Recursivist

[Recursivist @RecursiveL00p]

I struggle with the view that open source winning is somehow good for the AI ecosystem on a multi-year basis.

My understanding is that there are significant efficiencies in tokens per compute with each passing year on a given model (we are more efficient at serving GPT 4.0 then we were a year ago).

If we begin to hit frontier economics to the point where OAI/ANT are no longer incentivized (or cannot access capital) to train bigger and bigger models, why would progress on model size and capabilities not drastically slow.

If we get this setup of meaningful pressure on frontier economics in say early-'27, you would likely have a 6 month tailwind from lower-cost open source proliferating driving increased token usage (Jevon's Paradox), but then rapidly offset within 12 months by the rate of model size/capabilities rapidly slowing (allowing efficiencies on compute to really take effect).

This would then look very similar to the dot-com setup where internet users kept growing but the stocks all cratered as we had overbuilt capacity (and efficiencies began to take effect).

[Recursivist @RecursiveL00p]

@acemoney21 @jaytech877 Yeah I agree if by end of next week and definitely by end of the month we've made no progress have to change the view

[Recursivist @RecursiveL00p]

Agree on the intraday trading, feels like the markets sole goal is to burn theta and 0dte's after retail options volumes hit an all-time peak (June/July)

That said, market as a whole feels like its doing bear market structure (slow moves down and rapid rises) within a bull market (making higher highs on each impulse move). Think its too early to call the end of that (but the market structure clearly seems late stage cycle to me and we're debating when, with my view this comes to a head more in 1Q'27).

Looking at SPY inverse'd honestly helps the framing imo (doing a retest of the prior lows at 750 on the inverse chart right now), but we did break out of the downtrend from the March high (would ideally like to see a retest of that with SPY around 780-800 for a short setup into midterms)

[Recursivist @RecursiveL00p]

@chiefingza DVOL making new lows, BTC slowly drifting downward off a local high, ZEC making new highs. Seems awfully similar to the May 25th setup

[Recursivist @RecursiveL00p]

@Wild_Randomness Fake-outs complete and the road to 800 begins now

[Recursivist @RecursiveL00p]

@Wild_Randomness Expecting more fake-outs next week, but we appear to be tracking

https://t.co/ppvAt9KEPa

[Recursivist @RecursiveL00p]

@Wild_Randomness What if we hike and go higher anyway into Ant IPO. Imagine the scenes

[Recursivist @RecursiveL00p]

@acemoney21 @DynamicMoats He was why they were profitable ha (buying friday/tuesday), basically topped with his buys

[Recursivist @RecursiveL00p]

I agree MAGS are the cleanest. Positioning is clearly way offside in AI (as seen even in this weekend by the degree of whipsawing in sentiment on what I view as a relatively insignificant development).

However, think the next month or two of AI fundamentals probably going to be hard to ignore (Incremental ARR added likely to meaningfully accelerate in Sep and Ant probably drops an even stronger model heading into Oct to show signs of continuation into Oct). Also seeing some signs of token usage growth on consumer which I still view as low-value but helpful to keep things going (Meta now potentially getting rewarded for their spending)

Current base case is AI puts in a lower high (but QQQ puts in a new ATH on the back of MAGS) into EOM before a pullback into midterms. I do ultimately expect AI stocks to make new ATH in 1Q-2Q 2027 (and this may coincide with OAI finally going public) before its all over, but would not be surprised if the move happens after those Leopold contracts expire ha (January expiry)

[Recursivist @RecursiveL00p]

@acemoney21 @lfg_cap It's AI and MAGS post-FOMC imo, RSP/SPY looks broken down (I think you short any lower high when oil comes off with little hesitation)

[Recursivist @RecursiveL00p]

I'm all for AI being a bubble, but the idea that they'd all agree to just blow up the IPO's a month before it's slated seems quite unlikely (and gives them too much credit in terms of their morals imo).

All his article really says he wants some sort of regulatory body to monitor developments and particularly focuses on reducing China's access to chips/distillation, he explicitly says it does not mean slowing training. People getting ahead of themselves with pricing driving narrative rn imo.

