beyond mental models retweeted
BREAKING: A deal between Iran and Oman declaring the Strait of Hormuz "transit corridor" is expected to be announced today.
Iran has said talks with Oman are strictly bilateral and the agreement is not necessarily a guarantee that the waterway has become safe for vessels.
beyond mental models retweeted
Dylan’s takes on OpenAI, Anthropic, AMD, xAI, Oracle, Meta, & Google:
OpenAI → “Super awesome.”
Anthropic → “I'm actually more optimistic on Anthropic than I am OpenAI…Their revenue is accelerating way faster because what they're focused on is more relevant to that $2 trillion software market versus OpenAI is split between software, AI for science, AI for consumer, etc…Anthropic is definitely executing on the software side better.”
AMD → “I love them, but they're pretty mid.”
xAI → “They're in real danger of not being able to raise capital. Of course everyone's going to give Elon capital, but the scale of capital required for him to keep up—he can get the next bet…xAI can get to the next stage of compute. They won't have more compute than OpenAI. They don't have more compute than any individual company—Google, Meta, etc. But they will have the biggest individual data center. They'll have a very focused team and what they do with that, they have to do something really big. Otherwise they will fall behind in the race. And Elon will not let that happen. He can subsidize and fund this round. But as rich as he is, he can't go to a 3-gigawatt data center unless he gets capital, which he can't do unless he gets revenue and fundraising.”
Oracle → “Oracle is going to make so much money if you believe OpenAI is successful.”
Meta → “I think Meta's got the cards to potentially own it all…the only company in the world that has the full stack from good hardware—that is what Meta just showed with their glasses with the screen—plus the good models, plus the capacity to serve them, plus the knowledge and know-how around recommendation systems to know what content to put in front of the user. It's all four of these that you need, plus the capital. I think Meta is so close to being the only company that can do that.”
Google → “I was pretty bearish Google like two years ago, but I'm super bullish Google now. They're waking up on every front. They're taking the TPUs, they're selling them externally, they're taking their models and they're actually competitive on them, and they're training much better and better. They're being aggressive on infrastructure investments.
There's still a lot of dysfunction throughout the company, but they do have the hardware business that they can pivot into this. They do have Android, YouTube, [and] all these IPs, they have search that can come together when we turn to that next interface of consumer. But also they can dominate the professional sense too, potentially. And I think Google's well positioned to go capture both markets, or a meaningful share of both.”
My conversation with @dylan522p on everything you ever wanted to know about the incredible buildout happening in AI infrastructure
This is a whirlwind tour through the companies, deals, models, chips, power sources, partnerships, and people powering AI
Dylan is some sort of genius. I wish a version of him existed for every subject. I hope you enjoy this as much as I did.
Timestamps:
0:00 Introduction
0:39 The Infinite Money Glitch
4:42 Oracle's Bet on OpenAI
6:06 Nvidia's Strategic Investment and Deal Mechanics
9:17 Why Bigger Models Aren't Always Better
15:07 The Adoption Curve Problem
19:01 The Tokenomics of AI
23:59 Over-Parameterization and Model Learning
28:10 Building Environments for AI Training
32:16 AI in Everyday Life
34:46 The Future of Reasoning and Compute Scaling
44:34 The Spectrum of AI Optimism
46:29 Timeline to AGI
49:25 Talent Wars
58:37 Power Dynamics in the AI Ecosystem
1:18:42 AI for Material Science and Hard Tech
1:22:21 Building Infrastructure
1:29:55 US vs China: Who Really Needs AI to Win?
1:37:42 Favorite AI Bears
1:48:08 Speed Round: Company Impressions
1:55:11 The Death of Traditional SaaS Business Models
2:00:24 The Kindest Thing
beyond mental models retweeted
Mortgages are 7%. Your dollar is losing value every day. Higher Education is being threatened and talented immigration disincentivized or banned. The President is manipulating markets based on bad policy and ideology.
This is not a good road we are walking down
beyond mental models retweeted
$GOOG ATH as investors realize Wikipedia isn't disrupting Google.
beyond mental models retweeted
The critical CPI Rent category is proving to be like an 18-wheeler riding its brakes down the mountain. It’s going down, there’s no doubt about that. But the lagged-by-design methodology means that it’s just not going to decelerate as fast as real-life rents are in every major private-sector dataset.
And given the outsized importance of rents in the CPI, that might complicate the path toward the Fed’s stated goal of 2% inflation. To that point: When you remove the lagged shelter category, inflation has already been coming in around 2% or less for seven straight months. As of November, it was 1.4%. But with shelter, inflation measured 3.1% -- which is basically flat since June.
