Michael D. White

27.1K posts

Michael D. White banner
Michael D. White

Michael D. White

@here4impact

Bittensor Fund I • CIO @ Crain • Impact PE • former Risk Arb @ Citi • BJJ black belt

Katılım Eylül 2013
2.7K Takip Edilen6K Takipçiler
Sabitlenmiş Tweet
Michael D. White
Michael D. White@here4impact·
migration from $TAO=>#subnets is healthy and will continue. $TAO halving is less than 2 months out (early Dec). yes, there will be subnet holders who take profits to sell $TAO as it goes 5-10x in the next ~6 months. however, there will be an *overwhelming number* of new entrants to the #bittensor ecosystem as that price action occurs. $TAO will dominate every chart and will lead the most popular narrative (AI). and do you know what people buying at $2,000/$TAO *won't* be happy with? a 2x. they will move out on the risk curve. they will buy subnets. and then we will see the real game begin. it happened on $ETH. it will happen on $TAO. nfa/dyor/good luck! the next 3-6 months will be life-changing for those positioned accordingly. nfa/dyor. charts courtesy of @latentholdings / tao.app $TAO $dTAO #bittensor
Michael D. White tweet mediaMichael D. White tweet mediaMichael D. White tweet media
English
58
83
337
33.3K
Michael D. White
Michael D. White@here4impact·
@andyyy @tplr_ai is dominating the airwaves rn x.com/tplr_ai/status…
templar@tplr_ai

On the @theallinpod this week, @chamath asked @nvidia CEO Jensen Huang about decentralized AI training, calling our Covenant-72B run "a pretty crazy technical accomplishment." One correction: it's 72 billion parameters, not four. Trained permissionlessly across 70+ contributors on commodity internet. The largest model ever pre-trained on fully decentralized infrastructure. Jensen's answer is worth hearing too.

English
0
1
4
153
Andy
Andy@andyyy·
What are the best subnets on Bittensor???
English
26
3
40
3.3K
Micaela
Micaela@micaelabazo·
if ur one of those people that doesn’t want to be too early or too late… this is ur time to shine ✨ what more do u need? they’re mind blown and off by 68B
templar@tplr_ai

On the @theallinpod this week, @chamath asked @nvidia CEO Jensen Huang about decentralized AI training, calling our Covenant-72B run "a pretty crazy technical accomplishment." One correction: it's 72 billion parameters, not four. Trained permissionlessly across 70+ contributors on commodity internet. The largest model ever pre-trained on fully decentralized infrastructure. Jensen's answer is worth hearing too.

English
2
3
34
703
Michael D. White retweetledi
tommy 🌙
tommy 🌙@CryptoTomYT·
Tradfi gonna go wild over Templar when they realize the same narratives trading in 100b+ range is sitting under 100m on a TAO subnet.
English
6
6
50
2.8K
Austin Barack
Austin Barack@AustinBarack·
Templar was just discussed on the @theallinpod in a conversation with the founder of Nvidia Jensen Huang as a technical breakthrough for AI training. Jensen also said he believes decentralized and open models to be just as critical as closed source models. $SN3 looks set to make a new high and go on a real run as the market starts pricing it to the size of the opportunity
Austin Barack tweet media
Austin Barack@AustinBarack

Templar completed the largest distributed and open LLM pre-training run in history. $SN3 is trading at a $40MM mcap. Right now it is not yet listed on any CEXs. A lot of alpha just researching what's interesting and leaning into onchain friction points before it is easy

