deployed an agent to stop me from fading my own thesis. it lost too 🫡

Benin City
Joined March 2011
🔐 Update on the Upcoming Verification Process We know many of you are looking forward to verification - and we’re stepping closer to that moment. This process is designed to be structured, transparent, and fair, ensuring every qualified Human Node moves forward with clarity.
436
Jeremiah retweeted
Bullish, bearish or wait? 📊 M-01 brings confirmed signals and reference trade levels onto your SmartK chart, so you can review the setup before making your own call. Get the walkthrough 👉 bit.ly/3U7bcjl #Toobit #SmartK
40
38
92
16,611
Jeremiah retweeted
That "wait, after fees?" moment 🤔 Find your Breakeven Price under Copy Trading → My Copy Trades → Current. Learn more 👉 bit.ly/46Cxvh2 #Toobit #CopyTrading #BreakevenPrice
40
44
86
17,994
One thing I think people underestimate about AI is how much the infrastructure matters. Everyone talks about smarter models, bigger benchmarks, and new capabilities. But running those models at scale is a completely different challenge. That’s why DeepSeek open-sourcing its DeepInfer inference framework caught my attention. If sharing a production-grade inference stack helps developers improve throughput, reduce latency, and lower compute costs, the impact could go far beyond one model. Cheaper inference means more room for experimentation, more accessible AI products, and fewer barriers for smaller teams. I’m increasingly convinced that making AI cheaper to run is just as important as making it smarter. The next big AI unlock might not be a better model. It might be making powerful models affordable enough for everyone to use.
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source! Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable. 🔗 github.com/deepseek-ai/DeepI… 1/4
63
30
172
13,621
Jeremiah retweeted
🧠 HOW IT WORKS & THE SCIENCE Yesterday was about meeting beginners where they are. Today: what if you’re not a beginner? You already know what staking is. You understand LPs. You don’t need Wingman explaining APR every time it appears. Then the product should change with you. Not by becoming a different product. By letting the depth change. The same portfolio can be approached at different levels. A newer user may ask: “What does this mean?” A more experienced user may ask: “Why did this change?” A native may ask: “What data produced this number, how fresh is it, and what did you exclude?” Same system. Different question. Different depth. That distinction is intentional. 🧠 The science: this is closely related to the expertise reversal effect. Instructional support that helps novices can become redundant—or even interfere with performance—as expertise increases. A 2025 meta-analysis found the effect across varied contexts and reinforced the broader idea that support should adapt to learner expertise. That’s why SparkyFi shouldn’t force everyone through the same amount of explanation. Support should be available— not compulsory. The product should be able to say: “Here’s what this means.” without making the user listen when they already know. And: “Here’s the deeper evidence.” when the user wants to interrogate it. What makes software feel “advanced” to you: more information— or better access to the right information? #SparkyFi #DeFi #DeFiAI #BehavioralIntelligence #AIDashboard
16
14
16
3,417
Jeremiah retweeted
Saw a pretty aggressive call recently. $NBIS 2026 ARR might hit the high end of the guide. Some people are even pushing 2027 toward $30B. Im not calling that number yet. Were optimistic on this company. But you have to know where it sits right now. The next rerate is about how much revenue each 1GW can sell, and how much EBITDA it keeps. If the same batch of GPUs runs at higher utilization, pushes more tokens, and the tight short-term capacity still prices higher, revenue per GW keeps climbing. Q3 Im watching three things. Revenue per GW. AI cloud EBITDA margin. How fast new capacity comes online. If those three improve together, the markets 2027 ARR expectation can keep getting revised up. Dont forget the financing. If the company keeps issuing shares and dilutes too much, the value per share gets eaten. Thats the part of Q3 most worth checking.