[Recursivist @RecursiveL00p]

Fable was held back by ZDR (Fable 5.1 already increasing Fable penetration despite it being way overshadowed by Astra) in part imo.

I think that very well may end up being the right conclusion (returns to larger training runs will begin to slow), just think there's no chance they are concluding that right now (I think they all understand that without their consistent ~6mo edge vs open-source, they don't have a real business)

[Recursivist @RecursiveL00p]

@negligible_cap @RHouseResearch They're turning people away from $200 subscriptions, I'd be amazed if they're coming to the conclusion that there isn't ROI right now...

Just seems like a performative statement/slight shot at Anthropic

[Recursivist @RecursiveL00p]

@BigTrout300 BTM

[Recursivist @RecursiveL00p]

@Faisal_alHodhan ZIGHER

[Recursivist @RecursiveL00p]

@chartsMovesNews Think it could be around mid-Sep (Clarity act, fomc, japan interest rate decision), or potentially sooner?

[Recursivist @RecursiveL00p]

@BillyTheKid_59 Would prob be the most bullish outcome if right. Definitely seems like the ipo timing delay is to try improve their positioning (right now it's decelerating arr adds and second best model) heading into it

[Gavin Baker @GavinSBaker]

Anthropic pre-IPO gamesmanship post. Pure speculation but sharing as curious for thoughts.

Anthropic shifted from gross to net ARR accounting and stripped out both Meta and Chinese distillation from their $65 billion ARR number. Meta speculated to be over $5 billion in ARR so taking them out means they can easily weather it when Meta turns them off shortly after being public, which is widely expected. Also decreases the odds of Meta turning them off, watermelon quality dependent. All smart.

Then release Fable 5.1 so OpenAI feels confident releasing Astra. Vibes here on Astra are really good btw. I think that Astra was probably better than Anthropic was expecting.

Now there are whispers that Anthropic has solved Navier-Stokes, which would be super impressive.

Anthropic probably releases Fable 5.2, which should be better than Astra unless something is awry, sometime before the IPO. Likely also planning on showing a significant reacceleration in ARR in September which will of course leak to the press.

Grok 4.7, Meta’s Watermelon and ChatGPT 6.1 all likely coming in the next 6 weeks as well. All those labs are confident about their roadmaps in a way I have not seen in the last 18 months. And we will see about Gemini 4. Competitors get a vote in all these plans.

Grok Bot feels like the best agentic harness yet for enterprise use cases and Instinct is a promising agentic harness for consumer use cases. Should see variations of both from competitors soon. Grok Bot remains transformational for my use cases.

And all this is happening into a continued acceleration in overall AI demand.

Wild times.

As an aside, I think Krishna might turn out to be an exceptional CFO. His former Blackstone colleagues speak super highly of him. Going to be important as communicating clearly to Wall Street if they decide to shift their compute from inference to training will be difficult to digest the first time. Probably worth studying Amazon’s invest and then “check-in” margin strategy from 2010 through maybe 2016, which investors eventually understood.

[Recursivist @RecursiveL00p]

@Faisal_alHodhan @GlitchCapital You think it won't go back to 81k or think it goes higher than his stop?

[Recursivist @RecursiveL00p]

@Wild_Randomness What if we hike and go higher anyway into Ant IPO. Imagine the scenes

[Recursivist @RecursiveL00p]

@institLPGP @faryarshirzad @coinbase @SECGov Agreed on still needing to build off its own clients, think its why a partnership with Kraken was the most optimal route for them (rather than trying to build this on their own from scratch even if they do have a good marketing engine).

[Recursivist @RecursiveL00p]

Think this may just be a delayed burn. If they can't get a bid within the next week though, we may indeed be cooked.

A separate dynamic at play here that I am not sure how to read is Anthro setup into IPO appears subpar (seems like they overearned earlier this year by being first to coding and having the best model, but may be now losing their edge despite broader AI token trends still encouraging).

A bit of a weird setup given they are the one who's IPO'ing (though longer-term I think OAI winning remains the most bullish outcome as it confirms the theory that compute is the most valuable resource etc).