So in other words: The rest of the party has made it down the mountain, and they’re waiting on shelter (rents) to catch up.
But here’s where it gets interesting: By nearly every other account, shelter should be cooling more than the CPI shows. The Penn State/ACY Alternative Inflation Index, for example, shows inflation at -0.1% as of October by swapping out the CPI’s lagged rent data with their own real-time estimates. As its authors point out: “Housing rents exhibit large relative importance ranging from 16% to 42% depending on inflation measures,” so the swap-out makes a huge difference.
Rents are the largest variable in the CPI’s largest category (shelter), as the same rent survey is used to estimate costs for both renters and homeowners (with minor differences between “rent of primary residence” and “owners’ equivalent rent.”)
What is worrisome for the 2% inflation backers is that the CPI Rent slowdown looks far slower than it should be. Month-over-month change in CPI rents have been hovering around 0.5% since March, without further decelerating. While that’s well below 2021 and early 2022 peak hikes (hence the gradual drop-off in YoY rent inflation), it’s curiously still about double the norms for the 2010s decade. And yet, in most private sector datasets, rents have flattened out well below pre-COVID norms – partly due to a 50-year high in new apartment supply.
The CPI shelter and rent methodologies are quite complicated, but one possible factor could be this: Unlike private-sector measures, CPI Rent data is not collected on the same properties each month. Each sampled unit is surveyed only once per SIX months on the theory that rents don’t change that frequently. That’s true, BUT it also means the CPI won’t capture rents as they change. If you were surveyed in October and your rental costs changed in November, the CPI won’t pick that up until April.
It's also important to acknowledge that the CPI is attempting to measure the “contract rent,” or what the industry calls the in-place rent roll, and that will lag the asking rent for new lease data that private-sector data providers show and that make headlines. Contract rent includes all active leases (regardless when they’re signed) and renewals. That makes CPI Shelter the one major category that isn’t a real-time price, but an intentionally lagged one. But still: Renewals always follow new lease rents. Renewal rents will almost never get priced above new leases because that just encourages move-outs.
In fairness, the same methodology quirks created a lag going up the inflation hill, too. But it’s perhaps even more challenging going downhill – if only because, at this point, the destination is a pre-determined conclusion. We know it’ll get down the hill. We just don’t know how long it’ll take to get there.
beyond mental models retweeted
Best line from today's newspapers:
From ET:
"A Bengaluru based fintech saw its talks to raise funds fall apart after founders refused to share financials with potential investors".
beyond mental models retweeted
Asking all the FinTech influencers out there, For a young professional starting to write about FinTech. What will be your advice to get started?
@sytaylor @aakashgupta @AlexH_Johnson
@SamBobLev @PopularFintech @rshevlin
Please tag folks who can provide valuable input! Thanks!
beyond mental models retweeted
@SultanAhmedCopy would love your views as well
beyond mental models retweeted
1) I agree and I think most PFMs struggle:
👉To offer features that maintain daily user retention
👉To monetize services aimed at "expense management" and "saving"
👉To stand out when PFM is already available on most bank's platforms
Heard of a few people building a new version of @mint now that Intuit shut it down. I wouldn't recommend it if you want to build a venture-scale business.
There aren't that many people who want to actively manage their finances; startup graveyard is littered with PFM's.
beyond mental models retweeted
To be fair, Byju did turn out to be a proper education company- it has educated the world on so many issues including PE investing, private investing, due diligence etc.
beyond mental models retweeted
Steve Jobs was the single largest individual shareholder in Disney and it all started with this $5 million check.
The year was 1986 and Steve Jobs was in exile from Apple.
The company that would one day become Pixar was called Graphics Group. Graphics Group was owned by Star Wars creator George Lucas.
George Lucas was getting divorced and needed to sell off some assets.
Steve Jobs agreed to give George Lucas $5 million and invest another $5 million in the company.
This was not the first time Steve Jobs had expressed interest in the technology behind Graphics Group.
When he was still at Apple, Jobs tried to get Apple’s board to buy the company.
They turned him down.
This is Steve explaining why he was interested in acquiring Graphics Group:
“These guys were way ahead of us on graphics, way ahead. They were way, way ahead of anybody. I just knew in my bones that this was going to be very important."
Steve would eventually invest $54 million of his own money in Pixar.
After several years of failing to sell The Pixar Image Computer, Pixar decided to go all in on making films.
After years of poor sales and rising expenses they were short on cash. To solve this problem Steve signed a deal to produce 3 animated films for Disney.
The first of these movies was Toy Story —the world’s first computer animated feature film.
Now here is where this story gets really interesting:
Steve decided to capitalize on the publicity of Toy Story’s release and take Pixar public at the same time.