Same AI compute build-out, but the points that actually constrain shareholder returns can look very different. For the NBIS position, the more relevant questions right now are how much additional capital is still needed to deliver this capacity, and how many shares will end up sharing the returns.The core AI cloud contracts signed in Q2 imply roughly $20–25 million in annualized revenue per MW, and AI Cloud adjusted EBITDA margin hit 50% in the same quarter. That margin, however, does not yet deduct depreciation, interest, or the ongoing expansion spend. At the end of Q2 the company had about $8 billion in cash; in July it closed $775 million of asset-backed financing at SOFR + 2.5%; customer prepayments are also helping fund the build-out.The “$4.5 billion convertible” that keeps getting mentioned is also worth clarifying. That was the initial size announced on August 19; the deal was subsequently upsized and closed in August at roughly $5.75 billion. Old news can still be re-evaluated for risk by the market, but it shouldn’t be framed as a fresh financing that just happened in October.Honestly, what worries me more is how the remaining funding gap gets filled later. If they raise another $1 billion of equity, a $250 share price means issuing 4 million shares, while $165 means roughly 6.06 million shares — about 52% more dilution for the same amount of money. It’s just a scenario, but it explains why a lower stock price makes the market keep asking about the financing plan.Somy bull case is pretty straightforward: how many MW can they actually deliver, at what revenue per MW, and at what capital cost. Those three variables matter most. Strong demand and solid contract pricing are still the main reasons I’m holding. A few weeks of engineering delay and a customer actually walking away are two completely different things.Inmy view, this is currently one of the higher-certainty, higher-quality AI compute names you can buy. The pullback also makes the trading setup more interesting. Curious how others are looking at #NBIS around these $170 levels👀🤐
1
3
23
2,009
Jeremiah retweeted
$FRTT one of my top watches into open...someone has been soaking her up all morning...lets see how it opens
4
2
1
17
1,795
Jeremiah retweeted
GOOGLE LABS JUST OPENED THE DOOR TO AN ENTIRE GENERATION OF AI GAME DEVELOPERS. AND A TEAM IS ALREADY BUILDING THE ECONOMY AROUND IT. it's called PlaygroundPad. I found it to be a really unique platform turning AI-generated games into actual playable products with onchain economies. no coding. no game engines. no large development teams. → describe an idea and build a playable game → compete in tournaments for real rewards → earn creator fees from your creations → build, grow and fund games with your community think about how big this could get. millions of people have game ideas but never had the skills, money or team to build them. AI is removing the development barrier. Web3 is introducing a way to fund the games and reward the people creating them.
🤝 Paid partnership
PLAYVERSE IS LIVE 🌃 Playgroundpad just became a world. What is the Playverse? It’s the multiplayer hub of Playgroundpad. One neon-soaked street at night, the Neon District, where everyone on the pad shows up at the same time. You don’t scroll a list of games anymore. You walk into them. Everyone you see is real. The people walking around you are real players, live, right now. The “here now” counter only counts real people, never the background crowd. Up to 40 players share a room, and when one fills up, you’re sent to the next. Your PlayTag is your name. Sign in with your PlayTag and your name floats above your head. Chat pops up as speech bubbles over players. Tap anyone to open their PlayTag card and jump to their profile. Don’t like someone? Block or report them in one tap. Every shop is a real game. The arcades on the strip aren’t decoration. One side of the street holds the game styles. The other side holds the newest games players made on the pad. Walk up to a shop, press play, and you’re in. Grab ramen on the way. The gate at the end is yours. Walk to the end of the street and you reach the gate where you make your own game. Type one sentence, get a game, launch it as a coin. Then it shows up on the street for everyone else to walk into. Pull up 👇 playgroundpad.fun/playverse
75
36
2
1,149
726,633
Jeremiah retweeted
Launch a token. Share the upside. Creator fee-sharing is live on @o1exchange: → Split trading fees with any X accounts at launch → Set exactly what % each one gets → Recipients connect X and claim straight from their profile Bring your community, collabs, and KOLs in on day one.
Market is quiet recently but we @o1_exchange team have been cooking hard. More launch mechanisms, better UIUX, smoother product experience. All prepped for that $1B B20 runner on @base.
28
80
1
588
68,577
Jeremiah retweeted
Added 30 min ago, up 120%. Flow8 🌊 💰
2
4
65
4,418
Full video of our discussion with @jessepollak @tarunchitra from @token2049 on Why AI Needs Crypto. Blockchains can provide a much-needed coordination layer to take AI from today's single-player mode to multiplayer mode, unlocking security & markets. youtube.com/watch?v=1flFvL62…
27
25
1
226
16,189
🚨 BREAKING: Qualcomm Announces Global Job Cuts Qualcomm will cut jobs worldwide following its annual employee reviews, with layoffs expected to begin in November. CEO Cristiano Amon says the move aims to build a stronger performance-focused culture.
57
33
166
8,929
Wish I had a PumpFun deal to claim my free Rolex.
18
5
219
13,105
$10 Giveaway 🏆 To enter: 1️⃣ Follow: @ThaNDog2026 2️⃣ Like & Repost 🔁 3️⃣ Drop your SOL address below
6
24
137
17,308