The BasePodcastHard Fork53 min12,812 views

What Anthropic's S-1 Tells Us About the A.I. Business | Hard Fork

In a 53-minute episode, Hard Fork discusses Anthropic's IPO filing with New York Times reporter Erin Griffith: its losses, slowing growth, safety risk factors and governance. Later the panel covers the midterms, a PHP vulnerability, Microsoft buyouts and an Anthropic study of experts deferring to AI.

Open the source ↗
About Anthropic or ClaudeClaude, made by Anthropic, wrote this entry. It follows the same rules as every other entry, whoever the entry is about.

Why it matters

The same thing, for different readers.

What it means for you

Anthropic may ask the stock market to value it near $2 trillion, so savers and pension funds could soon hold a stake in a lab that warns of existential risk.

For lawmakers

Hard Fork's guests say Anthropic's founders keep 50.1% of the votes and a trust picks most of the board, so outside shareholders will have little say over its safety choices.

For researchers

An Anthropic randomized trial, as Jasmine Sun describes it, found expert analysts mostly adopted an AI's advice and rarely said they were influenced by it.

The question it raises

If experts follow an AI's reasoning without noticing, the guests ask whether human control is lost through quiet deference rather than a rogue system.

In their own words

Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.

From our own transcript of the episode’s audio, made by machine. Speaker names come from what is said, not from a voice match.

“An $80 billion net loss seems insane, but it's actually kind of a weird accounting quirk where they have to count some convertible debt that they issue to some of their shareholders as a net loss and that will essentially be cleared in the next quarter or then or by the time they go public.”
Erin Griffith, tech reporter, The New York TimesGriffith, answering the host's question on the S-1's main takeaways, explains the headline loss in Anthropic's filing.
Found word for word in the sourceHear it at 4:33 ↗Speaker known from context, not confirmed by listening
“the S1 that we saw did not have Q3 numbers in it and they've already kind of previewed those numbers to their investors and our reporting and others have shown that the growth is actually slowing in in the third quarter”
Erin Griffith, tech reporter, The New York TimesGriffith on the biggest risk investors are weighing: whether demand keeps pace with Anthropic's compute commitments.
Found word for word in the sourceHear it at 6:49 ↗Speaker known from context, not confirmed by listeningstarts inside a longer claim, right after “so”spoken clean-up: in (said twice)
“concerns have been raised both within our organization and by other researchers uh that sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, oversight and security measures, pose catastrophic or existential risks to humanity.”
Hard ForkA host reads one of Anthropic's S-1 risk factors aloud while discussing the filing's 90-plus-page risk section.
Found word for word in the sourceHear it at 10:27 ↗Speaker known from context, not confirmed by listeningspoken clean-up: uh
“not only did the experts uh I think over 80% of the time make the same decision as the AI, which are split half, they also over 70% of the time use the AI's arguments to justify their decision, but only 15% of the time did they say”
Jasmine Sun, writerSun describes, from memory, an Anthropic randomized trial of expert analysts advised by AI, when asked which story Sun is tracking.
Found word for word in the sourceHear it at 48:49 ↗Speaker known from context, not confirmed by listeningstarts inside a longer claim, right after “that”spoken clean-up: uh
“People feel very negatively about data centers. People feel very negatively about AI. This is bipartisan.”
Jasmine Sun, writerSun on public opinion before next week's midterms, saying AI is not yet a top voting issue.
Found word for word in the sourceHear it at 24:06 ↗Speaker known from context, not confirmed by listeningstarts inside a longer claim, right after “that”

The record

Questions
How fast is AI advancing, and where is it heading? · What happens to work, income and the economy? · Can AI systems act beyond what people intended or allowed?
People
Erin Griffith (Speaks, tech reporter, The New York Times) · Mike Isaac (Speaks, tech reporter) · Jasmine Sun (Speaks, writer) · Dario Amodei (Mentioned)
Also covered by
The Guardian
What we read
We read the whole source.
Found by
The Radar, run of October 31, 2026, via yt-hard-fork-youtube (tier 1: the source’s own channel or an official record)
Cite this
Hard Fork, “What Anthropic's S-1 Tells Us About the A.I. Business | Hard Fork”, October 30, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/hard-fork-2026-10-30-anthropic-s1-ai-business
Our notesTranscript has no speaker labels; speakers inferred from introductions, so speakerCertain false. The transcript names the host as 'Max Reed' and 'Aaron Griffith' (triage says Erin Griffith), likely transcription errors; host left out of people. Griffith's third-quarter 44% growth figure and Sun's trial figures are spoken from memory and not checked against the S-1 or the study. Another passage in the audio mentions a startup accessing Anthropic's Slack, not verified.