The Joe Walker Podcast (@joewalkerpod) header
The Joe Walker Podcast (@joewalkerpod) profile picture

Well-researched interviews on ideas, technology, and policy. (Sometimes Australian.) Hosted by @JosephNWalker.

Sydney
Joined January 2017
Photo by The Joe Walker Podcast (@joewalkerpod)Photo by The Joe Walker Podcast (@joewalkerpod)Photo by The Joe Walker Podcast (@joewalkerpod)Photo by The Joe Walker Podcast (@joewalkerpod)Photo by The Joe Walker Podcast (@joewalkerpod)Photo by The Joe Walker Podcast (@joewalkerpod)Photo by The Joe Walker Podcast (@joewalkerpod)Photo by The Joe Walker Podcast (@joewalkerpod)Photo by The Joe Walker Podcast (@joewalkerpod)Photo by The Joe Walker Podcast (@joewalkerpod)
Follow my primary account here: @JosephNWalker
1
5
1,950
The Joe Walker Podcast retweeted
This is brilliant
New episode with John Hempton (@John_Hempton), legendary Australian hedge fund manager and short seller. I wanted to get John’s thoughts on the likelihood of an overbuild / bust in AI compute, and what implications that has for the Australian government’s data centre push. John talks me through why he thinks the global compute build-out is “fibre optics Mark II” - i.e. why AI is “the most amazing technology of your lifetime”, yet compute looks like “the worst business in the world”. John’s recipe for the worst business in the world is: “high capital cost, low marginal cost, competition, and technical improvement”. With these ingredients, new entrants push prices down to marginal cost, businesses can't recover their capital costs, and they go bust. To get an adequate return on the $7 trillion of capex expected to be poured into compute by 2030, John estimates you’d need $2 trillion per year in revenue. That would most likely have to come from businesses replacing workers (or not hiring new ones), as consumers won’t get you there. But companies typically don’t fire people for a 10% saving. So to pay $2 trillion a year, they’d need to save at least twice as much, or $4 trillion in wages. The OECD wage bill is about $40 trillion per year. So you’d need to save, for argument's sake, 10% of the wage bill across the entire OECD. But that kind of saving seems politically and practically implausible in such a short timeframe. So expect excess supply and bankruptcies. (Again, none of that is to say the underlying tech is useless - far from it.) John also shares the thing that shocked him most about Anthropic's S-1: namely, how compute-efficient it is. By his estimates, Anthropic has spent only about $12 billion on training compute over the lifetime of the firm. Which makes the $7 trillion build-out look even worse. There's so much more of interest in this conversation, but let me share one last thing. There's an interesting question as to exactly what the productive infrastructure left over from any overbuild will be. It might not be the GPUs, but the fabs and/or energy generation. John tells a great story about how old chip fabs he thought were worthless ended up finding an application making the tiny sensors in car air conditioners and Tesla batteries. After all this, the policy implication for Aus govt is to happily facilitate the compute build-out, but just don’t subsidise the world’s compute the way China subsidised the world’s solar panels (another example of a bad business). Watch below, or wherever you get your podcasts! 0:00:00 – The most amazing technology of your lifetime 00:03:10 – A short seller discovers jaggedness (from first principles) 00:19:40 – The Situational Awareness blow-up 00:24:46 – The worst business in the world 00:31:02 – Fibre optics, Mark II 00:42:41 – The best model was shockingly cheap to train 01:01:26 – Why the $7 trillion build-out can't earn its return 01:11:13 – What survives a compute bust? 01:22:59 – A data centre is a better neighbour than a prison 01:27:30 – A bad business is not a bubble
4
20
1,575
The Joe Walker Podcast retweeted
Replying to @John_Hempton
@John_Hempton , really hope you see this message. I just wanted to say this is one of the best podcasts I have ever listened to. To listen to someone who is obviously very intelligent, who has accumulated a lot of scars and experience from a long career in markets and who can verbalise that knowledge in a simple, eloquent way is a real treat. You take very complex and well thought out ideas and get them across in a manner that is easy to understand and a pleasure to listen to. Thank you to you and to @JosephNWalker Great job Joe
1
2
43
2,670
The Joe Walker Podcast retweeted
