# ![@AgenticDataGuy Avatar](https://lunarcrush.com/gi/w:26/cr:twitter::1927365977043288064.png) @AgenticDataGuy Data Dave

Data Dave posts on X about bitcoin, ai, inference, agents the most. They currently have [--] followers and [--] posts still getting attention that total [---] engagements in the last [--] hours.

### Engagements: [---] [#](/creator/twitter::1927365977043288064/interactions)
![Engagements Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1927365977043288064/c:line/m:interactions.svg)

- [--] Week [--] -91%

### Mentions: [--] [#](/creator/twitter::1927365977043288064/posts_active)
![Mentions Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1927365977043288064/c:line/m:posts_active.svg)


### Followers: [--] [#](/creator/twitter::1927365977043288064/followers)
![Followers Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1927365977043288064/c:line/m:followers.svg)

- [--] Week [--] -9.50%

### CreatorRank: undefined [#](/creator/twitter::1927365977043288064/influencer_rank)
![CreatorRank Line Chart](https://lunarcrush.com/gi/w:600/cr:twitter::1927365977043288064/c:line/m:influencer_rank.svg)

### Social Influence

**Social category influence**
[finance](/list/finance)  [technology brands](/list/technology-brands)  [cryptocurrencies](/list/cryptocurrencies)  [stocks](/list/stocks)  [countries](/list/countries)  [social networks](/list/social-networks)  [ncaa football](/list/ncaa-football)  [travel destinations](/list/travel-destinations)  [exchanges](/list/exchanges) 

**Social topic influence**
[bitcoin](/topic/bitcoin), [ai](/topic/ai), [inference](/topic/inference), [agents](/topic/agents), [cloud](/topic/cloud), [market](/topic/market), [in the](/topic/in-the), [nfl](/topic/nfl), [wild](/topic/wild), [finance](/topic/finance)

**Top assets mentioned**
[Bitcoin (BTC)](/topic/bitcoin) [DoorDash Inc. (DASH)](/topic/doordash) [Uber Technologies, Inc. (UBER)](/topic/$uber) [Microsoft Corp. (MSFT)](/topic/microsoft)
### Top Social Posts
Top posts by engagements in the last [--] hours

"@KobeissiLetter Europe is basically short europe at this point. their own pension funds are making a massive bet that US companies will outperform the continent they're supposed to be funding retirements for"  
[X Link](https://x.com/AgenticDataGuy/status/2015949299834618135)  2026-01-27T00:45Z [--] followers, [----] engagements


"I analyzed students across [---] countries on their social media habits. What I found: There's a usage threshold where 98% report grade damage TikTok users sleep [--] minutes less than Instagram users WhatsApp is the most used platform at [----] hrs/day "Complicated" relationship status correlates with the least sleep Here's the full breakdown 🧵 https://twitter.com/i/web/status/2015901112247714185 https://twitter.com/i/web/status/2015901112247714185"  
[X Link](https://x.com/DataIsCool5/status/2015901112247714185)  2026-01-26T21:34Z [--] followers, [--] engagements


"THE 5-HOUR CLIFF This visualization shows something striking. The vertical yellow line marks [--] hours of daily usage. Look at what happens: Left of the line: Students are split between "Yes" and "No" on academic impact Right of the line: Almost everyone clusters at "Yes" The color gradient tells the story - as usage increases (blue red) students shift dramatically upward into the "affects academics" zone. Below [--] hours: 68% report academic impact Above [--] hours: 98% report academic impact The average student uses [---] hours daily. Most are right at the edge of this cliff"  
[X Link](https://x.com/DataIsCool5/status/2015901113883754867)  2026-01-26T21:34Z [--] followers, [--] engagements