Toy Story was released on November 22, 1995.
Pixar went public on November 29, 1995.
Steve's two partners — Ed Catmull and John Lasseter — did not want to go public.
They thought it was too early. They wanted to do a few more films first.
Then Steve laid out his strategy and told them what was likely coming from Disney.
Steve said let’s assume Toy Story is a big success (it was —grossing over $370 million).
When that happens the CEO of Disney — Michael Eisner — will realize he inadvertently created a rival for Disney Animation.
Pixar only had to make two more films for Disney and then they’d be free to do as they like.
Steve said Eisner’s next move would be to try and renegotiate with Pixar and keep Pixar close to Disney.
Steve said if they went public and had more money they’d be able to negotiate better terms with Disney.
Steve wanted a 50/50 split with Disney on returns. In order to request this Pixar had to be able to put up 50% of the production costs.
Going public would allow Pixar to afford this.
The time is now Steve said: "This is our moment."
And he was right. Toy Story was a massive success and Pixar's IPO raised $140 million -- the biggest IPO of 1995.
A few months later Eisner approached Jobs and said he wanted to keep Pixar as a partner.
Eisner accepted Steve’s 50/50 deal.
This partnership wasn't going to last. Steve Jobs and Michael Eisner couldn’t get along.
A few years later negotiations broke down completely and Pixar started looking for other partners.
Then Michael Eisner got fired. The new CEO of Disney was Bob Iger.
When Bob Iger found out he was going to be the next CEO he called his daughters first.
The next person he called was Steve Jobs.
He told Steve he valued Pixar and wanted to make a deal. This is Bob explaining why he felt Disney needed to buy Pixar:
“Michael [Eisner] didn’t understand that Disney’s problems in animation were as acute as they were. That manifested itself in the way he dealt with Pixar. He never felt he needed Pixar as much as he really did. Every negotiation needs to be resolved by compromises. Neither one of them [Eisner and Jobs] is a master of compromise.”
How did Bob Iger realize how badly Disney needed Pixar?
By watching a parade at Hong Kong Disneyland
Iger realized that the only characters in the parade that had been created in the past decade were Pixar’s:
“A lightbulb went off. I’m standing next to Michael [Eisner], but I kept it completely to myself, because it was such an indictment of his stewardship of animation during that period. After ten years of The Lion King, Beauty and the Beast, and Aladdin, there were then ten years of nothing.”
Bob begins to negotiate with Steve to buy Pixar and does something very unusual:
He tells Steve how badly Disney needs Pixar.
Even more surprising this tactic worked. This is how Steve reacted:
“That’s why I just loved Bob Iger. He just blurted it out. Now that’s the dumbest thing you can do as you enter a negotiation, at least according to the traditional rule book. He just put his cards out on the table and said, ‘We’re screwed.’ I immediately liked the guy, because that’s how I worked too. Let’s just immediately put all the cards on the table and see where they fall.”
Disney purchased Pixar for $7.4 billion in Disney stock.
And Steve Jobs went from the single largest shareholder in Pixar to the single largest individual shareholder in Disney.
beyond mental models retweeted
Everyone studying AI to generate alpha yet the greatest investing insight of the last decade is that Apple should trade at the same multiple as Clorox.
beyond mental models retweeted
Replying to @MetacriticCap
beyond mental models retweeted
Big plunge in 10yr T yield after $SIVB down 60% & 20%+ more in after hours. Concerns over VC bank’s solvency send shockwaves thru markets. But this is bullish for housing. 50 bps hike off table. Possible Fed pause
Last time 10yr hit 3.8% was Feb 16 & the 30 yr mortgage was 6.7%
beyond mental models retweeted
beyond mental models retweeted
Trucking and housing stocks trading near all time highs is amazing. Seems to be screaming soft landing
But don't get tricked by the narrative that speculation is back. Unprofitable tech stocks fell from the penthouse to the sub basement and have only bounced back to the basement
beyond mental models retweeted
$KKR out with a new deck today, showing they think the current earnings power of the biz = $6 per share. Should grow a lot over time...
beyond mental models retweeted
Scott Malpass (former CIO of Notre Dame, teaches a highly reviewed investing class)
podcasts.apple.com/us/podcas…
Sarah Samuels on analyzing hedge funds (does so for $1.5tn advisor)
podcasts.apple.com/us/podcas…
Larry Kochard on the LP CIO role
podcasts.apple.com/ca/podcas…
beyond mental models retweeted
Ram Parameswaran on internet investing
podcasts.apple.com/ca/podcas…




