New episode with John Hempton (@John_Hempton), legendary Australian hedge fund manager and short seller. I wanted to get John’s thoughts on the likelihood of an overbuild / bust in AI compute, and what implications that has for the Australian government’s data centre push. John talks me through why he thinks the global compute build-out is “fibre optics Mark II” - i.e. why AI is “the most amazing technology of your lifetime”, yet compute looks like “the worst business in the world”. John’s recipe for the worst business in the world is: “high capital cost, low marginal cost, competition, and technical improvement”. With these ingredients, new entrants push prices down to marginal cost, businesses can't recover their capital costs, and they go bust. To get an adequate return on the $7 trillion of capex expected to be poured into compute by 2030, John estimates you’d need $2 trillion per year in revenue. That would most likely have to come from businesses replacing workers (or not hiring new ones), as consumers won’t get you there. But companies typically don’t fire people for a 10% saving. So to pay $2 trillion a year, they’d need to save at least twice as much, or $4 trillion in wages. The OECD wage bill is about $40 trillion per year. So you’d need to save, for argument's sake, 10% of the wage bill across the entire OECD. But that kind of saving seems politically and practically implausible in such a short timeframe. So expect excess supply and bankruptcies. (Again, none of that is to say the underlying tech is useless - far from it.) John also shares the thing that shocked him most about Anthropic's S-1: namely, how compute-efficient it is. By his estimates, Anthropic has spent only about $12 billion on training compute over the lifetime of the firm. Which makes the $7 trillion build-out look even worse. There's so much more of interest in this conversation, but let me share one last thing. There's an interesting question as to exactly what the productive infrastructure left over from any overbuild will be. It might not be the GPUs, but the fabs and/or energy generation. John tells a great story about how old chip fabs he thought were worthless ended up finding an application making the tiny sensors in car air conditioners and Tesla batteries. After all this, the policy implication for Aus govt is to happily facilitate the compute build-out, but just don’t subsidise the world’s compute the way China subsidised the world’s solar panels (another example of a bad business). Watch below, or wherever you get your podcasts! 0:00:00 – The most amazing technology of your lifetime 00:03:10 – A short seller discovers jaggedness (from first principles) 00:19:40 – The Situational Awareness blow-up 00:24:46 – The worst business in the world 00:31:02 – Fibre optics, Mark II 00:42:41 – The best model was shockingly cheap to train 01:01:26 – Why the $7 trillion build-out can't earn its return 01:11:13 – What survives a compute bust? 01:22:59 – A data centre is a better neighbour than a prison 01:27:30 – A bad business is not a bubble
29
50
11
280
39,974
The Joe Walker Podcast retweeted
What would it look like if the data centre build-out actually lowered energy prices? Tim Nelson steps through the economic logic:
If Australia wanted to build out 10-15 GW of AI compute in the next three years, what would it actually take under the government's constraints of no net costs for consumers and 100% renewable backing? I had great fun trying to work this out with Tim Nelson, one of Australia's sharpest energy economists. The main idea we tested: what if the best place for training data centres is the outback -- far from cities, close to strong sun and wind, and maybe even off the main grid? Watch/listen below or on YouTube, Spotify, Apple Podcasts, etc. Enjoy! Timestamps: (0:00:00) – Introduction. (0:00:42) – Could AI data centres consume half the grid? (0:06:13) – Can we power the build-out with only renewable energy? (0:15:44) – What could bottleneck the build-out? (0:24:36) – How data centres could lower power prices (0:31:58) – What smelters teach us about building renewables quickly (0:35:52) – How would Tim build 10-15 GW of AI compute in 3 years? (0:40:38) – Special compute zones in the outback (1:02:02) – Renewable energy certificates and the time-matching debate (1:12:01) – Nuclear; unintended consequences; Jevons (1:20:19) – Post-interview chat
4
13
1,856
RT @JosephNWalker: AI factories in the outback: If Australia wanted to build out 10-15 GW of AI compute in the next three years, what woul…
1
9
The Joe Walker Podcast retweeted
In Melbourne with some time to kill tomorrow evening (>6:30pm, 2nd Sep) or Thursday arvo (2:30-3:30pm, 3rd Sep). Open to meeting anyone to talk about AI / compute / energy. Who should I meet? (Feel free to DM.)
3
4
30
2,927
The Joe Walker Podcast retweeted
In Canberra with some time to kill tomorrow evening (>6:30pm, 1 Sep). Open to meeting anyone to talk about AI / compute / energy. Who should I meet? (Feel free to DM.)
