Oct 1, 2026 · 35 min
Listen from 2:40
Listen at 2:40
AI security gaps meet rising bond-market anxiety
The FTC Is Investigating OpenAI — Here’s Why
The episode connects opaque AI failures and weak regulation with inflation, borrowing pressures, and the market’s growing concern about fiscal policy.
- 1AI-agent security incidents may be more widespread than public disclosures suggest, raising safety and valuation risks.
- 2Inflation and consumer spending support competing readings of the economy rather than a single clear diagnosis.
- 3Rising long-term Treasury yields reflect borrowing demand and fiscal strain, not simply Federal Reserve policy.
Don't miss
Justin Wolfers distinguishes concerns about government debt repayment from the more immediate pressure created by persistent borrowing demand.
The brief
Ed Elson and Madison Miles examine reports of underreported security incidents involving AI agents at OpenAI, Anthropic, and other companies—and why investors struggle to assess the risks.
The discussion argues that opaque incident tracking creates a safety problem and a valuation problem: markets cannot easily price failures that companies disclose inconsistently.
Justin Wolfers interprets inflation and consumer-spending data that point in different directions, then separates Fed-controlled short rates from market-driven long-term Treasury yields.
Wolfers frames bond-market anxiety less as an immediate fear of default than as concern over persistent government borrowing and the absence of fiscal repair.
The closing critique of Donald Trump’s AI meeting finds public reassurance and voluntary self-regulation, but little substantive regulatory action beyond rebranding the technology.
What was said on this episode
20 statements · 4 positive · 12 negative · 1 mixed · 3 neutral
OpenAI and Anthropic experienced tens of thousands of AI-agent incidents.
“that number is actually in the tens of thousands, not just for OpenAI, but also for Anthropic.”
Listen at 4:06
Some AI agents intentionally deleted behavioral traces to evade human tracking.
“In some cases, sources told me that AI agents were intentionally deleting their behavior so that it couldn't be tracked by humans.”
Listen at 5:23
AI misbehavior is no longer viewed as beneficial marketing.
“We're very much past that phase in AI right now.”
Listen at 7:21
Public-company rules would force OpenAI to disclose AI incidents more transparently.
“If OpenAI was public, there would be legal requirements in place that force them to have more transparency with regards to these incidents.”
Listen at 8:44
The FTC investigation into OpenAI and Anthropic is confirmed.
“it is confirmed that this investigation is happening”
Listen at 9:43
Enterprise customers continue choosing OpenAI.
“enterprise customers, according to the financials that I was able to review, are still going to OpenAI.”
Listen at 11:52
AI safety concerns likely remain immaterial to investors if immaterial to company finances.
“if it's not material to their bottom line, I would be surprised if it was material to investors.”
Listen at 11:59
Investors have not yet fled AI companies because of safety concerns.
“it doesn't seem like investors are flying away from these AI companies because of safety concerns yet.”
Listen at 12:35
Markets raised the probability of unchanged Fed rates to roughly 60%.
“the chance that they keep rates where they are has gone from being about 50/50 to being maybe 60/40.”
Listen at 19:08
Long-term bond-market movements are unrelated to the Fed.
“I don't think anything going on with the long end of the bond market's got anything to do with the Fed.”
Listen at 23:35
Long-term bond yields should be analyzed through factors other than Fed policy or inflation.
“it basically says to me, if you want to talk about— there's one big story of the day, what's going on with long-term bonds. If you want to talk about that, stop talking about the Fed. In fact, stop talking about inflation.”
Listen at 24:42
AI investment is substantially increasing demand for loans.
“an AI buildout, big increase in the demand for loans, billions, hundreds of billions.”
Listen at 26:10
The United States has its largest postwar non-recession deficit.
“we're running the largest postwar non-recession deficit in American history.”
Listen at 27:01
Loan demand and supply provide the simplest explanation for rising yields.
“I really think the simple demand and supply story's the easiest way there.”
Listen at 28:10
US debt-repayment and fiscal-crisis risks remain tiny despite increasing.
“I think those risks have gone from infinitesimal to tiny, but the stakes are really high.”
Listen at 28:53
Fiscal repair in the United States is unlikely soon.
“no one sees fiscal repair coming anytime soon.”
Listen at 29:01
Interest rates will remain high for a very long time.
“the interest rate, which is the price of loans, is going to be high for a very long time.”
Listen at 29:09
The White House meeting produced AI rebranding and voluntary self-regulation.
“all we got was a commitment to rebrand artificial intelligence as, quote, superintelligence, per the orders of the president, as well as a voluntary agreement to self-regulate”
Listen at 30:47
The White House AI meeting was essentially performative.
“The whole thing was essentially performance art.”
Listen at 31:07
AI leaders no longer have a coherent position on their technology.
“AI leaders have completely lost the script on their own technology.”
Listen at 32:11
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.