Joe Weisenthal
Co-host, Odd Lots podcast (Bloomberg)
aka @TheStalwart
“in traditional software, open source. One nice thing about open source is you don't have to pay for it. So it's free. It's a little bit different when we're talking about there's no really such thing as like free AI software. Because even if it's like free, you still have to pay for the depreciation of the chips and you have to pay for the electricity to run them. So there is no real such thing as like free open source AI software.”
“are the open source models cheaper on a per unit of intelligence basis? If we had some way of saying levelized cost of intelligence, which I don't know if the industry has yet, are open source models cheaper per IQ point?”
“as you use it more, it's just like everything else, the treadmill of expectations. Here's some tasks that you can do in 30 seconds which maybe several years ago would have taken you 30 minutes and you get impatient in that 30 seconds and you want it in 10 seconds. And that's just like that competition to shave down seconds. I think it's always going to be there. So no one ever gets satisfied with this... it always eventually becomes like, it feels like waiting.”
“there's going to be this like, okay, what really needs to be served fast, what really needs to be served on the most premium closed source models. And companies are probably going to get a lot more skilled at allocating from different forms of inference depending on the need.”
“he said, okay, inflation seems to be going in the wrong direction right now. He described policy as not restrictive. So then you fill in the blanks and it's like, okay, that means rate hikes. But he didn't quite say that.”
“the best outcome is the capabilities level off.”
“I'm on the Lockheed Martin looking at their Helios integrated weapon system and it's going through the specs and it's like deep magazine, low cost per kill, speed of light delivery. It's like it's gear. But, you know, these are the metrics.”
Joe Weisenthal
One-line summary: Co-host of Bloomberg's Odd Lots podcast (with Tracy Alloway). Tracked in the AI thread for thesis-grade framings on AI economics — the 'no free AI software' / levelized-cost-of-intelligence question and the treadmill-of-expectations on inference speed.
What they're known for
Long-time Bloomberg markets journalist; co-host (with tracy-alloway) of the Odd Lots podcast, which frequently treats AI infrastructure and economics as a markets-and-real-economy story.
Why they matter to artificial-intelligence
Tracked for thesis-grade framings on the economics of AI rather than its capabilities. On this source he sharpens the open-vs-closed question into two useful primitives: (1) there is no free AI software — even an open-weight model carries chip-depreciation and electricity cost, so "open source = free" doesn't transfer from traditional software; and (2) levelized cost of intelligence — the (not-yet-standardized) per-IQ-point unit that would let buyers compare open vs closed honestly. He also predicts a maturation toward inference-allocation skill: companies getting better at routing each task to the cheapest sufficient tier rather than defaulting to the premium closed model. See open-vs-closed-source-model-economics and inference-speed-as-a-pricing-premium.
Said
Speaker-attributed claims extracted from diarized sources. Each bullet mirrors one entry in quotes: frontmatter — keep them in sync.
-
On open-vs-closed-source-model-economics:
"in traditional software, open source. One nice thing about open source is you don't have to pay for it. So it's free. It's a little bit different when we're talking about there's no really such thing as like free AI software. Because even if it's like free, you still have to pay for the depreciation of the chips and you have to pay for the electricity to run them. So there is no real such thing as like free open source AI software." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On open-vs-closed-source-model-economics:
"are the open source models cheaper on a per unit of intelligence basis? If we had some way of saying levelized cost of intelligence, which I don't know if the industry has yet, are open source models cheaper per IQ point?" — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On inference-speed-as-a-pricing-premium:
"as you use it more, it's just like everything else, the treadmill of expectations. Here's some tasks that you can do in 30 seconds which maybe several years ago would have taken you 30 minutes and you get impatient in that 30 seconds and you want it in 10 seconds. And that's just like that competition to shave down seconds. I think it's always going to be there. So no one ever gets satisfied with this... it always eventually becomes like, it feels like waiting." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On open-vs-closed-source-model-economics, inference-speed-as-a-pricing-premium:
"there's going to be this like, okay, what really needs to be served fast, what really needs to be served on the most premium closed source models. And companies are probably going to get a lot more skilled at allocating from different forms of inference depending on the need." — 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s (2026-05-21)
-
On warsh-higher-for-longer-to-brokerage-nii-rerate:
"he said, okay, inflation seems to be going in the wrong direction right now. He described policy as not restrictive. So then you fill in the blanks and it's like, okay, that means rate hikes. But he didn't quite say that." — 2026-08-31-podcast-odd-lots-richmond-fed-s-tom-barkin-on-the-surprisingly (2026-08-31)
-
On solved-world-post-instrumentality:
"the best outcome is the capabilities level off." — 2026-08-20-odd-lots-nick-bostrom-on-what-happens-if-ai-solves-all-of (2026-08-20)
-
On hel-layered-iamd-to-lockheed-helios-mp-dopants, defense-industrial-base-magazine-depth:
"I'm on the Lockheed Martin looking at their Helios integrated weapon system and it's going through the specs and it's like deep magazine, low cost per kill, speed of light delivery. It's like it's gear. But, you know, these are the metrics." — 2026-09-04-podcast-odd-lots-why-laser-beams-are-the-hottest-new-tech-in (2026-09-04)
Sources
- 2026-05-21-odd-lots-why-cerebras-ceo-andrew-feldman-built-the-world-s
- 2026-08-31-podcast-odd-lots-richmond-fed-s-tom-barkin-on-the-surprisingly
- 2026-08-20-odd-lots-nick-bostrom-on-what-happens-if-ai-solves-all-of
- 2026-09-04-podcast-odd-lots-why-laser-beams-are-the-hottest-new-tech-in