"PLATFORM BREAKDOWN: The data across [--] platforms DAILY USAGE (Image 1): WhatsApp leads at [----] hrs/day - makes some sense for a "messaging" app. TikTok follows at [----] hrs. LinkedIn users average just [----] hrs. SLEEP (Image 2): Snapchat users sleep the least at [----] hrs/night. WhatsApp users get [----] hrs. Meanwhile LINE users average [----] hrs of sleep. MENTAL HEALTH (Image 3): Clear pattern - platforms with lower usage correlate with higher mental health scores. LinkedIn and LINE users score highest. ADDICTION SCORE (Image 4): WhatsApp Snapchat and TikTok cluster at the top (7.5/10). LinkedIn"  
[X Link](https://x.com/DataIsCool5/status/2015901117515784403)  2026-01-26T21:34Z [--] followers, [--] engagements


"THE SLEEP-SCROLL CONNECTION This scatter plot reveals one of the strongest patterns in the data. Each dot is a student. The color shows mental health score (yellow = good purple/black = poor). Notice: Top-left cluster (yellow/orange): Low usage high sleep good mental health Bottom-right cluster (purple/black): High usage low sleep poor mental health The trend line shows a clear negative correlation. Students using [--] hours/day sleep [--] hours Students using 7+ hours/day sleep under [--] hours And the color gradient makes it undeniable - mental health deteriorates as you move down and right. Three"  
[X Link](https://x.com/DataIsCool5/status/2015901120485527809)  2026-01-26T21:34Z [--] followers, [--] engagements


"10 out of [--]. The top [--] most efficient rushers in the NFL are all quarterbacks. Zero are running backs. Let me explain why this matters 🧵"  
[X Link](https://x.com/DataIsCool5/status/2015915244275609685)  2026-01-26T22:30Z [--] followers, [--] engagements


"Here's the full list. Mahomes Allen Dart Lamar Herbert Maye Lawrence Hurts Nix Caleb Williams. Every single one generates more value per carry than any RB in the league. EPA (Expected Points Added) measures how much each rush helps your team score. Higher = better"  
[X Link](https://x.com/DataIsCool5/status/2015915246699847818)  2026-01-26T22:30Z [--] followers, [--] engagements


"Qwen has been hyping their scaling progress: 10T parameters and trained on 100T tokens. Is this model already in training right now or are they talking about a 1T-parameter version instead It isn't available on their API yet (at least nothing documented so far). Performance-wise it doesn't clearly beat DeepSeek V3.2 which sits at 685B parameters and gets served by various providers for roughly $0.28 per 1M input tokens and $0.45 per 1M output tokens. I doubt Qwen will match those API prices anytime soon - theyre probably still relying on GQA like they did in their Qwen [--] MoE models. Still"  
[X Link](https://x.com/DataIsCool5/status/2015945847217774818)  2026-01-27T00:31Z [--] followers, [---] engagements


"Qwen's bragging about massive scaling: 10T parameters 100T tokens. Is the 10T model actually training now or is this still the 1T version It's not on their API yet (no docs at least). It doesn't really beat DeepSeek V3.2 (685B params) which vendors serve at $0.28/1M in and $0.45/1M out. Qwen probably won't match those prices soon - ikely still using GQA like their Qwen [--] MoEs. Still being roughly on par with DeepSeek across benchmarks is solid. Way better than them quitting LLMs entirely. https://twitter.com/i/web/status/2015946102751543537 https://twitter.com/i/web/status/2015946102751543537"  
[X Link](https://x.com/DataIsCool5/status/2015946102751543537)  2026-01-27T00:33Z [--] followers, [---] engagements


"@42Cyc Not to be that guy but shedeur's passer rating this year would've been dead last in [--] of the last [--] seasons. The one exception was [----] when there were multiple injured starters cycling through. Trevor's would've been top [--] in [--] of those years"  
[X Link](https://x.com/DataIsCool5/status/2015951043536777625)  2026-01-27T00:52Z [--] followers, [----] engagements


"There's a stat about Afghanistan that has stuck with me since I first saw it. Brick kilns. The country has thousands of them and a huge percentage operate on debt bondage. A family takes a loan they can't repay and suddenly they're working 14-hour days making bricks with their kids alongside them. Some families have been trapped for generations. The debt gets passed down. ILO estimates suggest something like [---] million Afghan children are in child labor with a significant portion in hazardous conditions. That was before the economic collapse of the last few years which made everything worse."  
[X Link](https://x.com/DataIsCool5/status/2015957942344794212)  2026-01-27T01:20Z [--] followers, [---] engagements


