemp once said. Artificial intelligence is challenging our leaders to think about those possibly left behind. Within the AI community itself, Anthropic CEO Dario Amodei expresses concern that we could see 5-10% GDP growth with 10% unemployment.
We cannot allow this concern about AI’s human costs to go totally unaddressed. “Pull yourself up by only your own bootstraps” is not an appropriate response when there is so much disruption throughout the entire economy—when the owners of capital will thrive as workers and would-be workers struggle. Ideas from both political parties are necessary.
A universal basic income is one possibility being pushed, but the cost may be prohibitive. A decent-sized UBI of $12,000 a year would cost trillions annually. A more realistic approach is needed.
There are two alternate proposals that have not passed Congress in the past, but they could be brought back up for consideration. First is the expansion of a concept called “wage insurance.” The second is an expanded child tax credit. I will contribute a third idea too.
Wage insurance was part of the Trade Adjustment Act’s safety net. A worker 50 or older who lost a job due to foreign trade qualified if the worker found another job paying less (but no more than $50,000). The benefit was half the difference in pay between the old and new jobs, up to $10,000 over two years.
The Obama administration proposed eliminating the trade-impacted requirement. Anyone over age 50 would qualify for wage insurance so long as they were in their previous job for at least three years. The cost was estimated at $3 billion a year. Unlike a universal basic income, Congress could find the savings to offset this spending.
Wage insurance directly addresses the immediate costs of an AI economy. The worker whose career is disrupted by what was only, until recently, science fiction finds a benefit other than only unemployment insurance. The worker may very well never have the old income level again and is approaching retirement. Help for such a person in the throes of capitalism’s “creative destruction” makes sense. We are addressing those in the spaces between what was and what is becoming.
I mentioned an expanded child tax credit as well. Former Sen. Mitt Romney proposed a kind of refundable and advanceable child tax credit—meaning paid even to those who do not owe taxes, and paid monthly before even filing a tax return—of $4,200 per child under age six and $3,000 for other children. This contrasts with the $2,200 per child, with less refundability, under President Donald Trump’s “one big beautiful bill.”
Sen. Romney proposed paying for the cost partly through tax reforms that largely, but not entirely, shift the tax code’s other credits and deductions in a more progressive direction. The “mostly” is why anti-poverty advocates seemed to give the bill two cheers instead of three. Overall, the proposal would net reduce childhood poverty but would cause some families lifted out of poverty to slip right back in.
Still, the Romney proposal could be a countervailing force against AI’s corrosive effects. Even more so if the bill’s pay-for is tipped in an even more progressive direction. We would have a kind of modest universal basic income for families with children. We would be addressing those of greatest need in the economy.
The third suggestion is this: Provide a tax break for work equivalent to the size of what we give to capital. The lower income tax rate for dividend income costs the federal government a little over $40 billion a year. Putting similarly-sized tax relief into workers’ pockets would produce an average tax savings of about $235 per worker.
An extra $235, or for a two-earner family, $470, is not a universal basic income, but it is affordable. The more progressive could very well connect the revenue that can be raised by eliminating the dividend preference to offsetting the cost of lowering the 10% bottom income tax rate, or some other way to put the money in pockets. A more ideologically conservative approach would seek the curtailment or elimination of tax breaks and government programs.
What is important is recognizing that the owners of capital will be great beneficiaries of the AI age. As a matter of public policy, we should look out for labor too. Patch the boat’s holes and let the tide rise. The human cost of not doing anything needs to be addressed along with the fiscal cost of doing something.
Scott Miller is a graduate of Widener School of Law, a former chief of staff in Congress, and the author of 'Christianity & Your Neighbor's Liberty.
Keep ReadingShow less

A Social Security inspector general found $1 billion in improper SSI payments tied to earnings alerts left unresolved for two years. Here’s what went wrong.
JJ Gouin/Getty Images
$1 Billion, 733 Days: Inside Social Security’s Improper-Payment Backlog
Oct 23, 2026
Social Security's inspector general opened its books this week and found a number that should stop any controller cold: roughly $1 billion in improper Supplemental Security Income payments, made to 333,800 recipients, sitting on top of a backlog the agency has known about for years. The cause was not a hacker, not a sophisticated fraud ring, and not a policy loophole. It was an earnings alert nobody closed out.
