Patrick Long retweeted
Trying to do 2 more pre-seed deals in US reindustrialization before end of year. We are early investors in @noxmetals, @DiracInc, Quintess, Sumeria Space, @YardRobotics. Also spent first half of my career in factories and studied mechanical engineering. Hmu 🫡
Patrick Long retweeted
Crypto has been under intense selling pressure the last few weeks.
While crypto fundamentals have been mostly positive (many banks announcing use of crypto and blockchains), price not followed.
Here are some reasons we see potentially behind weakness:
- AI/semi/$NVDA selloff triggered margin calls (particularly Asia retail)
- coming @AnthropicAI IPO a supply shock for equities (selling ahead so investors fund IPO purchase)
- $ANT to have heavy retail allocation @RobinhoodApp $HOOD (selling crypto to fund)
- funds from recent $6.3B hack recently released to account holders
- these account holders converting holdings to cash/ stables
Secondary factors:
- Anthropic roadshow talk about AI ‘breaking’ everything, all cyber defense and every company
- creates existential fears of anything technology or digital
- DoJ/ USA 🇺🇸 scrutinizing chip sales to Asia/China 🇨🇳 triggers concerns of crackdowns
- crypto sells off on fears of more bad Asia headlines
Our take:
Crypto remains the future of finance
- the settlement layer for transactions
- digital assets transform capital markets
- adds up to a ‘supercycle’ for crypto
- bigger than past cycles
The headwinds 💨 shall pass
$ETH is the future of finance
Patrick Long retweeted
re: fundraising
the best storytellers don’t specialize in turning non believers into believers.
they are capable of spotting the existing believers at scale and reeling them in. stories are mere catalysts of their own belief system.
Patrick Long retweeted
After setting a vision and inspiring others to join them, founders must then focus on setting an uncomfortably fast pace.
While it is an overgeneralization, the hull speed and standards of excellence do not improve from the first board meeting. The company's headcount grows, the business becomes more complex, processes creep, and coordination becomes a bottleneck.
Counteract it by:
1/ Running your weekly execution as speed workouts. Measure your inputs and outputs each week. Set the pace slightly faster each week.
2/ Breaking and reinventing processes. Process creep slows down companies. You complain about the long expense policy and employee handbooks; what processes have you created that are also guilty of this?
3/ Embracing and driving change. Think and act differently, and push your teams to achieve ambitious and near-impossible goals.
Patrick Long retweeted
Why big agent PRs are so hard to review
Subagents are a real improvement. Splitting a task lets agents work in parallel and keeps each one focused. But in most tools, the subagents report back to one orchestrating agent, and the orchestrator reports back to you once, at the end.
That means more work happens between the moments a person can see it. The faster the agents get, the bigger that blind stretch becomes, and the bigger the PR at the end of it.
The fix isn't fewer agents or slower agents. It's checking the work while it's being built.
lightsprint.ai/blog/verify-e…
Patrick Long retweeted
Phylax flips security from trying to catch every hole to proactively defining what shouldn’t happen (draining contract, minting infinite tokens, etc) and blocking transactions that would do that from settling.
Something like this is badly needed to address the defi safety problem. It’s why I was proud to become a tiny backer.
Patrick Long retweeted
Six years ago, we understood that the world’s financial infrastructure would never run on a single shared state machine.
We built Avalanche accordingly, and the rest of the industry is beginning to understand why.
Why Avalanche for NYSE?
It’s important to note that this was not a quick, monetarily-driven decision made without due diligence, as I’ve seen a few posts speculating that. After more than a year of evaluating and testing multiple blockchain networks, NYSE arrived at Avalanche for a multitude of reasons.
For those who have followed Avalanche closely, this shouldn’t come as much of a surprise. For everyone else, it’s worth understanding the fundamental difference between building for speculative crypto markets and building the infrastructure that underpins global capital markets.
For regulated financial institutions, high TPS and low transaction fees are the baseline. Most modern blockchains can offer some version of both, with the much harder problem being architectural.
Monolithic chains share consensus, execution, and blockspace. Conventional EVM L2s introduce dependencies on external settlement layers, sequencing, and data availability. Both models come with constraints that become particularly consequential when institutions need direct control over their infrastructure, especially when it comes to regulatory compliance.
Avalanche was designed around a different premise.
Avalanche L1s operate with independent consensus, dedicated validator sets, customizable virtual machines, configurable fee markets, and protocol-level permissioning. Institutions can define who validates, who participates, how transactions are processed, and how their networks are governed, with the flexibility to implement regulatory requirements directly into the infrastructure.
They can do so without inheriting the congestion or settlement dependencies of another chain, while retaining the ability to interoperate with other networks and access the broader Avalanche ecosystem’s assets and liquidity.
