underclassman retweeted
Replying to @balwiles
Well, we’ve already seen sporadic outtages in parts of the globe. Japanese banking systems, individual cities going dark in Europe, refrigeration failures across Europe, etc. So the question is really how bad it gets and if it’s like to hit your specific area.
IMO, I’m very confident we’ll see at least some major infrastructure “break” defined as a 7+ day outage in key infra like electricity, telecom, stock market, etc. The US is safer than most areas since we get the best cyber tools in the world to harden key systems and have better teams to bring stuff back online. But US is also a bigger target than most.
My guess: 90% that at least 200m people experience an outage like that described above within 2 years. 70% that 1 billion people experience it. 40% that 4 billion+ people do.
How about a much worse outcome? My base case is we get WW2 level devastation. Meaning 100m+ dead and a continent in disarray for years. I’d put that at ~60% to happen in the next 4 years.
How about much worse than that? I’d put odds of 1b+ dead in next 4 years at 20%.
These are all guesstimates ofc, built more on high level mental models than acute estimates.
underclassman retweeted
Replying to @AlexanderGrieve
The crazy part seems to me to be the federal regulator. Kalshi knowingly violated state laws. And those state laws existed for good reason. Have you watched a Kalshi ad? Fraudulent marketing preying on the weakest and most vulnerable. Coupled with the market manipulation and fraudulent volume numbers, it’s reasonable to call kalshi a criminal enterprise that exists to extract cash from consumers with deception.
But given the federal regulator’s support, I half agree with you. Unfair for a founder to be caught between two regulators. But then I go back to the fraudulent marketing…
underclassman retweeted
Replying to @star_okx @IamRamenPanda
Anyone still fooled by CZ at this point deserves it. He’s a serial scammer, fired for fraud, jailed for a tiny fraction of his crimes. He escaped a far longer jail sentence by giving DoJ dirt on everyone in crypto and bribing Trump.
underclassman retweeted
jensen could've been more responsible, but no, he chose fame & greed, & for that he will lose everything
underclassman retweeted
Replying to @TensaiCapital
i meant in the medium term (6-18 months), AI infra baskets are a bad bet because:
1. deep learning recipes are extremely simple
2. US labs are in a red queen's race that now involves China, which will weaken moats & compress multiples
3. the US will enter a recession within this timeframe
> Even if intelligence becomes commoditized, it will take WAY more compute than we currently have or even what is currently planned/forecasted to saturate demand
this is why i said 'AI will be completely transformative' & limited the 'AI infra bad bet' take to 6-18 months in my original response
imo SOXX/DRAM will underperform until we get close to densing/efficiency limits, & i also expect many bottlenecks to be routed around/managed to various extents (also, AI is now good enough to accelerate the search for the best & most efficient deep learning recipes)
> the counterargument being that getting to ASI first is maybe worth several trillion vs being 2nd or later
imo the ASI trillions meme might not be beneficial to shareholders (market participants will treat it cautiously) because the US Govt will likely immediately nationalize the entity that gets to ASI first -- they likely won't be allowed to sell or possibly even use the model the way they'd like to (they'll also lock it down to prevent leakage to China)
> This would definitely disrupt the long AI trade, but why do you think it will happen?
imo poor stock performance & rising unemployment rate will feed on each other, & eventually lead to single-digit multiples on a lot of marquee US stocks
underclassman retweeted
This kind of thing will have an enormously chilling effect on enterprise uptake of open weight models.
Fact: an open weight model running on domestic servers on your own hardware can still contain an encoded instruction to periodically “phone home” and share sensitive data it comes across with China (or anyone else). This can be inherently undetectable as the instructions can be hidden in the model weights. It’s possible that the instructions could survive finetuning or RLHF.
Such attacks have been demonstrated empirically and are entirely plausible.
For this reason, I don’t expect any serious enterprises will run Chinese open weight models even if it’s entirely their hardware stack. It’s too easy for these to be a vector of IP theft.
There will have to be a class of domestic open weight models.
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected.
Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions.
Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent.
But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models?
We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem.
Read the full report: cohere.com/blog/concerning-b…
underclassman retweeted
Replying to @per_asp3ra
real patriots know that billy is the hero of the story and the longhouse henpecker schoolmarm seyfried women is the villain
underclassman retweeted
Replying to @woke8yearold
GDP ain't gonna move until you have effectively closed loop supply chains.
underclassman retweeted
Jeff drank a bit too much of his own Kool-Aid in this recent Odd Lots clip. I think it's important to clear up a distinction that makes all the difference.
Democratizing access to wealth creation is NOT the same as democratizing access to speculation.
Many of us genuinely want broader participation in attractive investment opportunities. The growing wealth inequality produced by the current system is tearing at the seams of society, something I've spoken about many times.
But let's be clear: listing pre-IPO perpetual futures does absolutely nothing to democratize the wealth being created in private markets.
I've seen this misleading narrative pushed repeatedly by perp DEXs lately, and it needs to be called out.
A pre-IPO perp creates ZERO underlying wealth. It's a zero-sum contract before costs, and negative-sum after them. No capital flows to the company. No ownership changes hands. No new economic opportunity is created.
Every long has a short. There is no free lunch in simply going long something before its IPO, and funding costs can even determine the outcome more than any price moves. TRADING access is NOT INVESTING access.
In fact, in many ways pre-IPO perps are almost the antithesis of democratized access:
They create the illusion of participation in private-market wealth creation, while actually offering retail a speculative instrument whose mechanics they are often poorly equipped to understand.
