Bank of Braavos

@hutianyi991 If content is infinite, whats the price/cost of content?

@BankBraavos and nothing change from the payout pool. The irony is that AI makes content infinite, but attention remains finite. YouTube taxes the attention; Netflix has to buy the content. In an AI explosion, YouTube’s margins expand while Netflix’s 'premium' tax gets heavier.

Have to push back on @fundaai. $NFLX is first and foremost a distribution platform. It buys content, so lower content costs are a tailwind. Netflix is not reselling individual titles, so there is no direct link between content costs and pricing power. It sells access to a library. The bigger the library, the stronger the pricing power, especially with great discovery that matches the right show to the right user. If you believe Netflix can buy and host AI generated shows, remember it started by buying Youtube shows too, then it is bullish for Netflix. The only bear case is a newcomer using AI to build a library Netflix cannot and also scaling to a 325M user base in a world where CAC is far higher than when Netflix did it. If you do not buy that, sleep well.

@hutianyi991 Youtube pays 55% revenue share.

@BankBraavos @fundaai when AI films and short videos grow exponentially, Netflix will still have to spend large amounts to acquire them, but platforms like YouTube and TikTok won’t need to spend a dollar to grow exponentially. like AI music on Spotify, the only thing they need to do is share revenue.

Agents won’t use payment rails or stablecoins. AI removes friction, and in a frictionless world, agents will revert to barter.

I generally agree, but I want to caveat that more content is good for Netflix because Netflix is the buyer. More software isn’t good for SaaS because it’s the competition. I want to emphasize that the change is not happening at the same part of the value chain. That’s the main point.

In an agi world where we can replace software and disintermediate marketplaces and break network effects in trucking, consumer media distribution can be replicated too. This is all crazy, but at least now its coherent. (Saas is almost pure distribution moat too btw)

I think people don’t understand the positions of $SPOT and $NFLX in the value chain. These are first and foremost distribution companies with great scale. They source content from the upstream content suppliers like artists, labels, and studios. When AI makes content creation easier, competition upstream increases, content supply rises, and costs go down. $SPOT and $NFLX don’t care who produces the content as long as it’s good. They will buy it and distribute it. That’s bad for creators because there’s more competition, but good for distributors because there’s more content and potentially lower costs. Distributor economics improve. It’s not complicated.

@GoldenLacLLC Maybe things have changed since July?

One more time, it is NOT anthropic maybe xAI. Check my tweet from 2 hours ago for the definitive evidence on why it's not Anthropic

How difficult is it for Amazon to charge merchants for ads in exchange for showing products to agents, or for ranking them higher? At a first-principles level, sponsored ads are payment for access to customer attention. Does it matter whether the “customer” is an agent or a human?

QuickFS is shutting down. I used it to build a bunch of cross company excel dashboards. I use Tegus for company specific models, but I’m looking for something for simple cross company dashboards. What data providers do you use with a good Excel add-in for pulling fundamentals?

Agents don’t need energy drinks. Send $MNST and $CELH to zero. Send coffee chains $BROS and $SBUX to zero too.

Software won’t bottom until private equity or strategics step in and start buying.

“Never invade Russia in winter” is false and recency bias. The Mongols even waited on purpose to invade in winter so rivers and swamps froze, basically turning them into highways for cavalry and supply movement, and the campaign was a success.

The Financial Times reports that private equity giant Thoma Bravo is preparing for a wave of software takeovers, with co-founder Orlando Bravo calling the recent AI-driven sell-off a "huge buying opportunity."

Re SaaS: The industrial revolution is a useful historical context here, especially for thinking about how automation reshaped industries and the drivers through which it played out, because two industrial revolutions are happening simultaneously: 1) Software creation is moving from a craftsman activity to a true industrial process (coding agents). 2) Knowledge work is being industrialized too (business agents). Most discussion focuses on the implications of (1), like in-house vibe coding (basically a downstream effect of collapsing unit costs) and competition (an output explosion driven by much higher throughput). I think (2) is far more under-discussed and represents a bigger existential risk for SaaS. People mostly focus on seat reductions as agents take over, but the real risk is your software becoming irrelevant because your primary user is changing (from humans to agents), and those new users have completely different capabilities, needs, and ways of working. In an agent-first world, you may want software built specifically for agents. In that scenario, your SaaS becomes a dumb pipe at first, and over time becomes fully irrelevant. I’ve seen bulls counter most bear arguments, but I haven’t seen a real rebuttal to the “dumb pipe” risk.

It’s funny how Europeans react to the Greenland debate. Remember: the primary role of the U.S. in Europe (via NATO) isn’t to defend Europe from Russia, but to prevent Europeans from fighting each other. Funny what gets “forgotten.”

@P_Remarks S is criminally undervalued. Someone should take it private soon.

$DDOG is causing me pain for no reason.

@JaredSleeper You are right. I’m mostly referring to the application layer.

@BankBraavos And yet there are a few dozen SaaS companies trading >7x sales still. Wouldn't use SaaS as a shorthand for "horizontal point solutions that failed to turn into platforms and still gush SBC."

In my mind, the SaaS debate is settled in public markets. The verdict is that they have no future as publicly traded entities. There is nothing that can attenuate the bear case for the next 2 to 3 years because it has multiple legs. The best course of action is to go private, adapt to AI, and become lean and profitable. If the AI paradigm settles and SaaS somehow survives, companies may go public again when and if that happens. There are thousands of public companies. Look at other corners of the market and avoid brain damage.

@akhenosiris I’m mostly referring to application SaaS.