spent a little more time than needed but here's my full thesis & positioning for 2026!
https://t.co/UIoV36QTI6
predictions:
→ $BTC surpasses $150K
→ consumer apps >> blockchain infrastructure
→ launchpads momentum will fade
→ crypto gambling and sports betting will return
→ on-chain equities will be the next major vertical
→ @circle will challenge @tether's dominance
→ robotics will thrive, just not on-chain
→ DePIN interest returns
→ infofi will become irrelevant
→ crypto M&A and PE activity will accelerate
If I was a really sharp unnamed trading firm
>Momentum trading becomes increasingly profitable, and a large percentage of gross book
>These names are extremely illiquid, and liquidating your holdings would kill the market
>See fund that is extremely concentrated in single names
>LP into said fund, knowing the size of the raise and forecastable market impact in single name equities, as well as risk limits (VaR, Vol, DD limit)
>Unwind illiquid single name momentum book into the fresh new buying
>When the market starts to nuke, you know exactly where said fund will begin getting liquidated
>Scoop entire book in one block transaction and buyback the names you sold into the fund's top blasting
Just maybe...
The rate hold is not some sort of bullish panacea. At best its just not "another" headwind.
All it means is crypto is still going to be driven by liquidity flow and Clarity. Equities will continue to pass around the hot ball of money while market determines if this is just a semis bubble pop or we've reached an overall tech/AI top. And all assets will be driven by moves in the dollar and bond yields.
Any one of these headwinds could become a major tipping point and lead to renewed downside.
FED holds rates and delivers hawkish stance on inflation. A new hiking cycle is simply not viable but I don't have a read on if they will do a single hike this year for 25 bps or when they will do this.
Seems more likely to hold and posture hawkish than being aggressive from inflation coming from geopolitical disorder. Equally an aggressive hiking cycle will impact the US largest growth vector which is the capex expenditure pouring into AI infrastructure.
Nothing burger then wait until September.
quick fomc blast:
- fed funds pricing 35%, poly/kalshi closer to 25% chance of hike. imho this is too high.
- supposing we get a hike: this would likely be a hike pulled forward from later in the year. presser would reaffirm that they remain data-dependent and looking primarily at employment & inflation to justify another hike.
- if they hold: continued jawboning.
- why hike odds overpriced: july inflation was quite soft. oil is +20% but given the relatively minimal impact from earlier in the year, warsh is unlikely to hike based off of this alone. iran war is also a switch that trump can switch on/off relatively nimbly.
- main clause ppl are parroting in support of a hike is credibility building. would it necessarily build credibility as fed chair to hike despite the data not supporting such a decision? i see this as unlikely.
- note that Warsh's main initiative he announced last meeting was various committees/task forces--relevant here are existing data usage, productivity & jobs, and overall inflation framework. it would make very little sense for him to act drastically (and i do think a hike here would be somewhat drastic!) ahead of those committees forming and returning with revised definitions/policy suggestions.
- you need 6 votes in addition to warsh's to pass the hike. based on appearances & speeches in july, imho they get to 4-5 at best.
- a hold prob gives a small bounce but skeptical that btc sustains. am inclined to believe ewy/kospi is closer to a bottom than nq here.
- would not panic if we get a hike. lot of downside pxed in over last week (and this morning) and it's possible we head lower until the presser before Warsh comes in and discusses the pulling-forward and clarifies the rationale around the decision
we got the full Kimi K3, 2.8T params, running on 80x RTX 5090s.
20 tok/s single stream, day one, untuned. Last week we took GLM-5.2 from 30 to 110 tok/s on this same fleet. This number will climb.
A first for open weights: frontier intelligence served with zero HBM, the scarcest silicon in AI. Just GDDR7 gaming cards, plain ethernet, and the official MXFP4 weights, nothing requantized.
The most powerful open model on Earth, on the most abundant GPUs on Earth. Any lab, startup, or university can now own it, probe it, fine-tune it, run agents on it.
@Kimi_Moonshot
A few AI-slop adjacent thoughts i'm happy to be told I'm very wrong about but increasingly feel strongly about:
- Sending a claude-designed deck shows a lack of attention to detail or care for craft that might be existential to company building. At a minimum, I am morally opposed.
- Pangram is actually just an IQ/skill test that shows how good someone is with AI.
The fact that it is wildly easy to "beat" means if you can't, you're so bad at editing that you can't edit it out or so lacking taste that you don't care.
Either way, great public good and if your essay gets a 100% read you should be banned from the site temporarily.
- If internally you use AI to write or surface data that you don't contextualize, that's fine, but you should have to disclose it.
The biggest problem with nomads is that they are broke ass people in their own country and just leveraging on their currency to stay solvent in SEA, which is why you never see any “digital nomads” in Singapore. Cos those broke asses can’t afford it.