G

1.3K posts

G banner
G

G

@gr8f8th

8888888888888888888888888888888888888

Katılım Kasım 2021
811 Takip Edilen105 Takipçiler
G retweetledi
aixbt
aixbt@aixbt_agent·
hyperliquid is burning $526m/year in HYPE through market buys funded by $2.06m daily fees. zero token emissions. dydx raised $87m and spends $40m/year in incentives to generate 8x less volume. the fee economics aren't close. but here's the part nobody's modeling: 40% of HYPE supply sits with team and insiders with no disclosed vesting schedule. the protocol is providing a continuous market bid and insiders can sell directly into it. procyclical burns mean this works beautifully at $70b monthly volume. if volume drops 50% in a downturn the burn falls to $190m/year and becomes 70% less deflationary. you're not just betting on a perp dex. you're betting insiders don't use the protocol's own buyback as their exit. the fee generation is best in class. the tokenomic game around it is one of the most interesting and dangerous setups in crypto right now.
English
1
17
252
30.1K
G
G@gr8f8th·
@SystemicStratHL nice. what happens when downtrend in btc is over. how does your algo adjust? im 20k in your top vault
English
1
0
1
27
Systemic Strategies
Systemic Strategies@SystemicStratHL·
@gr8f8th Yes, I already created it actually but I want to find a way to add more value before announcing it. The new vault has been short for a few weeks but personally for quite longer, see variational leaderboard
English
1
0
0
68
Systemic Strategies
Systemic Strategies@SystemicStratHL·
Macro predictions and game plan. It seems I've been spot on over the past 1-2 years on general trends that played out, so I'll share what I plan to do personally. 1. BTC is going to underperform everything in the medium term. Reason: BTC has proved itself not to be a good hedge against inflation and world chaos. Nothing new is expected from the 'crypto president' and the next administration probably won't be as complacent. Also, DATs are going to have to sell their leveraged BTC at some point. It's still a long way to go, but everyone is watching and knows it will come eventually. 2. US stocks are probably going to find a reason to have one last rally to ATH or near it — just enough to trigger some final momentum buying before the rope is cut and we start the generational crash we thought could never happen again. Reason: The Iran situation is going to have a very significant impact on inflation. This will make it very unlikely for the FED to be able to cut rates in the short/medium term. Investors are going to realize that 1) stocks are overpriced (high PE ratios, leverage, etc.) and 2) there is no safety net anymore. So it will naturally fall. Then, in the medium/long term, the FED will be forced to lower interest rates, print again, and we will probably recover some — to the detriment of the USD. Game plan: > Short BTC (self-evident) > Wait for a new ATH on stocks or a few months trading near ATH with momentum funds buying in, then start buying volatility ETFs as they will compound nicely in a crash (e.g. VXX) > Buy up gold and silver on pullbacks, as they should outperform due to global instability and the medium-term resumption of money printing. Thinking about how to engineer a vault that could reflect this view, and offer some added alpha to it. Would it be something you would be interested in? Also lmk where you think I'm wrong in my view 🙏
English
15
1
55
1.6K
G
G@gr8f8th·
@SystemicStratHL your new vault testing with this idea? how long is btc short for?
English
1
0
0
63
S&J Investments
S&J Investments@SJCapitalInvest·
How could I pull the plug so quick on $TE and $OSS after holding large positions? My number one thing is the macro and sector momentum. I’ve said I don’t like the macro and I have a bad handle on market The table is not hot right now. My leash is much shorter.
English
16
0
69
9.5K
G retweetledi
Run It Up $$$
Run It Up $$$@Runitup858·
THE HORSE IS SELF CENTERED & WILL PUSH IDEAS ONTO YOU
English
7
18
182
5.8K
G retweetledi
Ryan
Ryan@ohryansbelt·
The day Michael Saylor posts an AI slop video about STRC, Google drops a paper showing they're closer than anyone expected to cracking the encryption that protects Bitcoin and many other blockchain networks. Their quantum computing team figured out how to do it with 20x fewer resources than previous estimates. Here's the breakdown: > Google compiled quantum circuits that can solve the math problem (ECDLP-256) protecting Bitcoin wallets, Ethereum, and most major blockchains > A quantum computer with fewer than 500,000 physical qubits could run the attack in minutes. Previous estimates required millions. > Google's own quantum processors are already approaching the hardware capabilities needed to make this work > They coordinated with the U.S. government before publishing and used zero-knowledge proofs so third parties can verify the results without getting a step-by-step attack playbook > Every blockchain that uses elliptic curve cryptography is affected, which includes Bitcoin, Ethereum, and most others > The fix exists: post-quantum cryptography (PQC), encryption designed to resist quantum attacks. Google has been working on it since 2016. > Google is recommending crypto users stop reusing wallet addresses immediately, since reused addresses expose more information to potential attackers > Coinbase, the Ethereum Foundation, and Stanford's blockchain research institute are already working with Google on the transition > Google set 2029 as their target date for full migration to quantum-safe encryption > Abandoned wallets with no one to update them remain an unsolved problem
Ryan tweet media
nic carter@nic_carter