6
2
40
4,657
The Joe Walker Podcast retweeted
Australia has "a month or months, rather than a year" to sort out copyright if it wants to become the second home for major labs' AI training:
The geostrategic case for compute in Australia: Janet Egan (@janet_e_egan) has been doing a bunch of quiet work trying to nudge Australian policymakers to think more seriously about AI. She works at the Center for a New American Security in Washington, but was back home in Australia over the past month. Was lucky to catch her on the day before she returned to the US. We pressure-test various strategic arguments for a large-scale compute industry and talk about the steps we would need to take to enable it. Watch below, or wherever you get your podcasts! Timestamps: (0:00:00) – Introduction (0:00:33) – Lessons from the AdBlue crisis (0:09:02) – How seriously do Australian policymakers take AI? (0:12:48) – How much does frontier access matter? (0:28:55) – Compute-for-access (0:43:08) – What does sufficient leverage look like? (0:53:00) – What does AI sovereignty actually mean? (1:03:15) – Taiwan, undersea cables and data centres as military targets (1:10:30) – Could compute buy Australia influence over AI safety? (1:15:19) – Verifying international AI agreements (1:24:00) – Copyright: the red line? (1:37:49) – The maximally ambitious path
5
7
4
61
17,994
The Joe Walker Podcast retweeted
Janet Egan has spent the last month back home in Australia talking to policymakers about AI. I asked her: Any differences between how Aus and US policymakers are thinking about AI? Janet: "The first thing that is so distinct between Washington, D.C. and Australia is how seriously people take AI." "... There are a lot of very senior people in positions in [the Australian] government who still aren’t grappling with the realities of AI progress."
The geostrategic case for compute in Australia: Janet Egan (@janet_e_egan) has been doing a bunch of quiet work trying to nudge Australian policymakers to think more seriously about AI. She works at the Center for a New American Security in Washington, but was back home in Australia over the past month. Was lucky to catch her on the day before she returned to the US. We pressure-test various strategic arguments for a large-scale compute industry and talk about the steps we would need to take to enable it. Watch below, or wherever you get your podcasts! Timestamps: (0:00:00) – Introduction (0:00:33) – Lessons from the AdBlue crisis (0:09:02) – How seriously do Australian policymakers take AI? (0:12:48) – How much does frontier access matter? (0:28:55) – Compute-for-access (0:43:08) – What does sufficient leverage look like? (0:53:00) – What does AI sovereignty actually mean? (1:03:15) – Taiwan, undersea cables and data centres as military targets (1:10:30) – Could compute buy Australia influence over AI safety? (1:15:19) – Verifying international AI agreements (1:24:00) – Copyright: the red line? (1:37:49) – The maximally ambitious path
9
15
2
92
12,262
The Joe Walker Podcast retweeted
The geostrategic case for compute in Australia: Janet Egan (@janet_e_egan) has been doing a bunch of quiet work trying to nudge Australian policymakers to think more seriously about AI. She works at the Center for a New American Security in Washington, but was back home in Australia over the past month. Was lucky to catch her on the day before she returned to the US. We pressure-test various strategic arguments for a large-scale compute industry and talk about the steps we would need to take to enable it. Watch below, or wherever you get your podcasts! Timestamps: (0:00:00) – Introduction (0:00:33) – Lessons from the AdBlue crisis (0:09:02) – How seriously do Australian policymakers take AI? (0:12:48) – How much does frontier access matter? (0:28:55) – Compute-for-access (0:43:08) – What does sufficient leverage look like? (0:53:00) – What does AI sovereignty actually mean? (1:03:15) – Taiwan, undersea cables and data centres as military targets (1:10:30) – Could compute buy Australia influence over AI safety? (1:15:19) – Verifying international AI agreements (1:24:00) – Copyright: the red line? (1:37:49) – The maximally ambitious path
8
8
4
86
34,435
The Joe Walker Podcast retweeted
Great interview between @JosephNWalker and @Charlton_AB on Australia & AI compute: "10 years from today... tokens flowing around our economy will be as least as important, potentially much more important, than oil... the question for us is how do we think about that dependency?"
Andrew Charlton (@Charlton_AB) is the key federal politician shaping Australia's AI strategy. In episode #1 of my 'Compute in Australia' series, he shares his thinking on how Australia can turn compute into sovereign AI -- "the oldest industrial policy trick in the book" -- and much more. Listen below, or wherever you get your podcasts. (0:01:17) How a minister uses AI, and thinks about its trajectory (0:10:28) Is diffusion a sufficient strategy for middle powers? (0:22:33) Comparative advantage versus sovereignty (0:26:41) Where in Australia's national accounts will compute show up? (0:32:06) Australia as a global inference compute hub (0:39:28) How big will compute rents be? (0:47:01) Time to power (0:55:07) Taxing compute? (0:58:03) Data centres and the social licence (1:10:58) The geostrategic case for compute (1:20:50) Compute for access (1:26:15) Parlaying compute into sovereign AI (1:36:03) What does sovereign AI actually mean? (1:58:52) Orbital compute
2
5
17
3,247
The Joe Walker Podcast retweeted
Thread on @Charlton_AB podcast with @JosephNWalker: * Charlton mostly talks like an economist, not like a minister reciting talking points. Joe does an excellent job. He’s not going for grabby headlines but exploring Charlton’s core thinking. It’s not often you get a ... 1/
1
2
8
2,120
The Joe Walker Podcast retweeted
Typically great chat! My main takeaway is that the government seems committed to bartering with data centre builders (for more power, cheaper compute, better access etc) rather then just taxing high end chips and using said money to buy stuff. Taxing seems much better to me.