English
9
1
49
7.2K
Michael D. White retweetledi
Zeus | SN 18
Zeus | SN 18@zeussubnet·
Here’s the first 2m temperature run since the Zeus upgrade went live. 🌍 Previously, miners were forecasting weather across scattered bounding boxes, thousands of times per day. Now, they generate full global predictions in four synchronized runs at 00, 06, 12, and 18 UTC, each with hourly resolution. Why? Industry standard. Now, it's directly compatible with downstream trading workflows.
English
3
7
42
2K
The Rollup
The Rollup@therollupco·
“Bittensor is recreating a lot of the early interest people saw during DeFi Summer, where multiple projects were launching.” Andy asks Synth founder James Ross for his bull thesis on Bittensor. Hundreds of millions of dollars could flow into the Bittensor ecosystem, and TAO could go 10x, 20x, or even 50x from here. We saw something similar with Templar, which generated major interest and helped drive TAO up roughly 30–40%. Will we continue to see this growth across the Bittensor ecosystem?
English
17
14
95
12.3K
Michael D. White retweetledi
Openτensor Foundaτion
The largest decentralised LLM pre-training run in history. SN3 @tplr_ai trained Covenant-72B across 70+ contributors on open internet infrastructure. Now it’s being discussed by @chamath with @nvidia CEO Jensen Huang. Distributed, open-weight model training on Bittensor is getting started.
English
30
162
694
23.2K
Michael D. White retweetledi
templar
templar@tplr_ai·
On the @theallinpod this week, @chamath asked @nvidia CEO Jensen Huang about decentralized AI training, calling our Covenant-72B run "a pretty crazy technical accomplishment." One correction: it's 72 billion parameters, not four. Trained permissionlessly across 70+ contributors on commodity internet. The largest model ever pre-trained on fully decentralized infrastructure. Jensen's answer is worth hearing too.
English
63
268
1.2K
170.5K
Michael D. White
Michael D. White@here4impact·
@Investing88ai's retail fund is launching by end of month. **you will need sn88 alpha to participate** I committed 100T to their strat on 12/22/25. 47% in < 80 days. I have not touched the wallet: all automated. absolutely banana-level insane. thank you for your attention to this matter. #bittensor $TAO $dTAO
Michael D. White tweet media
Investing88@Investing88ai

With 88's public retail fund launching in the coming weeks, 88 + 118 will be the major force that stabilizes the subnet ecosystem and drives growth x.com/Subnet118/stat…

