Chu Bo Yan retweeted
Advertisements are also easy RL environment like coding and math because of the verifiable A/B testing mentioned (number of conversions, clicks, etc).
Ads AI (Meta AI) should be as valuable as coding AI (Claude)
$META
Ben Thompson says the real money in AI is being made by Meta, not OpenAI or Anthropic
"The biggest impact of the models, the biggest monetization right now is probably not in Anthropic or OpenAI. It's the incremental gain that is happening for Google and Meta."
"They can actually validate their image creation and their text creation in a way no one else can. Their validation is the ad marketplace, like running a gazillion A/B tests to see what works. Most ads don't convert.”
"What do LLMs do? LLMs predict. We're going to move to this world where Meta is going to look at people and say, this person probably wants to see this next. And they're going to go find that thing and show it to them."
"The potential upside in terms of just showing people better ads that are more relevant to them, they only need to increase like a few percentage points for the returns to be billions and billions of dollars. This alone is worth them investing in being on the leading edge."
"This is why advertising is great. And Meta's advertising in particular is awesome. I get frustrated that Meta doesn't talk about that. Mark Zuckerberg has never really talked about the societal benefits of advertising except in passing in 20 years."
Chu Bo Yan retweeted
🚨 Word is that Ilya Sutskever's SSI began benchmarking its upcoming model at a certain prominent AI evaluation startup last month
Chu Bo Yan retweeted
OpenAI and Anthropic currently have more demand than compute. Since GPU capacity is limited, they need to maximise how much revenue and gross profit they earn from every GPU-hour.
If they can maintain high margins, they can achieve higher ARR and can justify trillion-dollar IPO valuations to raise more equity, secure even more compute and extend their frontier lead.
Meta can disrupt this by selling near-frontier intelligence at a much lower price.
Because compute is already constrained, OpenAI and Anthropic cannot simply cut prices and make up for it with unlimited additional volume. They would either have to keep prices high and lose market share to Meta, or cut prices and weaken their ARR growth, gross margins and IPO narrative.
Meta can afford to tolerate low margins on its models because its advertising business generates enough cash to fund the datacentres. It does not necessarily need to make large profits from AI immediately.
So Meta’s strategy may be to start a price war, weaken the labs’ economics and slow down their ability to raise capital and expand compute. This prevents OpenAI and Anthropic from becoming economically unassailable and gives Meta more time to close the frontier gap.
$META
SNDK's gross margin plateau is a welcome move. It means they are not squeezing customers to the extreme, and that they are looking at a long term vision instead. The disciplined, balanced capex execution I have been arguing for all along, the kind that prevents a company from investing to extremes at the top of the cycle and taking severe damage in the downcycle, requires SanDisk's NBM model. There is no way around it.
So is now the right time to buy SanDisk? The bears argue that the gross margin plateau marks the peak of this cycle. That is also why the stock is falling right now.
As I wrote in my SK Hynix earnings review(You can find it in the Citrini Substack chat), this is the point where we should be looking at Q, not P. As long as you believe in the AI supercycle, as long as big tech keeps spending, as long as compute companies like Nvidia remain in short supply, there is no scenario where the shortage eases for memory alone. That is the corresponding effect of LTAs (NBM) and disciplined capex. Even if gross margins stall, as long as Q expands and the floor set by those LTAs prevents an excessive collapse, the total profit pool will grow enormously.
Layer massive shareholder returns such as buybacks on top of that, and memory companies should be rerated, and they will be rerated. From cyclical stocks to essential infrastructure of the AI world.
$SNDK
Chu Bo Yan retweeted
TLDR my $META thesis:
(1) Hyperscalers should see meaningful FCF inflection from past AI capex, but META is the main exception.
(2) META's AI success hinges on building a true frontier model to sell intelligence directly while owning the infrastructure. Success would let META monetize both AI products and capex, control upstream and downstream, and generate superior free cash flow.
(3) META holds advantages with more training-focused compute, massive engineering talent for real-world applications, and vast distribution.
(4) If META reaches frontier later this year, it could trade $1,000 to $1,500 as markets price in outsized returns.
Chu Bo Yan retweeted
Judging by my tl there is a growing gap in understanding of AI capability.
The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code.
But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along.
So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions.
TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.
Chu Bo Yan retweeted
(1) I do not care if the market does not believe in AI ROIC yet. That just means there is more money to be made before everyone wakes up. Current models are already insanely capable, and this is before Rubin, vastly larger training clusters and another wave of software improvements. We have crossed the Rubicon: Trillions of dollars are being invested for datacentres, and the leading labs are raising what may be their final major rounds because they believe the endgame is close.
(2) Once agents can automate AI research and improve the next generation of agents (Which i believe to be next year 2027), progress will not move linearly -it will bring forward by orders of magnitude.