Andrew Charlton (@Charlton_AB) is the key federal politician shaping Australia's AI strategy. In episode #1 of my 'Compute in Australia' series, he shares his thinking on how Australia can turn compute into sovereign AI -- "the oldest industrial policy trick in the book" -- and much more. Listen below, or wherever you get your podcasts. (0:01:17) How a minister uses AI, and thinks about its trajectory (0:10:28) Is diffusion a sufficient strategy for middle powers? (0:22:33) Comparative advantage versus sovereignty (0:26:41) Where in Australia's national accounts will compute show up? (0:32:06) Australia as a global inference compute hub (0:39:28) How big will compute rents be? (0:47:01) Time to power (0:55:07) Taxing compute? (0:58:03) Data centres and the social licence (1:10:58) The geostrategic case for compute (1:20:50) Compute for access (1:26:15) Parlaying compute into sovereign AI (1:36:03) What does sovereign AI actually mean? (1:58:52) Orbital compute
4
4
29
4,279
The Joe Walker Podcast retweeted
An impressive interview that I recommend all Australians listen too. What is probably most interesting about this is the fact that a prominent politician sat down for an extended period of time on a single topic. Having not thought much of Andrew, I now atleast have a clearer understanding of his thoughts on a topic I care about. This alone makes me respect him more. If I had to nitpick, I’d say he was still a touch too political and wasn’t prepared (or able) to go deep enough into the topic. My guess is that’s for political reasons, rather than his ability to. Would love to see more Australian politicians doing long form, it’s a great precedent to set.
I’ve long been a fan of Joe Walker’s podcast, so it was great to sit down with him and have a wide-ranging conversation about compute, AI and what Australia needs to do to be a maker—not just a taker—in the AI economy. 📺Watch: youtu.be/0MGUR4giXJ0?si=1MnB…
8
3
46
7,985
The Joe Walker Podcast retweeted
Andrew Charlton MP (@Charlton_AB) on how the Australian government wants us to leverage a domestic compute industry into sovereign AI:
Andrew Charlton (@Charlton_AB) is the key federal politician shaping Australia's AI strategy. In episode #1 of my 'Compute in Australia' series, he shares his thinking on how Australia can turn compute into sovereign AI -- "the oldest industrial policy trick in the book" -- and much more. Listen below, or wherever you get your podcasts. (0:01:17) How a minister uses AI, and thinks about its trajectory (0:10:28) Is diffusion a sufficient strategy for middle powers? (0:22:33) Comparative advantage versus sovereignty (0:26:41) Where in Australia's national accounts will compute show up? (0:32:06) Australia as a global inference compute hub (0:39:28) How big will compute rents be? (0:47:01) Time to power (0:55:07) Taxing compute? (0:58:03) Data centres and the social licence (1:10:58) The geostrategic case for compute (1:20:50) Compute for access (1:26:15) Parlaying compute into sovereign AI (1:36:03) What does sovereign AI actually mean? (1:58:52) Orbital compute
7
5
2
28
5,181
The Joe Walker Podcast retweeted
I’ve long been a fan of Joe Walker’s podcast, so it was great to sit down with him and have a wide-ranging conversation about compute, AI and what Australia needs to do to be a maker—not just a taker—in the AI economy. 📺Watch: youtu.be/0MGUR4giXJ0?si=1MnB…
31
12
4
64
14,775
The Joe Walker Podcast retweeted
This is a fantastic episode. What a great kickoff to @joewalkerpod 's series on Compute in Australia.
Andrew Charlton (@Charlton_AB) is the key federal politician shaping Australia's AI strategy. In episode #1 of my 'Compute in Australia' series, he shares his thinking on how Australia can turn compute into sovereign AI -- "the oldest industrial policy trick in the book" -- and much more. Listen below, or wherever you get your podcasts. (0:01:17) How a minister uses AI, and thinks about its trajectory (0:10:28) Is diffusion a sufficient strategy for middle powers? (0:22:33) Comparative advantage versus sovereignty (0:26:41) Where in Australia's national accounts will compute show up? (0:32:06) Australia as a global inference compute hub (0:39:28) How big will compute rents be? (0:47:01) Time to power (0:55:07) Taxing compute? (0:58:03) Data centres and the social licence (1:10:58) The geostrategic case for compute (1:20:50) Compute for access (1:26:15) Parlaying compute into sovereign AI (1:36:03) What does sovereign AI actually mean? (1:58:52) Orbital compute
5
25
4,440