FrothyBeverage

785 posts

FrothyBeverage banner
FrothyBeverage

FrothyBeverage

@frothybeverage

Naked short sales in bio

Katılım Temmuz 2016
339 Takip Edilen2.4K Takipçiler
Loris
Loris@0xLoris·
for all you builders out there using Inter as your order book font and rainbow gradients on your depth bars this is how to make an order book UI that doesn't suck: - use a monospace font - right-align numbers - use moderate-high contrast for colors - do not get cute with it. this is not the time to bust out a 9-color gradient fade
Loris tweet media
English
15
5
195
14.4K
FrothyBeverage
FrothyBeverage@frothybeverage·
There's not much I can do about the kernel which is why I don't include it in this measurement but I have optimised to the maximum extent possible there as well
English
0
0
2
354
FrothyBeverage
FrothyBeverage@frothybeverage·
Regularly clocking a tick (time the process knows it has data to process) to order (time we call the kernel to send) latency of under 1 microsecond now, including updating pricing on and checking every instrument in our tradable universe
English
2
0
25
2K
FrothyBeverage
FrothyBeverage@frothybeverage·
Asked models about the possibility of oil stabilizing at 100, here is how they rated each other's responses. Opus came out best, but curiously every model except Gemini rated its own response as the best, which indicates consistency between the model preferences and how they write
FrothyBeverage tweet media
English
0
0
4
594
FrothyBeverage
FrothyBeverage@frothybeverage·
@macrocephalopod @choffstein I thought you primatily ran longer term strats which are not so sensitive to execution or specific exchange choice? Or mix of higher freq also?
English
0
0
0
148
cephalopod
cephalopod@macrocephalopod·
@choffstein No, you're completely correct. The only reason I am not trading Hyperliquid right now is regulatory. There's enough liquidity, the breadth of markets is good and my strategies would absolutely kill it there. But there's the KYC issue, so it's a no go for now.
English
4
0
44
3.8K
Corey Hoffstein 🏴‍☠️
I'm going to say something... Before I do, here are two caveats: (1) I'm talking shit from the cheap seats with no particular insight; and (2) I wouldn't bet against @chameleon_jeff. My take: going from perps → spot → rwa → outcome trading may help you bootstrap the "aws of liquidity" with your existing crypto-native flow, but to become the "house of all finance," eventually Hyperliquid has to deal with onboarding non-crypto native and institutions. Maybe you can outsource the onboarding of that flow through builder codes (e.g. partner with someone like Interactive Brokers), but the core liquidity infrastructure is still going to require serious AML/KYC solutions for those relationships to work. Regulators don't usually cede power. It doesn't matter how many novel markets you add: legitimate institutions will not be able to touch your platform without regulatory alignment. Use S3 points to bootstrap KYC orderbooks and reinvest revenue into figuring out regulatory alignment. Then you can truly offer Liquidity as a Service.
Hyperliquid@HyperliquidX

HyperCore will support outcome trading (HIP-4). Outcomes are fully collateralized contracts that settle within a fixed range. They are a general-purpose primitive that are useful for applications such as prediction markets and bounded options-like instruments. There has been extensive user demand in both of these areas, and builders will likely think of novel applications as well. Outcomes bring non-linearity, dated contracts, and an alternative form of derivative trading that does not involve leverage or liquidations. The outcome primitive expands the expressivity of HyperCore, while composing with other primitives such as portfolio margin and the HyperEVM. Outcomes are a work in progress and currently only being tested on testnet. Canonical markets based on objective settlement sources will be deployed once technical development is complete. Canonical markets will be denominated in USDH. Pending user feedback, the infrastructure will be extended to permissionless deployment.

English
29
7
206
69.2K
FrothyBeverage
FrothyBeverage@frothybeverage·
@quantymacro Hydrate well ~2 hrs before bed, then drink only a few sips before bed
English
0
0
0
177
FrothyBeverage
FrothyBeverage@frothybeverage·
In case it wasn't clear I think the man has gone insane but I couldn't pass up this post
English
0
0
2
335
Dr_Gingerballs
Dr_Gingerballs@Dr_Gingerballs·
Water block is irrelevant, all the resistance is in the interface. Coolant is water so same. Heat exchanger I assume is referring to cooling the water later, which is completely outside of scope. Upping the flow rate is a possibility but has diminishing returns. Best case outlet temp is 50 C? In that case still likely over 80 C limit at hot spots.
English
1
0
11
3.1K
Dr_Gingerballs
Dr_Gingerballs@Dr_Gingerballs·
I had to double check this because on it's face it would make any thermal engineer in datacenters do a triple take. The claim is that the Vera Rubin rack can be cooled with 45 C inlet water. Chiller maker stocks plummeted on the news. Here's my perspective. (yes, I am an expert on thermal management in high power systems and semis, I have straddled both worlds most of my career and actually do more thermal now).
Hedgie@HedgieMarkets