(3) That is why I am not worried about Chinese or open-source models commoditising the frontier. They may commoditise yesterday’s capabilities, but once the leading labs reach escape velocity, open models may never catch up. Frontier labs also do not have to release their strongest models or expose them for distillation, so the capability gap could become insanely wide. At that point, China may be forced to rely increasingly on espionage, stolen weights, chip circumvention or pressure on US infrastructure simply to avoid falling permanently behind. Meanwhile, cheaper lower-end models will still expand AI adoption and increase demand for compute, memory and inference. The winners will be those who own the stack , models, chips, foundries, datacentres and distribution
(4) While companies on the wrong side of AI risk getting sonic-boomed into dust.
$NVDA $META $TSM $NBIS $SFTB.NE $SKHY $SNDK
Chu Bo Yan retweeted
My take is that Meta should double down on this. Meta has the best and smartest people and the resources to do so. Hit FUD articles prevent Meta from raising more equity and giving up on frontier AI. My view is that Meta is much closer to the frontier than most people think and has the ability to do so.
Engineers who think that creating high-quality training/eval data that existing models cannot solve well is the “Gulag” should have their fitness questioned if that lazy entitled engineer is supposed to be at Meta in the first place.
Meta has 80k of the brightest people and should mobilize that army to do things that build tremendous value for the company.
$META
Many people think any given ML project is 99% training.
In reality, it’s 50% evaluation, 40% data cleaning, 8% integration, and 2% training.
The first two set the noise floor for learning. No ML magic matters; the model cannot lower the noise floor, as that’s the optimal bound of Shannon encoding of your data.
Thus, not a single day goes by without me thinking about ontology. Even the old labels have to be constantly reviewed.
Chu Bo Yan retweeted
LFG! Huge respect to Zuck for being willing to let free cash flow turn negative, take the PR hit, and push through early employee backlash. True founder-CEO and man in the arena!
My take is that Meta should double down on this. Meta has the best and smartest people and the resources to do so. Hit FUD articles prevent Meta from raising more equity and giving up on frontier AI. My view is that Meta is much closer to the frontier than most people think and has the ability to do so.
Engineers who think that creating high-quality training/eval data that existing models cannot solve well is the “Gulag” should have their fitness questioned if that lazy entitled engineer is supposed to be at Meta in the first place.
Meta has 80k of the brightest people and should mobilize that army to do things that build tremendous value for the company.
$META
Chu Bo Yan retweeted
Replying to @oldschooldegen @ManuInvests
Value will accrue at the model layer as compute cost eventually decreases and model improves. $Meta has the rich proprietary data, across social and intent graphs that no other companies have. TAM will expand much further and meta will capture all that value into the stock price
Chu Bo Yan retweeted
A bull cried so much before being slaughtered that the earth beneath him became wet with his tears. 💔💔
He struggled, he trembled, he looked around with fear in his eyes as if begging for mercy, as if asking one silent question
What wrong have I done to deserve this?
His tears were real,His fear was real,His pain was real.
When a helpless living being stands before us shaking with fear, crying for life, trusting us to spare him… can ending that life truly be called an act of faith?
Can any religion ask for compassion on one side and ignore tears on the other?
Before the knife falls, before blood touches the ground listen to the cry of that bull.
If his tears can wet the earth, shouldn’t they also touch our hearts?
Will such slaughter in the name of religion ever be justified?
Chu Bo Yan retweeted
Interesting comparison. Like to hear some thoughts from @evrgn11112231
Taken from Bytedance fundraising deck May 2026
$META
Chu Bo Yan retweeted
Street consensus $Meta is priced for 17% average earnings CAGR for the next 5 years. Stupidly cheap (8.6x 2030 EV/E) compared to the other Mag 7.
Street is betting:
1. Earnings grow average 17% (I think will be higher)
2.Earnings are not durable/ uncertain ( hence 8.6X 2030)
I think both assumptions will be wrong. $Meta will more than 2x from here
Chu Bo Yan retweeted
What $meta FUD bears are claiming lol:
1. Meta lacks a public cloud business, so it cannot monetize excess AI compute like Google, Amazon, or Microsoft.
2. AI capex has a less direct payoff path, because Meta must earn returns mainly through its own apps, ads, and engagement.
3. Meta is late on custom ASIC chips, which could leave it with higher AI infrastructure costs.
4. Free cash flow pressure is legitimate, as heavy AI spending may reduce cash generation even if the core ad business stays strong.
5. User growth may be saturating, meaning future growth may depend more on extracting higher revenue per user than adding new user
Chu Bo Yan retweeted
Replying to @oldschooldegen @evrgn11112231
Key risks
1. AI execution flops and fail to monetize and earnings compress via depreciation
2. “Tobacco” litigation
3. Hormuz risk affects consumer spending spillover to ads revenue decrease
Chu Bo Yan retweeted
Long thesis
1. $META is the best applicator of AI and it will show up in ads revenue growth
2. Value will accrue at the model layer as compute cost decreases and models improve. Meta is best position to capture this value
3. Insta + reels + social moat intact
@evrgn11112231

















