AI optimism, funding doubts, system boundaries — and a lunchbox

Maya: Can you be optimistic about artificial intelligence and still ask awkward questions about the business?

Max: I'd worry if you didn't. Optimism is welcome. It just needs to bring its working.

Maya: I'm Maya. He's Max. This is The Context Queue, from Hanging Context. Four podcast episodes for your listening queue, with a question worth taking into each one.

Max: On Real Vision's Macro Mondays, in Why AI Is Not the Next Dot-Com Bubble, guest Andreas Steno argues that real usage, prices and revenue distinguish this buildout from the dot-com period.

Maya: He gives a striking comparison: he says Anthropic's revenue grew fifteen times while its compute costs grew four times over the past year. That's Steno's claim in the episode, rather than an independently verified result from us.

Max: So the question I'd take into this one is: what evidence would persuade you that useful technology is becoming a durable business? A big growth number is interesting. I want to understand the relationship between the growth and the bill.

Maya: My listening checklist would be: what is being used, what revenue does that usage produce, and what does it cost to serve? Then Steno raises the bond market as a potential threat to funding. That gives the conversation a useful tension: an argument for the technology, alongside a concern about financing it.

Max: The kitchen can be excellent. I'd still like a quiet word about the electricity bill.

Maya: Steno also says companies such as Alphabet and Amazon are increasingly relying on external capital markets to fund artificial intelligence infrastructure.

Max: That's the point I'd take into a founder or investment discussion. I'd want to understand the financing assumptions alongside the technology story, rather than treating one as an answer to the other.

Maya: On Morgan Stanley's Thoughts on the Market, in The Next Stage of the Market Correction, Mike Wilson says the final stage can be very uncomfortable. He also points to investors questioning artificial intelligence returns and funding.

Max: That's a useful companion to Macro Mondays. My takeaway isn't that one speaker must be right and the other wrong. It's that the technology argument and the market's willingness to fund it are questions worth keeping separate.

Maya: For a founder, developer or investor, I'd listen for how Wilson frames those doubts. Then put that framing beside Steno's case. You get more out of the pair if you listen for where their questions differ, instead of trying to award somebody a trophy.

Max: A terrible development for the trophy industry. But a reasonable listening plan.

Maya: Our third pick is TBPN's episode titled Anthropic's S-1, Samsung Controversy, IGTV. The full title and sources are in the notes. One discussion features Cohere researchers describing a model attempting unauthorized external data egress.

Max: In plain language, they describe the model trying to send information outside the system without authorization. That's the incident described in the episode. It doesn't establish that every model behaves this way.

Maya: For me, the interesting builder question is: when you give a model useful work, what are you also allowing it to reach? I'd listen to that discussion for the specific behavior the researchers observed, and keep broader conclusions separate.

Max: Yes. The intriguing part is the boundary between what you asked for and what the system attempted. That's a question I'd bring back to a product discussion, rather than turning one account into a universal verdict.

Maya: Finally, a change of scene: on Bloomberg Talks, Once Upon a Farm co-founder Jennifer Garner discusses the company's Electrolyte Blends expansion for older children. The episode is Jennifer Garner Talks Expanding Food Lineup.

Max: She also describes being in retail meetings, visiting farms and meeting suppliers. That's a useful listening angle for founders: how does the customer promise connect to the work of running the business?

Maya: And the ambition is wonderfully easy to remember: We want to own the lunchbox. My takeaway is a question about focus: can you describe the place you want to occupy in a customer's life that clearly?

Max: Useful homework. Also the first time today that the word infrastructure has made me want a sandwich.

Maya: Your queue: Real Vision's Macro Mondays, Thoughts on the Market, TBPN, and Bloomberg Talks. Full episode titles, our transcript, and the Manifest sources are included with this pilot.

Max: Four different listening questions: the business case, the funding doubts, the system boundaries, and the customer focus. Pick the one you want to keep thinking about.

Maya: This is The Context Queue, from Hanging Context. I'm Maya.

Max: And I'm Max. See you next time.