Life, in the end, is a journey to discover your own alpha.
Joined July 2009
- Tweets252
- Following57
- Followers52
- Likes894
Puymoon retweeted
gossip (read) priority fees on hyperliquid are broken
currently the team that is benefitting the most is wintermute (insert fist shaking red emoji guy meme)
this is not publicly known but can be determined with certainty by mapping out transfers to labeled wallets from the top winner of the gossip priority fee auctions (they have won 79% of the slot 0 auctions in the past 24h)
tldr; wahhh wahhhh wahhh hl gossip priority, not fair. add more slots or scrap the gossip auction
my proposed solution:
1) increase gossip priority to 20 equal-priority slots
2) make it so winning an auction slot grants IP whitelisting to a foundation-run peering slot
i think this basically solves it and would likely *increase* total gossip priority fees
---
some context
the purported intention, as stated in the docs, is to "reduce the need for market makers to hyper-optimize their infrastructure to remain competitive"
if you win the gossip auction slot, you get order book / node data 50ms (!!) faster than anyone else.
yes. one IP, across the *entire network*, gets the feeds 50ms faster. it's the whole ballgame
so in theory you could just buy the faster data slot instead of spending 10s of thousands monthly on high octane ec2 instances and [insert hft magic]
or you can do BOTH, which is what wintermute is doing
so despite the stated intentions it does the exact opposite and gives 1 wallet a dominant advantage over every other participant
further reducing the democratization of this mechanism is the fact that, practically speaking, you can only use a won gossip slot if you have a direct peering relationship with the foundation node.
multiple providers we work with couldn’t pass gossip priority benefits through to us without slowing their other peers (all independently confirmed this behavior - strange design)
and to get a foundation peering you to do 0.5% of the 14d HL maker volume (we don't quite meet this threshold so our application was rejected) as well as stake 10,000 HYPE ($780,000 here), or be like, a node service provider
---
without revealing too much i can say with 100% certainty that wintermute has maxed out BOTH the infrastructure optimization + expenditure AND is monopolizing the gossip fee auction
they're spending on average $26,000 *per day* ($10M/yr) to win this auction
so rn they're just eating everyone else's lunch
and the only way that approach is economic is if you are MM at the scale that they are, which only a handful of wallets are doing
Puymoon retweeted
마크롱이 예전에 한 말
“프랑스의 문화는 없다. 프랑스 안에 여러 문화가 있을 뿐이다”
이러한 주장에 대한
알렉스 카프의 관점.
1. 모든 것을 허용하는게 옳다는 방식의 사고방식은 결국 국가 정체성에 혼란을 가져오고, 무엇이 옳은지에 대해 이야기하는 것 조차 논의를 시작할 수 없도록 만든다
2. 결국 국가의 공익을 위해 정작 필요한 것, 공유되는 가치관에 대한 이야기의 공백이 발생하고, 이를 시장이 다 채워버린다.
3. 시장에서 소외된 못 가진자들이 결국 “나는 옳은데 왜 나는 못 가지는가”에 대한 불만을 갖게 됨
4. 시장주의가 팽배한 포용력없는 사회에서 이런 사람들을 받아줄 순 없고, 결국 이런 사람들은 불법적이고 폭력적인 모습이 되어가고 사회는 더욱 분열된다.
Puymoon retweeted
The presumption that the Fed raising short-term rates reduces inflation is predicated on the belief that higher rates reduce demand and investment.
But what if higher rates don’t reduce demand and investment because the demand for intelligence and energy is unaffected by higher rates because winning the race for super intelligence has a near infinite ROI and the demand for compute will remain incalculable.
Why won’t higher rates at this unique moment in history therefore lead to more inflation as interest costs are embedded in everything?
And the problem is compounded as the more the Fed raises rates, the more inflation we will have and the more the Fed will need to raise rates further and so on.
But what if the old models don’t apply to the current paradigm and the Fed is wrong?
I think the Fed might have just made a mistake. Am I right or am I wrong?