Here is how the mechanism actually works, and why it matters more than the headline figure. When a recipient's reported earnings change, the Social Security Administration's system generates an alert flagging that the payment amount may need adjustment. An employee is supposed to resolve that alert, verify the wage documentation, and correct the payment before the gap compounds. The inspector general's office pulled a sample of 100 such cases. Only 20 had been resolved. Sixty-three had gone untouched entirely, left pending for an average of 733 days, just over two years. Seventeen more were marked "cleared" without the wage documentation ever being fully verified, meaning the case was closed on paper while the underlying question remained open.
Multiply that failure rate across the full caseload and the agency's own auditors landed on 417,291 recipients with alerts still pending, and roughly $1 billion in resulting overpayments — $344 million of it still recoverable, and $664 million gone for good under the administrative finality rules that bar the agency from clawing back money once too much time has passed. That second figure is the one a controller would flag first. It is not merely money the government failed to collect. It is money the government forfeited its own legal right to ever collect, because the paperwork sat too long.
Put the number in context. The SSI program pays out $5.8 billion a month to roughly 7 million recipients, many of them elderly, disabled, or with no other income. A $1 billion improper-payment figure is a rounding error against annual outlays in the hundreds of billions, and that is precisely the trap: individually forgivable, structurally significant, and consequent, cycle after cycle, on the same unresolved bottleneck. Michelle Anderson, the assistant inspector general for audit who signed the report, put it plainly: the agency "has the opportunity to strengthen its controls to ensure employees timely and accurately resolve earnings alerts." That is auditor-speak for "someone stopped doing their job, and nobody noticed for two years."
This is where the story becomes bipartisan rather than partisan, and that is worth saying explicitly. Nobody designed a policy to let $664 million walk out the door permanently. No administration, no party, no piece of legislation set out to create a 733-day queue. This is what happens when a caseload grows faster than the staff assigned to work it, and when a backlog becomes background noise rather than a crisis, because it never generates a single dramatic headline of its own. It simply accretes, report after report, until an inspector general adds it up and the total is substantial enough to notice.
The agency's own response is instructive, and to its credit, unresisting. SSA agreed to implement all three of the inspector general's recommendations: review and correct the 63 unresolved alerts, review and correct the 17 incomplete verifications, and identify what is actually driving the pattern so it stops recurring. That last item is the one that determines whether this report changes anything. Reviewing the specific 63 cases is a one-time fix. Diagnosing why alerts sit for two years at a time is the only step with any chance of preventing the next inspector general's report from finding the same thing with a different case number attached.
There is a broader lesson here for how Washington measures accountability. A single agency losing $664 million to its own processing delays will not move a federal budget measured in trillions. But add up every program running the same kind of quietly aging backlog — disability claims, unemployment insurance overpayments, veterans' benefits recalculations — and the aggregate is not a rounding error anymore. It is a pattern of institutions that are diligent about writing rules and much less diligent about enforcing them against their own inertia. A pragmatic fix is not more legislation. It is adequate staffing against the alert queue that already exists, tracked with the same rigor a private-sector controller would apply to any accounts receivable aging past ninety days.
Congress does not need a new program to address this. It needs an oversight hearing that asks SSA a narrower and more useful question than "how much fraud is there": how many earnings alerts are open right now, how old is the oldest one, and what is the staffing plan to close the gap before the next $1 billion accrues. That is a question with an answer a controller could evaluate in an afternoon. Whether anyone in Washington asks it before the next inspector general's report comes due is the real test of whether this one changes anything.
Jose E. Navarro, MBA, is a financial controller and founder of The Navarro Report, a public finance and government accountability publication based in San Diego, California.
Human-Directed AI Journalism: This article was researched and drafted with AI assistance under the direction, review, and editorial control of the author.
Keep ReadingShow less
Load More
