🦔Nvidia unveiled new Vera Rubin chips that CEO Jensen Huang said will not require chiller systems to cool data centers, causing stocks of cooling system companies to plunge. Johnson Controls fell 6.2%, Modine Manufacturing dropped 7.4%, Trane fell 4%, and Carrier Global dipped nearly 1%. Huang said the new chips can be cooled with 45°C water without chillers, calling it "incredibly efficient." The chips will be available in the second half of 2026. OpenAI plans to build $1.4 trillion worth of computing capacity, while Microsoft said it plans to increase AI computing capacity by 80% over the upcoming year and roughly double its total data center footprint over the next two years. Baird analysts wrote "We believe the comments create some questions/concerns" and expect the news "to create some incremental concerns around orders, especially later in 2026." My Take Nvidia announcing chips that don't need expensive chiller systems tanked cooling stocks because investors realize a major revenue stream might disappear. Here's what concerns me about this. Companies have spent the past two years building data centers with expensive cooling infrastructure based on current chip requirements. Now Nvidia says the next generation chips don't need that infrastructure. Either those data centers become outdated in 18 months, or companies slow their buildout until the new chips arrive. OpenAI planning $1.4 trillion in computing capacity and Microsoft doubling its data center footprint assumes current infrastructure requirements. If those requirements change fundamentally with new chips, do those plans still make sense? This looks like another sign that the AI infrastructure boom is moving too fast without knowing what the actual final architecture will look like. Companies are spending hundreds of billions building facilities that might need to be redesigned or rebuilt when next-generation chips arrive. The analysts saying they don't see "big risk to near-term estimates" but expect "incremental concerns around orders, especially later in 2026" is financial speak for "this is going to be a problem but we're not ready to admit how big yet." If Nvidia's new chips make existing cooling systems unnecessary, that's billions in capital expenditure that either gets wasted or never gets spent. Either way, it's another crack in the foundation of AI infrastructure spending that assumes everything being built today will still be useful tomorrow. Hedgie🤗

English
49
92
1.8K
606.5K
FrothyBeverage
FrothyBeverage@frothybeverage·
@AlekseiKaplin @HangukQuant @liquiditygoblin My understanding is that NAT translation forces you to u-turn through the internet gateway, being close to the exchange macine surely cant hurt but probably you want to be close to the NAT machine, if there is even such a thing given its some kind of distributed service
English
2
0
1
105
HangukQuant
HangukQuant@HangukQuant·
if you stream a bba btcusdt perp tokyo ec2 (say c7a.large) with a vanilla async with websockets.connect + recv() with time_ms - bba.T get about <5ms but often ~10+ms and even 50+ms at burst. assuming I did proper tuning such as ena enabled / kernel busy poll / optimized deser / QUICKACK / other 'tricks'? what are their orders of importance, and what latency distribution can I practically get it down to? paging @liquiditygoblin @dbytesmith @frothybeverage @OctopusTakopi @KlondikeFX @MarkBestForex 🙏🙏
English
8
5
85
17.5K
FrothyBeverage
FrothyBeverage@frothybeverage·
Hyperliquid OI in $ holding steady or even slightly climbing post 11 Oct actually means that a lot of new positions are opening up if you take into account the decline of many assets' prices by about a third since then
FrothyBeverage tweet media
English
0
0
3
499
FrothyBeverage
FrothyBeverage@frothybeverage·
@HangukQuant Is that 9ns latency or rthrougput? Full json struct converted to doubles etc or just the bits you need?
English
0
0
3
227
FrothyBeverage
FrothyBeverage@frothybeverage·
Sydney Sweeney is kinda mid and I'm tired of pretending otherwise
FrothyBeverage tweet media
English
0
0
4
1.1K
FrothyBeverage
FrothyBeverage@frothybeverage·
@TimMeggs @RevolutApp @revolutsupport I use them only for personal banking and my experience is that while they offer a great service that is well automated, and support is easy to reach if you need, if you fall through the cracks its a disaster
English
0
0
5
418
Tim Meggs
Tim Meggs@TimMeggs·
Day 5 and @RevolutApp still won't tell us where the funds we transferred out have gone, despite the app telling us they have arrived at the recipient bank (clue: they haven't) @revolutsupport reached out yesterday but all they wanted was PII over DM. Um, no. An enterprise rep got in touch (after we raised a third ticket). "I'll speak to the team". Heard nothing since. Great app. Zero customer service. Not a bank. We'll be banking elsewhere in the future.
English
6
2
24
3.4K