This combination of sovereignty, customization, performance, and interoperability is what makes Avalanche uniquely suited for true capital markets. It’s one thing to offer institutions a place to deploy applications, anyone can do that. It’s another thing to give them the ability to operate the underlying infrastructure on their own terms.
We’ve spent years building around the conviction that the world’s largest financial institutions would eventually need purpose-built blockchains of their own, rather than forcing entirely different businesses and regulatory environments onto a single shared network.
NYSE’s decision is a powerful validation of that thesis.
And I believe we’re only beginning to appreciate just how important these architectural decisions will be as global financial markets continue to move onchain.
Patrick Long retweeted
Replying to @_julianma
Security on the protocol, yes.
Safety for users... We have a long way to go.
But once we make it easy for normal people to safely use Ethereum, $ETH will hit $10k no prob.
Patrick Long retweeted
The annual assembly of the Kleros Cooperative! One more year building decentralized justice for the world!
Patrick Long retweeted
Most think vibecoding at scale is slop. Here’s a counterexample.
Today we release a 16M-line formalization of the entire Stacks Project, critical infra underlying modern math, with optimized code quality and its highly interconnected structure preserved.
A mathematical city 🌆
Patrick Long retweeted
Digital minimalism
vitalik.eth.limo/general/202…
Patrick Long retweeted
This is a fascinating (LONG) essay by @WillManidis that paints a lucid picture of the future. The TLDR, and the implications (I think?):
1) Labs are a cost center.
2) The upside will mostly go to whoever controls the channels AI reaches the economy through. Kind of like railroads and oil wealth going to the routes and pipeline owners vs. the inventors.
3) Capital allocators will have disproportionate advantages
4) The future of companies is much smaller, and they themselves will operate a lot more like capital allocators w/ models doing the bulk of the work.
Implications from this for normal people (I think):
- Don't try to pick AI labs. The authors argue that's what will be commoditized. If you invest in index funds, you effectively already own the AI labs anyway.
- Lean toward businesses that tax AI usage. Things like payments, distribution, data, regulated gatekeepers. The pipes basically.
- It was always a good idea to move from income to ownership long-term. But this is more fuel (and urgency) for doing so.
- Two places where you can "own a chokepoint" yourself are trust and relationships. Intelligence is cheap and access is expensive.
- Could explore things like profit-sharing to simulate the "paid like capital" model the essay describes.
- Build owned audiences (newsletters, substacks, personal brands, etc.) Become a conduit, not a vendor.
In short: invest in diversified toll-takers, spend your energy building chokepoints you control.
commanding-heights.com
Patrick Long retweeted
Introducing SCRUB-1: teaching a robot arm to support the scrub tech with instrument work. Tested on the bench.
We fine-tuned π0.5 from @physical_int on three tasks: 300 demonstrations each, 403 trials scored live.
Fine-tuning took the handoff from 0% to 79%.
Patrick Long retweeted
Replying to @ycombinator
@ycombinator's latest RFS highlights a breadth of opportunities, from robotics and specialized AI models to infrastructure and financing the physical AI economy.
@patricktlong and I have been angel investing both pre- and post-YC, with experience across fundraising, GTM, and growth. One of our favorite parts is helping founders navigate the early stages!
With W27 applications due November 2, we’re opening up a few office hours over the next few days. Whether you’re exploring an idea in these areas, refining your pitch, or looking for feedback on your YC app, we’d love to help and share perspectives.
Feel free to message us!
ycombinator.com/apply
ycombinator.com/rfs
Patrick Long retweeted
Replying to @patricktlong
@patricktlong crypto OG since 2011, and we’re excited to see what’s next, especially at the intersection of AI + finance. We’re opening up a few 30-minute office hours for founders building in this space. Would love to hear what you’re thinking about and brainstorm together. Feel free to reach out!
At @ycombinator, we're excited to fund more crypto startups solving financial needs including AI capital raising, stablecoins, trading, and agents.
Here's our latest crypto request for startups. Our next application deadline is Monday.
Patrick Long retweeted
A historic day for AI. Congrats to the @AnthropicAI team and the whole ecosystem! 🧡
Anthropic has filed an S-1 registration statement with the SEC related to its proposed IPO.
Read more: anthropic.com/news/s1-sec
Legal disclaimer: A registration statement relating to these securities has been filed with the SEC but has not yet become effective. These securities may not be sold, nor may offers to buy be accepted, prior to the time the registration statement becomes effective. This post does not constitute an offer to sell or the solicitation of an offer to buy these securities, nor shall there be any sale of these securities in any state or jurisdiction in which such offer, solicitation or sale would be unlawful prior to registration or qualification under the securities laws of any such state or jurisdiction.































