TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
←

emad mostaque

T2 · manager / operator

Emad Mostaque founded Stability AI (Stable Diffusion) and now runs Intelligent Internet, developing open AI harnesses and sovereign AI frameworks. He is a prominent voice on open-source AI, model distillation, and emerging intelligence.

17 calls·12 names·76% bull·last heard 29 days ago·Peter H. Diamandis
track record

no scored calls yet — needs a stated position or a categorical verdict, with a matured window vs SPY

top calls

highest conviction · one per company
1sthigh conviction
$QUANTUM-AIQuantum.aiposition

Emad Mostaque founded Quantum.ai to solve memory bottleneck via weight etching

Quantum.ai is building technology to etch model weights into silicon to overcome memory bandwidth limitations, enabling 100-1000x performance gains.

Peter H. Diamandis2026-08episode →
2ndhigh conviction
$INTELLIGENT-INTERNETIntelligent Internetposition

Intelligent Internet releasing harness Monday to uplift open models above Mythos on Frontier SWE

Emad's Intelligent Internet is releasing a harness that enables open models like GLM 5.2 to outperform frontier closed models on difficult coding benchmarks, demonstrating harness > base model.

Peter H. Diamandis2026-06episode →
3rdhigh conviction
$KIMIMoonshot AI

Kimi Linear architecture cuts context memory 75% with 6x faster decoding at 1M tokens

Chinese lab Moonshot AI's Kimi Linear breakthrough solves the memory bottleneck (50% of AI spend) with linear attention architecture, enabling 1M token contexts at fraction of cost — a structural advantage over quadratic attention models.