English
7
2
45
6.3K
Michael D. White retweetledi
Whale.Guru
Whale.Guru@Whale_Guru·
🚨🚨🚨 THE BIGGEST MONEY PRINTING EVENT IN HUMAN HISTORY IS ABOUT TO START. THE NUMBERS WILL MAKE 2020 LOOK LIKE A JOKE. 🚨🚨🚨 Let me show you exactly what is coming. 💀 US national debt: $38.8 TRILLION — heading to $40 trillion 💀 US annual interest payments: $970 BILLION — nearly TRIPLED since 2020 💀 US deficit FY2026: $1.9 TRILLION projected — already $1 trillion in FIVE months 💀 Pentagon Iran war bill: $200+ BILLION — just requested TODAY from Congress 💀 Global debt: $348 TRILLION — all-time record high as of 2025 💀 Europe ReArm Plan: €650 BILLION unlocked for military spending 💀 Germany defense budget: $107 BILLION in 2025, up 24% year-over-year ⚠️ COVID stimulus globally was $5.54 trillion. This is already bigger and the printing hasn't even started. ⚠️ The US is spending $11.3 billion PER WEEK on the Iran war. ⚠️ At that rate, in one year that is $587 billion in war costs alone. Now look at what each country is about to do. 🇺🇸 UNITED STATES: $200B war bill hits Congress this week. Fed holds at 3.5-3.75% but cannot hold forever. When bond yields spike past 5%, they will restart QE to cap them. They have NO CHOICE. $39 trillion in debt means every 1% rate rise = $390 billion more in interest. 🇨🇳 CHINA: PBOC governor literally said this week: "we will flexibly and efficiently use rate cuts and RRR cuts." Already cut FX reserve ratio to 0%. M2 money supply set to expand ~7% this year. 🇪🇺 EUROPE: €650 billion ReArm Europe plan cannot be funded through taxes alone. ECB will restart QE. Germany's defense budget already up 24%. France and Italy cannot afford this without printing. 🇯🇵 JAPAN: 237% debt-to-GDP. Every rate hike is catastrophic. BOJ will be forced to SLOW normalization. That is de facto printing. The uncomfortable question nobody is asking: What happens when all four central banks ease AT THE SAME TIME? Here is the logic chain they don't want you to follow. Pentagon gets $200B → Treasury floods bonds → yields spike → Fed buys bonds (QE) → dollar weakens → China cuts rates to stay competitive → yuan falls → Europe has to ease to compete → every currency races to the bottom → who wins? Not you. Not your savings account. Not your 3% money market fund. GOLD already went to $5,595 this year. The market already knows. BITCOIN is at $71,000. Down from $94,000 peak. Historically this is the last buying opportunity before the next wave. In 2020-2021, COVID printing sent Bitcoin from $7,000 to $69,000. This cycle has MORE debt. MORE war. MORE structural pressure. MORE competitive devaluation. If global M2 expands by even half of what it did in COVID, you are looking at Bitcoin targets that sound insane right now. $200K? $300K? Those numbers made people laugh in 2019 too. The conspiracy nobody talks about: governments WANT inflation. It is the only way to erase $39 trillion in real debt value. Your savings account loses 5% a year. Their debt burden shrinks 5% a year. That is the plan. They are not incompetent. This is the exit strategy. Prepare accordingly. 🚨🚨🚨
Whale.Guru tweet media
English
31
142
418
26.4K
Michael D. White retweetledi
Chutes
Chutes@chutes_ai·
RESI Appraisals is now live on Chutes. @resilabsai (SN46) is building institutional-grade property appraisals for US real estate, and they're running inference through Chutes infrastructure. Lowest cost. End-to-end encryption. Increased distribution to agents. Subnet building on subnet. This is the Bittensor flywheel in action.
Chutes tweet media
English
8
39
177
9.8K
Michael D. White retweetledi
based16z
based16z@based16z·
Everyone stay in your homes to keep oil prices down. 2 weeks to open the strait
English
102
979
12.6K
742.6K
Michael D. White retweetledi
Sasha
Sasha@Sashawright_1·
Reading the $TAO whitepaper felt like reading Bitcoin and Ethereum for the first time. That same energy. That same feeling that something real is being built. Not another token. Not another narrative. A genuinely new system. Cypherpunk is the word people use to dismiss it. It’s also the word they used about Bitcoin in 2009.
English
52
40
249
16.3K
Jim Cramer
Jim Cramer@jimcramer·
Very oversold...
English
2.7K
618
6.4K
3M
Investing88
Investing88@Investing88ai·
@CryptoZPunisher @here4impact Details in the coming weeks. @TrustedStake will have a strategy page for the public fund where everyone can join. Wallets can have sn88 alpha pre-loaded, or auto-purchased based on a ratio Essentially it'll be turnkey wallet growth + auto alpha buyback. Win-win for everyone
English
3
2
18
688
Michael D. White retweetledi
James Ross
James Ross@jamesrosst·
I get this daily… Why does @SynthdataCo focus on volatility modelling and not price point predictions? Volatility is a tractable mathematical problem that machine learning models can optimize for, and it is the missing primitive required for AI-native financial intelligence. It has taken us 12 months and we are now starting to see significant demand for Synth’s API from LLMs, agents, and trading systems globally. We expect this to 10-100x in the coming months as AI based trading accelerates and Synth continues to expand intelligence across equities, crypto and commodities.
James Ross tweet media
English
3
7
40
2.8K
Michael D. White retweetledi
Erfan Miahi
Erfan Miahi@erfan_mhi·
Used autoresearch to make @grail_ai GRPO trainer 1.8x faster on a single B200. I kept postponing this for weeks since the bottleneck in our decentralized framework was mainly communication. But after our proposed technique, PULSE, made weight sync 100x faster, the training update itself became the bottleneck. Even with a fully async trainer and inference, a slow trainer kills convergence speed. A task that could've eaten days of my time ran in parallel while I worked on other stuff. Unlike original autoresearch, where each experiment is 5 min, our feedback loop is way longer (10-17 min per epoch + 10-60 minutes of installations and code changes), so I did minimal steering when it was heading in bad directions to avoid burning GPU hours. The agent tried so many things that failed. But, eventually found the wins: Liger kernel, sequence packing, token-budget dynamic batching, and native FA4 via AttentionInterface. 27% to 47% MFU. 16.7 min to 9.2 min per epoch. If you wanna dig deeper or contribute: github.com/tplr-ai/grail We're optimizing everything at the scale of global nodes to make decentralized post-training as fast as centralized ones. Stay tuned for some cool models coming out of this effort. Cheers!
Erfan Miahi tweet media
English
1
13
55
13.5K