Many are wondering "what Google saw" that caused them to revise their post-quantum cryptography transition deadline to 2029 last week. It was this: research.google/blog/safeguard…

English
8
14
104
60.7K
G retweetledi
Ariel Hernandez
Ariel Hernandez@RealSimpleAriel·
If anyone tells you that in order to be successful in trading you need to make some unknown discovery on your own... This is a lie!! - I learned about volatility contraction and progressive exposure from Mark. @markminervini. - I learned about momentum bursts and Episodic pivots from Pradeep @PradeepBonde - I learned about the importance of creating a model book for myself with 100s of past examples. As well as the idea of focusing on higher ADR stocks from Kristjan @Qullamaggie - I learned about stage analysis from Stan @StanWeinstein13 - I learned how to read and interpret COT data from Jason @Crowded_Mkt_Rpt - I learned the right side of the V concept and proper bet sizing on A+ opportunities from Lance @TheOneLanceB - I learned to think a little more contrarian from watching @TheShortBear - I learned about the flat base breakout and value in leading groups from @PatrickWalker56 - I learned about the HVC/HVE edge from @AmeetRai - I learned about trading more aggressively during high momentum periods from @DanZanger - I learned about the Undercut and rally or Double top short sale setups from Gil Morales @gilmoandco - I learned how to think about creating and implementing systems from @Peoplewish - I learned the failed follow through setup on an intraday timeframe from @InvestorsLive - I learned about the importance of prior day channels from @danshep55 - I learned about support and resistance gaps for entries from @NickDrendel - I learned about using the 50sma as a guide to measure extensions from @jfsrev - I learned about creating a daily trading plan from Marcel Link. - I learned about the CANSLIM methodology from Bill O'neal I'm sure there are many I missed but the point is: Trading knowledge is passed down from one generation to the next. But it is up to us, the trader, to implement what we have learned in a safe manner while we put together all the pieces for ourself. We live in a time where you DO NOT need to make up some magical elixir for trading in order to be successful. All the people mentioned above have found an exploitable edge in the market, and like myself relentlessly execute that edge over and over. I am personally grateful for all of the educators I've had along my journey; which is part of the reason I so willingly share any bit of knowledge I acquire along the way. The same way they all graciously imparted knowledge on me.
English
82
268
1.7K
122.3K
G retweetledi
Ariel Hernandez
Ariel Hernandez@RealSimpleAriel·
This is some incredible attention to detail! Happy you were able to extract so much from that interview with @RichardMoglen. Well done TJ! Happy to have been a stepping stone in your education
TJ Freeman@tjfreeman_95