"@unusual_whales @josephwang @jam_croissant @super_macro @MPelletierCIO @LastBearStandng @grok what are the odds on a rate cut this week"  
[X Link](https://x.com/DataIsCool5/status/2015958252576522501)  2026-01-27T01:21Z [--] followers, [--] engagements


"The $550 million figure in that DCWP report is pretty striking but the comparison group is what makes it interesting. Grubhub kept tipping at checkout and their per-delivery tips stayed around $2.17. DoorDash and Uber moved it to after checkout and theirs dropped to [--] cents. Same city same customers same time period. The regulation is basically forcing the second group to match the first. https://twitter.com/i/web/status/2015964192717578685 https://twitter.com/i/web/status/2015964192717578685"  
[X Link](https://x.com/DataIsCool5/status/2015964192717578685)  2026-01-27T01:44Z [--] followers, [--] engagements


"Worth noting the gold/silver ratio context here. Even at $110 silver is still historically cheap relative to gold. The long-term average ratio is around 50-60:1. During the [----] peak it got down to 32:1. Lots of room to run if that mean reversion thesis plays out and these swings are just noise on that longer timeframe. https://twitter.com/i/web/status/2015980471214117260 https://twitter.com/i/web/status/2015980471214117260"  
[X Link](https://x.com/DataIsCool5/status/2015980471214117260)  2026-01-27T02:49Z [--] followers, [---] engagements


"16040 square miles. That's the entire Netherlands. West Virginia is 50% larger. The United States has [---] times its landmass. Yet this postage stamp of a country exports $140 billion in agricultural products annually. Second largest food exporter on the planet. Behind only America. The Dutch grow a pound of tomatoes with half a gallon of water. The global average is [--] gallons. Not rainfall. Actual tap water pumped from aquifers and reservoirs. Chinese tomato farmers use [--] gallons per pound. The Netherlands uses [--] times less. [-----] acres of crops grow under glass. Twice the size of"  
[X Link](https://x.com/DataIsCool5/status/2016137616337355187)  2026-01-27T13:14Z [--] followers, [--] engagements


"@WallStreetMav The timing is wild. This is Indias 4th major trade deal since last May. UK Oman New Zealand and now EU. Theyre speedrunning trade diversification while sitting under 50% US tariffs"  
[X Link](https://x.com/DataIsCool5/status/2016146508526002505)  2026-01-27T13:49Z [--] followers, [---] engagements


"@AdamSchefter Joe Bradys offenses have averaged [----] PPG (2019 LSU) [----] PPG (2024 Bills) and [----] PPG (2025 Bills). Every time he calls plays for a full season the offense finishes top-5 nationally or in the NFL. The pattern is pretty hard to ignore"  
[X Link](https://x.com/DataIsCool5/status/2016191467224469720)  2026-01-27T16:47Z [--] followers, 10.3K engagements


"@espn @AdamSchefter Brady got fired by the Panthers after going 5-11 and 5-7 then quietly rebuilt his career as a QB coach in Buffalo. Three years later hes an MVP quarterbacks head coach. The NFL coordinator carousel is genuinely wild"  
[X Link](https://x.com/DataIsCool5/status/2016191816421269590)  2026-01-27T16:49Z [--] followers, [----] engagements


"@KobeissiLetter We got a preview of what partial unwinding looks like in August [----]. Nikkei dropped 12.4% in a single day VIX spiked above [--] worst Japanese trading day since [----]. And that was just a warning shot from a surprise [--] basis point hike"  
[X Link](https://x.com/DataIsCool5/status/2016194166212948148)  2026-01-27T16:58Z [--] followers, [--] engagements


"@KobeissiLetter Wild that this wipes out more value than Nvidia lost in the September [----] crash ($279B) but gets a fraction of the attention. Healthcare just isnt as sexy as AI"  
[X Link](https://x.com/DataIsCool5/status/2016214668138750052)  2026-01-27T18:20Z [--] followers, [---] engagements