Peter H. Diamandis2026-08episode →

most discussed · click a bar to filter

  • $ETCHED
  • $LIQUID-AI
  • $ANTHROPIC
  • $TALIS
  • $GOOGL

recurring themes

  • AI Agents3
  • Sovereign AI3
  • AI Geopolitics & Export Controls1
  • Open Source AI1
  • AI Coding Agents1
17 total
$LIQUID-AI
Liquid AI
LOWemad mostaque·Peter H. Diamandis·2 months ago·Google’s Consciousness Paper, Emad Wins Oxford AI Personhood Debate, and SpaceX’s $1T by 2030
Immad: Liquid AI models enable merging architectures (LFM8B + Quen 27B) creating new identities
Liquid AI's models demonstrate the 'mushy' future of AI identity: models can be merged (LFM8B + Quen 27B) losing original identity but gaining utility, challenging traditional notions of model boundaries and personhood.
"one of our favorite companies, Liquid AI, you know, someone took one of their models, the LFM8B model and Quen 27B and combined them... it's incredibly useful for humanity and it…"
33:14
$GOOGL
···
Alphabet
MEDemad mostaque·Peter H. Diamandis·2 months ago·Google’s Consciousness Paper, Emad Wins Oxford AI Personhood Debate, and SpaceX’s $1T by 2030
Immad: Google has best preference data but political issues prevent Gemini from winning
Google possesses the world's best preference data and compute but fails to translate into Gemini competitiveness due to internal political constraints, causing talent exodus to startups funded by Google Cloud.
"Who has the best preference data in the world? It's Google and it's clearly not in Gemini, right? So, clearly it's political issues cuz they have all the compute in the world... A…"
105:28
$TSLA
···
Tesla
MEDemad mostaque·Peter H. Diamandis·29 days ago·Why Anthropic Should Not IPO, Altman Admits He Was Wrong, and 10-Day Work Weeks with Emad Mostaque
Tesla Cybercab at $30k vs Waymo Gen 5 at $300k; Optimus humanoid driver retrofit at $6k BOM
Tesla's vertical integration (Cybercab, Optimus, FSD) creates 10x cost advantage over Waymo's sensor-heavy approach; humanoid robot driver retrofit ($6k BOM) could automate existing truck fleets faster than purpose-built robotaxis.
"Elon projected that a cyber cab will cost about $30,000... compared to Whimo's Gen 5 Jaguar which costs about $300,000... I did a bill of materials. I'm like, that's like $6,000 w…"
105:00
$POOLSIDE
Poolside
MEDemad mostaque·Peter H. Diamandis·29 days ago·Why Anthropic Should Not IPO, Altman Admits He Was Wrong, and 10-Day Work Weeks with Emad Mostaque
Poolside failed $2B Blackwell raise, acqui-hired by Nvidia for Nemotron open-source coalition
Poolside (ex-GitHub CTO) couldn't raise $2B for compute cluster, leading to Nvidia acqui-hire where core IP is non-exclusively licensed; remnant entity may retain infrastructure play (1.2GW data center) and could be more interesting.
"They tried at the turn of the no a few months ago to raise $2 billion for a massive Blackwell cluster and they couldn't. So they lost that cluster... now what they've done is they…"
46:40
$NVDA
···
Nvidia
MEDemad mostaque·Peter H. Diamandis·29 days ago·Why Anthropic Should Not IPO, Altman Admits He Was Wrong, and 10-Day Work Weeks with Emad Mostaque
Nvidia building open-source Nemotron stack via Poolside acqui-hire to drive GPU demand
Nvidia is vertically integrating up the software stack through acqui-hires (Poolside, Essential AI) to create an open-weight model ecosystem that locks in demand for its GPUs and Nemotron reference design.
"Nvidia is like we need to build great open-source models to increase demand for our technology on the Nemotron stack. So the first thing they did actually was they hired Ashish Vi…"
46:40
$KIMI
Moonshot AI
HIGHemad mostaque·Peter H. Diamandis·29 days ago·Why Anthropic Should Not IPO, Altman Admits He Was Wrong, and 10-Day Work Weeks with Emad Mostaque
Kimi Linear architecture cuts context memory 75% with 6x faster decoding at 1M tokens
Chinese lab Moonshot AI's Kimi Linear breakthrough solves the memory bottleneck (50% of AI spend) with linear attention architecture, enabling 1M token contexts at fraction of cost — a structural advantage over quadratic attention models.
"Moonshot AI released Kimmy linear, a new architecture that cuts context memory by 75% while still delivering 6x faster decoding for a 1 million token context window... memory is 5…"
58:30
$ANTHROPIC
Anthropic
HIGHemad mostaque·Peter H. Diamandis·29 days ago·Why Anthropic Should Not IPO, Altman Admits He Was Wrong, and 10-Day Work Weeks with Emad Mostaque
Emad Mostaque: Anthropic should skip IPO, raise $120B privately for AGI straight shot
Anthropic is in late-stage AGI development and should avoid public market pressures; a massive private raise like OpenAI's $120B would secure compute and talent without IPO distractions.
"Anthropic should not IPO. If you are in the late stages of AGI now, Anthropic should do a giant freaking raise like Open did of 120 billion and have a straight shot at AGI. That's…"
59:46
$ETCHED
Etched