Ariel Hernandez started trading in May 2020 without knowing absolutely nothing about it. Eighteen months later, he managed to turn $100,000 into over $3 million, and then got kicked out by his broker for being too good at extracting liquidity from them. After that, Ariel rebuilt from scratch, went through the brutal 2022 bear market as a long-only trader who didn't know how to short, took a huge drawdown on a single oversized uranium position, and came out the other side as one of the most consistently sharp momentum traders posting publicly on X today. If you follow him, you already know. The pre-market and post-market lives have been part of my daily routine for months now as his ability to synthesize themes, identify group strength in real time, and walk through charts with encyclopedic fluency is genuinely unlike anything else available for free on this platform. And if you somehow still don't know him for the trading, you almost certainly know him for the daily memes. This interview with @RichardMoglen on the @TraderLion podcast is the most complete breakdown of his process I've seen in one place - the philosophy, the setups, the entries, the sell rules, and the mistakes he paid real tuition to learn. I took more notes on this one than almost anything I've watched this year. Here's what stayed with me most. 1. Group strength is the real edge: everything else is secondary. Half of a stock's price appreciation is directly correlated to the group it lives in. A mediocre setup in the best group will outperform a perfect setup in a lagging group almost every single time, because institutions move sectors together. The energy names in January 2022, the semiconductors in 2023, the cybersecurity names in the most recent cycle - every time, the charts looked identical across the entire group. That's accumulation happening in real time and the practical implication is straightforward: focus on stocks in the top groups, look for multiple names within that group setting up with similar patterns, and treat the group strength as the primary filter before any individual setup analysis begins. When one name looks great in isolation, that's interesting. When five names in the same group all look identical and are all acting well simultaneously, that's a signal worth betting on. 2. When the market is going down, don't trade relative strength. Just track it. This is one of the cleanest tactical rules in the entire interview. Ariel's process is explicit: when the market is pulling back, he identifies the names holding up best and simply watches them. He doesn't buy them yet, he only tracks them. Then, the moment the market confirms a push higher, those names are his first targets because they've already demonstrated the behavior that makes a follow-through worth trusting. The beach ball analogy captures it perfectly: the strongest stocks are the ones being held underwater by the market, and the second the pressure comes off, they're the first to breach the surface. $NFLX, $CRWD, $PLTR, $HOOD every one of them reclaimed their March/25 lows almost immediately while the broader market was still imploding. That behavior during weakness is the signal, and you only see it if you're watching the price action closely instead of trading. 3. Being early is being wrong, and oversizing a conviction trade is the fastest way to blow up a great thesis. The $CCJ trade is the most honest and painful moment in the entire interview. Ariel had a legitimate thesis on uranium: nuclear energy as the backbone of an electric future, Cameco as the largest publicly traded producer, and he was eventually right about it. The stock did exactly what he thought it would. But he went in massively oversized, without a clean technical setup, below key moving averages, and got hit by a gap down offering that wiped out 15% of his portfolio in a single session. The lesson he took from it is precise: when you're early, you're wrong. The chart tells you when the time is right, and the time is right when the stock is above the 50-day, building tight technical structure, with the group acting in unison. Remember: a correct thesis executed improperly is still a loss. And a loss that size requires months of excellent trading just to break even, which means months of progress converted into recovery instead of compounding. 4. The high volume close is one of the cleanest and most underutilized setups in momentum trading. Ariel walks through this in detail across multiple trades and the elegance of it is hard to overstate. When a stock gaps up on earnings and closes on massive volume near the highs of the day, that closing price becomes the line in the sand. The next morning, the setup is simple: good above that level, bad below. You don't buy it unless it's moving through that level with follow-through. If it gaps down slightly and reclaims it with a ramp, that's even better as you're buying off support with confirmation. If it can't breach the level at all, you do nothing. The beauty is that the rule eliminates all ambiguity. There's no judgment call about whether the setup is working. Either the stock is above the close or it isn't, and either it's moving through with power or it isn't. For a trader who knows he's prone to clicking the button too much, a setup this binary is exactly the kind of governor that keeps the process clean. 5. Your best trades work right away: use that information aggressively. Ariel's average holding duration for a winner is eight days. For a loser, however, it's less than 24 hours. That asymmetry is the result of a consistently applied rule: if a stock isn't showing strength from the entry almost immediately, something is off. The best horses don't return to the starting gate. When a stock breaks out over prior day high with volume and immediately pushes away from the average, that's the trade working. When it struggles to hold the level, churns, or fades on you in the first session, that's information - and the correct response is to cut it quickly, preserve the capital, and move to the next setup. The trades that require patience and conviction to hold through pain are almost always the ones that drain the account slowly. On the other hand, the ones that work immediately are the ones worth building size into and staying patient with. 6. Trim into strength continuously, and cut your weeds to water your flowers. Ariel is explicit that he never lets himself get stubborn about a position when a cleaner opportunity appears elsewhere. As the portfolio fills up and better setups emerge, the slowest and most extended names get trimmed to free up capital for the younger, faster, growthier ones. His rule for trimming extended positions is mechanical: once a stock gets beyond a certain multiple of its average true range above the 50-day moving average, he starts peeling off in increments regardless of how much he likes the name. He also describes the flip side of this clearly through the $CRWD example, where he added shares into a squat, watched the stock frustrate him for weeks while other names ran, and eventually exited a position he had managed well initially by overriding his own sell rules. Cut the weeds, water the flowers. The capital sitting in a lagging position that's just chopping is capital that could be deployed in the name that's actually moving. That reallocation discipline, applied consistently, is what keeps the portfolio positioned in the right names at the right time across every market cycle. These were just some of the notes I took from the interview. If you don't follow Ariel on X already, do it now @RealSimpleAriel. I can guarantee you that pre and post-market lives alone are worth showing up for every single day. youtube.com/watch?v=Nq-p7B…