"@alifarhat79 If the Fed actually gets a more dovish chair in May how much further can this go Are we looking at a sustained multi-year decline or a Q2 bottom"  
[X Link](https://x.com/DataIsCool5/status/2016232148764610987)  2026-01-27T19:29Z [--] followers, [---] engagements


"@stats_feed Mondays dollar trading volume through DTCC was the second highest on record. Surpassed only by April [--] [----]. Thats how aggressive this selloff has been"  
[X Link](https://x.com/DataIsCool5/status/2016243578561839286)  2026-01-27T20:15Z [--] followers, [---] engagements


"@thecalball2 @grok Please do @grok"  
[X Link](https://x.com/DataIsCool5/status/2017044016911536559)  2026-01-30T01:15Z [--] followers, [--] engagements


"@bryan_johnson I heard it does better if you give it complete and total access to all your files and accounts. Especially your credit card information"  
[X Link](https://x.com/AgenticDataGuy/status/2016213834147496005)  2026-01-27T18:16Z [--] followers, [--] engagements


"The timing of when empires break from hard money is weirdly consistent. Rome did it when military expansion costs outpaced loot from new conquests. Britain did it to finance WWI. US did it when Vietnam + Great Society spending collided with gold outflows. The trigger is almost always the same: expenses that cant be covered by honest taxation. https://twitter.com/i/web/status/2016242565050249359 https://twitter.com/i/web/status/2016242565050249359"  
[X Link](https://x.com/AgenticDataGuy/status/2016242565050249359)  2026-01-27T20:11Z [--] followers, [---] engagements


"@Alibaba_Qwen Funny how "80B total 3B active" would've sounded like a gimmick two years ago. Now it's the whole point. Local inference is the bottleneck not raw capability. The model that wins isn't the smartest one. It's the smartest one that actually runs on your hardware"  
[X Link](https://x.com/AgenticDataGuy/status/2018734930407072245)  2026-02-03T17:14Z [--] followers, [----] engagements


"Local coding agents have been "6 months away" for two years. Qwen3-Coder-Next might actually deliver. 80B params total. 3B active at inference. Trained on 800K verifiable coding tasks. Here's what this means for your local setup: 🧵"  
[X Link](https://x.com/anyuser/status/2018736695454126198)  2026-02-03T17:21Z [--] followers, [--] engagements


"Who should actually test this: Devs working on proprietary code who can't use cloud APIs. Teams in regulated industries (finance healthcare defense). Anyone tired of paying per-token for coding assistance. People who just want their tools to work offline. The open-weight ecosystem just got a lot more practical"  
[X Link](https://x.com/anyuser/status/2018736701393215795)  2026-02-03T17:21Z [--] followers, [--] engagements


"@Shekhar_Builds Im excited to test it too Qwen seems to be working hard at optimizing for local inference"  
[X Link](https://x.com/AgenticDataGuy/status/2018737576538997119)  2026-02-03T17:25Z [--] followers, [--] engagements


"$800 billion in Bitcoin market value. Gone in four months. In October people were calling for $200000. Today it briefly traded at $69800 and the Fear & Greed Index hit [--]. Lower than during the FTX collapse. This is the story of how the market broke: 🧵"  
[X Link](https://x.com/anyuser/status/2019423127735005366)  2026-02-05T14:49Z [--] followers, [--] engagements


"It started with leverage. Billions in borrowed money betting Bitcoin would keep climbing. When it didn't exchanges auto-sold those positions. That selling pushed prices lower. Which liquidated more people. CoinGlass data: $5.42 billion in positions liquidated since January [--]. Today alone $775 million evaporated when BTC broke $70K. Then Strategy (formerly MicroStrategy) crossed below their $76037 average buy price. Panic spread that they'd be forced to sell their massive BTC stack. The fear made everything worse"  
[X Link](https://x.com/AgenticDataGuy/status/2019423131631247596)  2026-02-05T14:49Z [--] followers, [--] engagements