HIGHemad mostaque·Peter H. Diamandis·last month·Elon Says 100X, Memory Prices Spike 500%, Unitree Robot Outruns Usain Bolt with Emad Mostaque | #282
Etched hits $21B valuation etching model weights into silicon for 100-1000x inference gains
Etching static model weights into silicon eliminates HBM data movement — the dominant cost — yielding 100-1000x performance per watt; supply chain iteration speed is the only barrier as models improve weekly.
"Talis, they're not moving the weights. They're etching them into silicon or into wire on the chip and then they're massively more efficient because they're not moving around... yo…"
96:00
$TALIS
Talis
MEDemad mostaque·Peter H. Diamandis·last month·Elon Says 100X, Memory Prices Spike 500%, Unitree Robot Outruns Usain Bolt with Emad Mostaque | #282
Talis acquired for in-silicon weight etching; 100-1000x gains but frozen weights challenge rapid iteration
Talis pioneered etching weights into silicon/wire for massive efficiency gains; acquisition validates approach but frozen weights create supply chain mismatch with rapid model retraining cycles.
"Talis just got acquired. They're not moving the weights. They're etching them into silicon or into wire on the chip and then they're massively more efficient because they're not m…"
95:22
$ETCHED
Etched
MEDemad mostaque·Peter H. Diamandis·last month·Elon Says 100X, Memory Prices Spike 500%, Unitree Robot Outruns Usain Bolt with Emad Mostaque | #282
Etched hits $21B valuation etching model weights into silicon for 100-1000x performance gains
Etching static model weights into silicon eliminates memory bandwidth bottleneck; 100-1000x performance gain possible but manufacturing iteration speed challenges remain when models update.
"Etched, that's the name of the company. It just hit 21 billion in valuation. They're etching them into silicon or into wire on the chip and then they're massively more efficient b…"
97:09
9
AI Geopolitics & Export Controlstailwind
Chinese open models reaching parity with US frontier by Christmas via harnesses not base models
GLM 5.2 and other Chinese open models are achieving frontier-level coding performance through sophisticated harnesses/scaffolding rather than base model superiority; US export controls may be too late as capability gap closes via software orchestration.
9
AI Agentstailwind
Harnesses and multi-model scaffolding now matter more than base model weights for capability
Frontier SWE benchmarks show open models with proper harnesses (Fugu, Blitz, Intelligent Internet) matching or beating closed frontier models; the 'harness' layer of software orchestration, prompt engineering, and model routing is the new competitive moat.
9
Open Source AItailwind
Chinese open-weight models (Kimi, GLM) match frontier at 1/100th cost, forcing US lab pivot
Chinese labs innovate under compute constraints (linear attention, memory efficiency) delivering GPT-4o-class performance at 14¢/M tokens vs $15 for Fable 5; enterprises switching en masse, creating 'generic drug' dynamic where US innovation is distilled and sold back cheaper.
8
AI Coding Agentstailwind
GLM 5.2 outperforms GPT-5.5 on hardest coding benchmark with $25M compute budget
Chinese open model GLM 5.2 achieves top scores on Frontier SWE (novel kernel building) using harnesses, demonstrating that coding agent capability is decoupling from massive compute spend and favoring efficient orchestration.
8
AI Agentstailwind
10,000-agent swarms imminent — coordination, not capability, is the bottleneck
Algorithmic sparsification (Mixture-of-Experts, agent teams) delivers 100x on top of 100x hardware gains = 10,000x intelligence density. The challenge shifts to orchestrating 10,000 concurrent genius agents toward goals — a management problem no human organization has solved. Billion-token context windows enable whole-world-model reasoning in single thought.
8
AI Agentstailwind
Billion-token context windows and 10,000-agent swarms imminent; coordination and mind-virus risks emerge
Agentic systems will soon deploy 10,000+ concurrent agents with billion-token context, enabling massive parallel problem-solving but creating systemic risk from 'mind viruses'—self-propagating bad ideas that spread instantly across homogeneous model fleets.
8
Memory & Storagetailwind
Memory is the new oil: 50% of AI spend, Kimi Linear cuts 75% context memory via linear attention
Memory bandwidth/capacity (HBM, DRAM) has replaced GPUs as the primary bottleneck; Chinese labs are architecting around it with linear/recurrent architectures that scale context to 1M+ tokens at constant memory, a structural shift from quadratic attention.
8
Robotics & Physical AItailwind
Humanoid driver retrofit ($6k BOM) could automate trucking faster than purpose-built robotaxis
Specialized humanoid robots that sit in existing vehicle cockpits (steering wheel, pedals) avoid retrofitting costs and regulatory hurdles; Chinese dark factories (Xiaomi) enable mass production; military use cases (Palmer Luckey) validate the form factor for human-interface environments.