The result is a playground for the FEW sophisticated traders who understand funding, oracle design, and liquidation mechanics to extract value from the MANY who don't.
To give credit where it's due, Robinhood's efforts toward sourcing actual employee shares, placing them into SPVs, and tokenizing genuine spot ownership are much closer to meaningful democratization.
That can unlock real economic value by allowing employees to exchange some of their expected upside for liquidity, while giving new investors actual economic exposure. A legitimate financial trade-off, without the funding-rate and liquidation shenanigans.
I'm strongly aligned with expanding access to investment opportunities. And I believe advances in AI and blockchain can dramatically lower the barriers to discovering, structuring, and distributing them.
But let's not BS ourselves that launching another zero-sum casino contract is moving the needle on wealth inequality.
What would actually move the needle is bringing real capital formation and ownership onchain. Giving investors access to productive, illiquid assets like data centers, GPUs, and select startup investing, where capital can actually finance growth and share in the resulting value creation.
These opportunities are harder to structure, underwrite, and bring to market, and we haven't seen many compelling implementations onchain yet.
Hopefully this will change soon and, as the industry matures, we can direct more of our energy toward actually expanding access to wealth creation.
Clip below. Was glad to see @tracyalloway push back on this framing, and to Jeff's credit, after this he does seem to walk back his original position somewhat.
Jeff Yan: Wealth Is Often Created Before a Company Goes Public
On October 15, 2026, Hyperliquid founder Jeff Yan @chameleon_jeff told Bloomberg's Odd Lots podcast @podcasts that when a company goes public at a valuation of several trillion dollars, most of the wealth has often already been created before its IPO.
Access to those investment opportunities is limited to a small group, reflecting the limitations of the existing financial system. He argued that barriers to participation should be lowered to give more people equal access to financial opportunities, alongside stronger financial education to help investors assess whether those opportunities suit their risk tolerance.
underclassman retweeted
I'm not sure if @naval and I are as directly at odds on this.
->There is the single-player game. Everyone can learn to self-develop, enhance their talents, and be an ideal version of themselves.
->Then there is the adversarial, more brutal efficiency game of societal impact. You can't just move at your own pace there. In this game, the brutal truth is without a natural headstart in a category you will never be able to reach your society's frontier.
Jordi Alexander doesn't believe everyone has a special talent.
He estimates only 10–15% of people can realistically push themselves to make a meaningful impact.
That puts him directly at odds with Naval’s philosophy that everyone has a unique talent.
Is Jordi being brutally honest about human potential?
But if you tell people that most of them simply don’t have what it takes, does that create realism... or kill the belief that they can become something greater?
underclassman retweeted
To anyone disappointed at how slowly this played out - my apologies. I’ve never “danced around” when challenging fraudsters before. But I also never challenged such a politically powerful and aggressive legal bully before. I was also super burnt out when this first happened. Lastly I had fiduciary obligations to prioritize fund recovery when that looked viable by negotiation. I’m feeling more energized now. I have no fiduciary duties nor conflicts of interest, so I’m going to try to push for the full truth to come out quickly.
underclassman retweeted
This blew up more than I expected, since I’ve written about this vaguely before.
I can’t share proof yet since multiple legal processes still unfolding and Coinbase already tried to sue me into silence (unsuccessfully, but I need to be smart about how I fight the 800 pound legal bully.)
Every word I’ve written is true, I have no ulterior motive, nor position long nor short Coinbase or related equity. I’m still heavily long crypto, so if anything, I’d be incentivized to help Coinbase cover this up as many industry leading firms have “for the good of the industry”, but I won’t do that. I also think of coinbase as one of the better companies in the industry and have many friends there. No vendetta here. But when the “best” companies in crypto are funneling billions in customer assets to Lazarus group and lying about it to investors, customers, regulators, and the secret service?
Coinbase has represented to investors and customers (including pensions doing official due diligence) that they were never hacked (at least as of 2 years ago, last I checked). Not only is that radically false, coinbase has shockingly extensive internal documentation of them knowingly “misleading” customers and investors about this, deliberately, over time.
I used the word “misleading”, but I’m being…gentle. @grok what is the legal term for a financial company inducing customers and investors to give them capital on the basis of deliberately false material financial representations?
For coinbase’ legal team. I’m not going to dare you to sue me again, but I’ll say this. It would be in the interest of the industry and the world…but not your executives for us to go through discovery and air all this out. I encourage you to assert this as libel so we can go through discovery and see if a single word of my writing is untruthful.
underclassman retweeted
Coinbase “lost” $25m of my firm’s a couple years ago. Turned out they were actually covering up massive and repeated hacks. Still wouldn’t return our money. We traced this to at least a dozen other affected firms and over $1b covered up. That’s all I can say for now as multiple legal processes still ongoing.
underclassman retweeted
underclassman retweeted
Looking to hire employee #1 for my startup, which will be launching publicly this summer.
The role is head of engineering. You're a good fit for the role if you:
- are an elite programmer
- upgrade your agentic coding workflow constantly
- have previously built software that interacts with blockchains or financial markets
- actively participate in financial markets and love trading culture (either crypto or tradfi)
- (optional but preferred) are physically located in the bay area or are willing to relocate there
If you or someone you know fits this description, shoot me a DM. And please RT for visibility!
underclassman retweeted
since the cost of building great things is collapsing, & great builders will only ship greater things at an accelerated pace, i'd rather focus on developing internal context while regularly sampling the frontier to maximally leverage the steady flow of great composable primitives