"Then the real economy showed up. The Fed paused rate cuts January [--]. Trump nominated Kevin Warsh a known hawk as the next Fed Chair. Markets heard "higher rates for longer" and dumped everything risky. Tech earnings were brutal. Microsoft fell 10%. AMD dropped 14%. PayPal crashed 19% after earnings (CNBC). The Nasdaq sank 2.2% in a single session. Bitcoin trades like a tech stock now. Per CoinDesk it's down 44% from its October peak while gold (the "safe haven" BTC was supposed to replace) is up 23% year-to-date. https://twitter.com/i/web/status/2019423133996925416"  
[X Link](https://x.com/AgenticDataGuy/status/2019423133996925416)  2026-02-05T14:49Z [--] followers, [--] engagements


"Claude Opus [---] just dropped and the benchmarks are wild. Beats GPT-5.2 by [---] Elo points on real-world knowledge work. Beats its own predecessor by [---]. Here's what actually matters: 🧵"  
[X Link](https://x.com/anyuser/status/2019479569795817694)  2026-02-05T18:33Z [--] followers, [--] engagements


"First Opus model with a 1M token context window. That's not a gimmick. That means entire codebases full regulatory filings months of internal docs all processed in a single pass. Output bumped to 128K tokens too. This thing can write a book in one shot"  
[X Link](https://x.com/AgenticDataGuy/status/2019479571788132538)  2026-02-05T18:33Z [--] followers, [--] engagements


"The coding results are absurd. Highest score ever on Terminal-Bench [---] (agentic coding eval). Top score on Humanity's Last Exam (multidisciplinary reasoning). Top score on BrowseComp (finding hard-to-locate info online). It's leading on basically everything that matters"  
[X Link](https://x.com/AgenticDataGuy/status/2019479573746868720)  2026-02-05T18:33Z [--] followers, [--] engagements


"The real shift though Agent teams. Instead of one AI grinding through tasks sequentially Opus [---] splits work across multiple agents coordinating in parallel. NBIM tested this with [--] cybersecurity investigations. Opus [---] won [--] of [--] in blind ranking against [---] models. Each run used [--] subagents and 100+ tool calls. Read more here: https://code.claude.com/docs/en/agent-teams https://code.claude.com/docs/en/agent-teams"  
[X Link](https://x.com/anyuser/status/2019479577865658419)  2026-02-05T18:33Z [--] followers, [--] engagements


"Pricing stayed the same. $5/$25 per million input/output tokens. Better model same cost. That almost never happens. Its Available now on the API and all major cloud platforms. Which of these capabilities are you most excited to test http://claude.ai http://claude.ai"  
[X Link](https://x.com/anyuser/status/2019479579346248092)  2026-02-05T18:33Z [--] followers, [--] engagements


"Claude Opus [---] just dropped and the benchmarks are wild. Beats GPT-5.2 by [---] Elo points on real-world knowledge work. Beats its own predecessor by [---]. Here's what actually matters: 🧵"  
[X Link](https://x.com/anyuser/status/2019479569795817694)  2026-02-05T18:33Z [--] followers, [--] engagements


"The real shift though Agent teams. Instead of one AI grinding through tasks sequentially Opus [---] splits work across multiple agents coordinating in parallel. NBIM tested this with [--] cybersecurity investigations. Opus [---] won [--] of [--] in blind ranking against [---] models. Each run used [--] subagents and 100+ tool calls. Read more here: https://code.claude.com/docs/en/agent-teams https://code.claude.com/docs/en/agent-teams"  
[X Link](https://x.com/anyuser/status/2019479577865658419)  2026-02-05T18:33Z [--] followers, [--] engagements


"Pricing stayed the same. $5/$25 per million input/output tokens. Better model same cost. That almost never happens. Its Available now on the API and all major cloud platforms. Which of these capabilities are you most excited to test http://claude.ai http://claude.ai"  
[X Link](https://x.com/anyuser/status/2019479579346248092)  2026-02-05T18:33Z [--] followers, [--] engagements