Puymoon retweeted
Today we announced the Claude-led discovery of a molecular machine that we suspect could represent a new gene editing mechanism. Its precise function, biotechnological utility (if any), or level of significance is not yet clear, but at minimum it is work I would have been proud to do as a PhD student. The work was done mostly, though not entirely, by Claude: our life sciences team suggested a broad area of research, Claude read through the literature and a bunch of genome data and discovered something interesting, then Claude proposed experiments to verify the discovery and our team carried them out.
It’s easy to dismiss this as a one-off or curiosity, but we’ve repeatedly seen a pattern where AI performance in new intellectual domains goes from weak to superhuman in a matter of a few years. In 2023 models struggled to do math at the level of an average high-school student. In 2024 they started to do well on math competitions for the best high-schoolers in the country, in 2025 they started to solve minor open problems, in early 2026 more significant open problems, and in late 2026 they are beginning to solve the top few open problems in all of mathematics. We believe AI for biology is on a similar exponential trend.
The main difference between biology and mathematics, of course, is that math can be done purely theoretically, while biology requires experimentation. Some have used this to draw the conclusion that AI’s utility in biology will be limited. We think this is wrong. As we’ve demonstrated today, humans can collaborate with AI to perform the experiments, validate key results in a few weeks and, if necessary, work with the AI to iterate on what they find. Eventually it may even be possible for Claude itself to safely perform the experiments by autonomously controlling lab equipment, with appropriate safeguards in place, but we aren’t doing that today (our lab is also a BSL1/BSL2 facility that doesn't handle materials dangerous to humans).
More broadly, biomedical advancement has many stages — from fundamental biology discoveries, to translational research, to drug discovery, clinical trials, and finally the actual delivery of medicines and health care to patients. We are also interested in these later stages, but even simply accelerating the first stage of fundamental biological discoveries has the potential to speed up and broaden the entire pipeline. Improving our understanding of biology and sharpening biologists’ tools can drive forward all of the later stages, for example by identifying new drug targets, finding new therapeutic modalities, allowing for more precise measurement, and speeding up the experimental loop which itself further accelerates our understanding of biology. This will not in itself speed up clinical trial times, but if it succeeds it could greatly increase the number of promising candidates that go into the pipeline — an increase in throughput even though latency remains.
In Machines of Loving Grace, I wrote about AI’s potential to “cure most diseases in 5-10 years” — a goal that sounds impossible, but one I believe is just barely possible if AI is applied to every stage of the pipeline. The first step is showing that AI can first help with, and then drive, biological discoveries.
Claude’s discovery is the latest in a line of related prior work that goes back decades, beginning with systems like CRISPR, and continuing with discoveries like the bridge recombinase and VIPR in the past few years. Recently, there has been heightened interest in systems based on reverse transcriptase (RT) enzymes, the enzyme underlying the system Claude identified. And most recently, a Stanford team working independently described a novel RT system with an associated non-coding array that is in some ways similar to the one Claude found, though they are distinct systems that evolved independently from each other. I believe that we’re at the very beginning of finding such systems and developing them into powerful tools for biotechnology.
I’m proud of the resources Anthropic has invested in accelerating the public benefits of AI through the life sciences, and we’re aiming both to grow our life sciences team and to work with other scientists to extend this approach to a broad range of problems. If you have a proposal for a research collaboration or are interested in joining our life sciences team, please reach out.
Claude has discovered a previously unknown enzyme system hidden in the DNA of bacteriophages. Beside the enzyme’s gene sits a long array of repeating DNA—a structure that looks somewhat similar to CRISPR.
We don’t yet understand what this system does, but only a handful of known systems share its features, and all of them are able to cut, copy, and paste DNA. Historically, the discovery of such programmable systems has helped revolutionize medicine. CRISPR, for instance, is now the foundation of genetic medicines. But it will take much more work to learn what this system does, and whether it can be put to similar use.
Read more: anthropic.com/news/claude-di…
어제 오늘 AI 진영에서의 충격적인 발표로 다들 난리인 거 같다.