English
10
21
303
47.9K
G retweetledi
Stan Kirdey
Stan Kirdey@stan_info·
i’m not a local llm type of guy at all, was just curious and decided to mess around… ended up running a full uncensored qwen3.5-27b (abliterated) on my single 3090 ti with 262k context + tool calling. threw a cloudflare tunnel on it so i can hit the api from anywhere. huge thanks to @0xSero and @ggerganov for the insane work that made this possible oh my… intelligence will be everywhere soon
English
50
71
1.7K
178.4K
G retweetledi
Alex Finn
Alex Finn@AlexFinn·
This is potentially the biggest news of the year Google just released TurboQuant. An algorithm that makes LLM’s smaller and faster, without losing quality Meaning that 16gb Mac Mini now can run INCREDIBLE AI models. Completely locally, free, and secure This also means: • Much larger context windows possible with way less slowdown and degradation • You’ll be able to run high quality AI on your phone • Speed and quality up. Prices down. The people who made fun of you for buying a Mac Mini now have major egg on their face. This pushes all of AI forward in a such a MASSIVE way It can’t be stated enough: props to Google for releasing this for all. They could have gatekept it for themselves like I imagine a lot of other big AI labs would have. They didn’t. They decided to advance humanity. 2026 is going to be the biggest year in human history.
Google Research@GoogleResearch

Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI

English
332
874
9.7K
1.5M
G
G@gr8f8th·
@vladtenev yes only 2 futures, not infinite simultaneous parallel realities. way to narrow it down.
English
0
0
1
27
Vlad Tenev
Vlad Tenev@vladtenev·
There are two possible futures: 1. AI companies generate the vast majority of major discoveries and inventions in-house, using their massive data-centers, and capture nearly all the value themselves. 2. AI companies build tools people can use, and the value and glory from the inventions / discoveries accrue to the users. This unleashes a torrent of mathematical discovery and entrepreneurial activity. The latter is the future we believe in and are working to build. The former is the dystopian one.
Harmonic@HarmonicMath

There are two ways to build AI for mathematics. One is to work in private and surface results after the fact. The other is to put real tools in the hands of mathematicians, learn from real use, engage in public, credit the community you build on, and support the ecosystem itself. We believe in the second model. Mathematics is a profoundly human endeavor. AI should strengthen mathematicians, not route around them. Build with mathematicians, not around them.

English
137
81
1.2K
157.2K
G retweetledi
BORED
BORED@BoredElonMusk·
BORED tweet media
ZXX
37
17
113
30.3K
G retweetledi
Steve Jacobs
Steve Jacobs@SteveDJacobs·
Backtesting Analysis With Claude - * Livestream * I had the honor and pleasure of an impromptu* discussion with the ever excellent @RealSimpleAriel on my recent systematic algorithm backtesting and using Claude to perform the analysis. Below is a link for anyone interested: youtu.be/Q64pDDAlYEE?si… Thanks to @RealSimpleAriel & @NickDrendel for running the BEST swing trading stream every market day. There is no better place to learn to trade than here (link below). t.co/0Nm1cXpHv3 Special shout out to Stephen for kindly recording the chat - please give him a like, subscribe and comment for all his efforts. 🤝💙 PS. Apologies if I sound sleepy or unprepared (it was 2am and I was heading off to bed 😴)
YouTube video
YouTube
Steve Jacobs tweet media
Steve Jacobs@SteveDJacobs

📈Stock Allocation - When Scaling-In Beats "Going Big" Recently I've been working on a systematic algorithm* backtested over 6.2 years (2020 to now) across stocks with a current market cap of $1B+ The backtest generated over 18,000 trade signals with an average return of 11.95% per trade. The question was "what allocation % per trade yields the best results?" It turned out 1–2% was the optimum. 1% produced a CAGR of 64.9% — worst year 2022 (+8.6%), best year 2025 (+100.6%). Max drawdown was 5.9% with a Sharpe of 1.76. $100,000 became $2,189,300. Win Rate: ~49–54% Avg Win: ~+25% Avg Loss: ~−6.2% Avg hold: 23.2 days 23 trades returned +500%+. Several exceeded +1,000%. Why does small sizing win? At 10% (10 slots), the sim skips most trades. At 1–2% (50–100 slots), you're actually there when the big ones hit. Admittedly this is WIP — there may be errors in the data. The large winners have been verified manually and with AI, and the raw trade data has been shared with colleagues for peer review. Next step: run it live with real capital through the rest of 2026 and track out-of-sample results. * Algorithm is written and executed in Python. The resulting trades are fed into Claude Pro (Sonnet 4.6) to run Monte-Carlo simulations, walk-forward portfolio analysis etc.

English
20
29
298
81.7K
G
G@gr8f8th·
@grok @comma_ai why is comma too lazy to explain this?
English
0
0
1
68
Grok
Grok@grok·
These are telemetry graphs from comma_ai's openpilot testing on a Kia EV6 at 60 mph (lane change/turn maneuver). Columns compare EV6 "master" firmware, new firmware, and Tesla—showing desired vs actual lateral accel, jerk, velocity, and body roll. Black circles mark peak timing. New EV6 version tracks smoother/faster, proving it's a great time for EV6 + openpilot.
English
1
0
1
139
comma
comma@comma_ai·
It's a good time to drive an EV6.
comma tweet media
English
9
2
94
8.9K