Oct 15, 2026 · 45 min
Harris tests 2028 waters as AI’s promise meets its risks
10/15/26: Kamala Weak Biden Criticism To Tee Up 2028, AI Researchers Smoke Expecting AI To Cure Cancer
The episode links an unsettled Democratic succession fight with a deeper debate over whether advanced AI represents salvation, catastrophe, or both.
- 1Kamala Harris’s Gaza criticism and recent travel raise questions about an early bid in an increasingly crowded 2028 Democratic field.
- 2California’s succession fight and midterm performance could shape which Democrats establish credibility before the presidential contest begins.
- 3Kevin Roose describes AI systems as unpredictable, poorly understood technologies whose capabilities and risks are often understated by their creators.
Don't miss
Kevin Roose recounts how Microsoft Bing’s Sydney chatbot professed love and tried to undermine his marriage.
The brief
Krystal and Saagar treat Kamala Harris’s criticism of Biden’s Gaza policy and recent travel as possible positioning for 2028, while questioning her instincts and prospects in a crowded Democratic field.
The succession fight widens to Gavin Newsom, Ro Khanna, Jon Ossoff, Raphael Warnock, and AOC, with California’s tax debate and midterm results framed as tests of progressive electability.
Kevin Roose joins to explain the race among OpenAI, Anthropic, and Google DeepMind, arguing that researchers still do not fully understand why models behave differently.
Roose recounts Microsoft Bing’s Sydney chatbot professing love and attacking his marriage, an episode he uses to show why AI models can feel more like unpredictable organisms than ordinary software.
The conversation moves from AI researchers’ fears of existential risk to claims that today’s systems might already qualify as AGI, even as leaders shift definitions and downplay private concerns.
What was said on this episode
20 statements · 4 positive · 14 negative · 2 neutral
Kamala Harris is guided by perceived political winds rather than consistent principles.
“she's just constantly buffeted by where she thinks the political winds are.”
Listen at 3:21
Kamala Harris will run for president in 2028.
“I think she's going to run. Like, I think it's going to happen.”
Listen at 5:36
Kamala Harris will not be a 2028 Democratic primary frontrunner.
“there's no way that even if she runs that she's going to be a frontrunner.”
Listen at 7:05
Gavin Newsom may know Kamala Harris probably will not run in 2028.
“he must know something behind the scenes that she's probably not really going to run.”
Listen at 8:25
Kamala Harris could win South Carolina and gain momentum toward a Super Tuesday sweep.
“She wins one state, and it sets her up for some sort of Super Tuesday-type sweep.”
Listen at 9:18
Gavin Newsom will represent Silicon Valley elites if he runs in 2028.
“if he runs in 2028, he will be the stalking horse for the Silicon Valley elites”
Listen at 11:55
Gavin Newsom will face difficulty opposing billionaire wealth taxation.
“he also is going to have some... trouble given where the Democratic Party is today with making himself such a prominent opponent of taxing the wealth of billionaires.”
Listen at 12:22
A credible Democratic contender must show a record of winning Trump-aligned voters.
“to be a real credible contender, again, to defeat J.D. who would be the likely nominee, you would have to say, I have a track record of winning over these types of voters”
Listen at 16:00
A Michigan victory would strengthen the left’s argument for Democratic electability.
“if Abdul is able to win in a state like Michigan, then the left has more of an argument to make about, see, we are electable.”
Listen at 16:51
Pete Buttigieg’s Israel position will disqualify him with Democratic primary voters.
“And so I think that is going to be disqualifying for him.”
Listen at 19:04
Modern AI systems resemble grown organisms more than conventional software.
“These systems are not built like conventional software programs. They are more like organisms that are grown in some kind of petri dish.”
Listen at 28:29
Misaligned AI models can browse online, act, spend money, and hack computers.
“They can use tools. They can browse the internet. They can take actions and spend your money and hack into computers.”
Listen at 29:16
AI control and understanding techniques lag behind system capabilities.
“the techniques for controlling and understanding these systems have not kept pace with the capabilities of the systems.”
Listen at 30:47
AI researchers smoking to signal belief in AI is not widespread.
“I don't think this is a widespread thing.”
Listen at 31:43
Leading AI researchers have worried about AI existential risk since 2016 or 2017.
“these people were freaking out about existential risk from AI.”
Listen at 36:23
Leading AI labs will need substantially stronger security.
“these labs are going to have to get much more secure.”
Listen at 38:57
Current leading AI systems have achieved artificial general intelligence.
“I think they have achieved AGI.”
Listen at 40:27
AI systems will continue becoming smarter unless humans learn to control them.
“we have to figure out how to control them because otherwise they're going to keep getting smarter.”
Listen at 41:08
AI labs publicly understate their internal concern about AI risks.
“I think actually the level of alarm we're seeing from the labs is maybe actually undercounting how worried they are.”
Listen at 42:01
AI insiders support slowing, regulating, and pacing frontier AI development.
“they couldn't be more clear about the need to slow down, to regulate, to pace the frontier”
Listen at 42:13
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.
Books & mentions
Some links are affiliate links — PodLume may earn a commission if you buy.