다들 이걸 두고 달리는 기차에서 내려버린 사건이라고들 이야기 하는데.. 정말 그들이 그런 의도 였을까?
오히려 반대 아닐까? 증류 등으로 인한 기술 유출을 막고, 입지를 굳히기 위한 전략이라고 생각.
» 사이버보안·바이오·AI 통제 상실 위험을 낮추기 위한 취지지만, 감사·인증·보안 및 규제 대응 비용 증가는 자본과 인프라를 갖춘 OpenAI·Anthropic보다 스타트업과 오픈소스 진영에 더 큰 부담
Puymoon retweeted
According to Korean media reports, Samsung Foundry's Taylor fab has begun operating earlier than the originally planned November start and has recently reached a utilization rate of 30%.
Utilization crossed 20% as of last month and has approached 30% this month.
Samsung Electronics held an equipment move in ceremony at the site in April and completed a tape out in July, the step that hands a chip design over to actual wafer fabrication.
A tape out is normally followed by prototype production and verification, but Samsung is said to have pulled its mass production schedule forward to keep up with surging AI chip orders. With the foundry lines at its Hwaseong and Pyeongtaek campuses in Korea, which handle the leading edge nodes (2 to 4nm), running at full capacity, the company accelerated the production schedule at Taylor.
The customer that triggered the early ramp at Taylor is Tesla. The key product to be made there is Tesla's next generation AI5 chip. Tesla, which is accelerating its investments in full self driving, robotics, and AI data centers, is understood to be demanding a rapid supply expansion of AI5, which it plans to use across the Model 3 and Model Y, the Cybercab robotaxi, the Optimus humanoid robot, and its AI data centers.
With Taylor utilization, currently around 30%, climbing quickly, some in the industry expect it to reach 100%, effectively full utilization, by around year end.
Alongside raising utilization at Taylor, Samsung has also begun preparations to build a second Taylor fab, targeting groundbreaking by year end. The industry speculates that the second fab, expected to come online around 2030, could introduce Samsung's next generation 1.4nm process.
The expansion at Taylor also ties into Samsung's broader strategy to strengthen its foundry business. Taiwan's TSMC, the industry leader, has seen most of its advanced node capacity absorbed by a flood of AI chip orders. Samsung plans to take advantage of this, using the early ramp at Taylor as a springboard to absorb Tesla's volume reliably while pursuing additional customers. If Taylor utilization rises quickly, the foundry division's return to profitability is also expected to come sooner.
Puymoon retweeted
“모호함에는 가치가 있다.”
<MS 글로벌 채권 헤드 Andrew Sheets>
*01
저는 이맘때를 좋아합니다. 공기는 조금 선선해지고, 수많은 학교에서 새로운 학년이 시작되면서 새 옷과 새로운 친구들을 만나는 설렘이 있습니다. 그리고 무엇보다 드디어 미국에서는 미식축구 시즌이 시작됩니다.
2025년 미국에서 시청률이 가장 높았던 TV 방송 100개 가운데 무려 90개가 미식축구 경기였습니다. 세상이 갈수록 분열되고 콘텐츠 소비 역시 파편화되고 있다는 점을 생각하면 놀라울 정도의 집중도입니다.
미식축구가 이렇게 인기 있는 데는 여러 이유가 있겠지만, 시간이 지날수록 제가 높이 평가하게 된 특징 중 하나는 이 스포츠가 가진 전략적 복잡성입니다.
그중에서도 가장 중요한 것이 바로 ‘모호함의 가치’입니다.
경기장에 나온 선수 구성만 보고 상대방이 러닝 플레이를 할지 패스 플레이를 할지 알 수 없도록 하는 것, 누가 블리츠를 들어올지 혹은 아무도 들어오지 않을지를 숨기는 것처럼 말입니다.
코치와 선수들은 가능한 마지막 순간까지 선택지를 열어두기 위해 노력합니다.
그리고 9월에 접어든 지금, 이런 전략은 미식축구에만 국한된 이야기가 아닙니다.