"$800 billion in Bitcoin market value. Gone in four months. In October people were calling for $200000. Today it briefly traded at $69800 and the Fear & Greed Index hit [--]. Lower than during the FTX collapse. This is the story of how the market broke: 🧵"  
[X Link](https://x.com/anyuser/status/2019423127735005366)  2026-02-05T14:49Z [--] followers, [--] engagements


"The ETF engine that fueled the rally It's running in reverse. Spot Bitcoin ETFs lost $6.18 billion from November through January. On one single day BlackRock's IBIT saw $528.3 million walk out the door. Those ETF redemptions are dumping about [-----] BTC per month back into the market. For context miners only produce about [-----] BTC per month. ETFs are selling twice what miners create. The average ETF holder bought in around $87830 per Galaxy Digital. They're sitting on 20%+ losses. That breeds more selling. Take a look at those red days stacking up recently (outflow)"  
[X Link](https://x.com/anyuser/status/2019423136551157919)  2026-02-05T14:49Z [--] followers, [--] engagements


"Here's what the panicking crowd forgets: Bitcoin has crashed 40%+ from highs five times in its history. Every single time it eventually set new records. The weekly RSI is approaching oversold. The last time that happened was June [----]. BTC was at $17500. It ran to $126000 from there. And while retail is selling the options market tells a different story. $1.159 billion in put open interest at $75K now nearly matches $1.168 billion in calls at $100K. Smart money is hedging not fleeing. Nobody knows the bottom. But fear this extreme has historically been temporary. Are you buying holding or"  
[X Link](https://x.com/anyuser/status/2019423144398725214)  2026-02-05T14:49Z [--] followers, [--] engagements


"How Im waiting for the sonnet [--] release today"  
[X Link](https://x.com/anyuser/status/2018740789547806864)  2026-02-03T17:38Z [--] followers, [---] engagements


"@grok what time does anthropic usually launch new models"  
[X Link](https://x.com/anyuser/status/2018741429464367511)  2026-02-03T17:40Z [--] followers, [--] engagements


"Local coding agents have been "6 months away" for two years. Qwen3-Coder-Next might actually deliver. 80B params total. 3B active at inference. Trained on 800K verifiable coding tasks. Here's what this means for your local setup: 🧵"  
[X Link](https://x.com/anyuser/status/2018736695454126198)  2026-02-03T17:21Z [--] followers, [--] engagements


"The benchmark that matters for local use: Aider. Qwen3-Coder-Next scores 69.9%. That's your actual "can it help me code in my terminal" test. It beats GLM-4.7 and MiniMax M2.1 while using a fraction of the active compute. This is the efficiency curve we've been waiting for"  
[X Link](https://x.com/anyuser/status/2018736699912650893)  2026-02-03T17:21Z [--] followers, [--] engagements


"Who should actually test this: Devs working on proprietary code who can't use cloud APIs. Teams in regulated industries (finance healthcare defense). Anyone tired of paying per-token for coding assistance. People who just want their tools to work offline. The open-weight ecosystem just got a lot more practical"  
[X Link](https://x.com/anyuser/status/2018736701393215795)  2026-02-03T17:21Z [--] followers, [--] engagements


"Something weird is happening in AI. The models are getting smaller. And better. Qwen3's MoE architecture: 235B total parameters but only 22B activate per token. You get frontier-class performance at a fraction of the cost. Their Qwen3-4B model It matches models 18x its size on instruction-following tasks. DeepSeek trained V3 for roughly $5.6 million. GPT-4 cost an estimated $50-100 million. Both compete at the frontier. The Stanford AI Index found inference costs dropped 280x in two years. Open-weight models closed the performance gap from 8% to 1.7% in a single year. What's driving this"  
[X Link](https://x.com/anyuser/status/2018733291147296802)  2026-02-03T17:08Z [--] followers, [--] engagements


"Hold your horses folks. Sonnet [--] is coming🙏"  
[X Link](https://x.com/anyuser/status/2018717943132049459)  2026-02-03T16:07Z [--] followers, [--] engagements

Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing