Welcome to Nofinity

• On June 12, the US government, for the first time, shut down live AI models: Fable 5 and Mythos 5 by Anthropic.
• Reason: A jailbreak (security vulnerability) allowing identification of software flaws – reportedly discovered by Amazon researchers.
• Consequence: Global blackout within 90 minutes, affecting even US citizens.
• New dimension: A commercial chatbot was classified as an export item under the Deemed Export Rule. Not a geographic filter, but a citizenship-based access revocation.

• Decentralized AI networks like Bittensor (TAO) have no central kill switch – no CEO to receive a letter.
• After June 12, capital flowed into crypto-AI assets: TAO +28–30%, Venice (VVV) also rose.
• Core thesis: Censorship-resistant infrastructure as protection against state control.

• An IC3 study (Cornell, Princeton, etc.) highlights three issues:
  1. Efficiency gap: Decentralized networks cannot match centralized data centers for large models.
  2. Junk problem: Many AI tokens are pure branding without real decentralization.
  3. Creeping centralization: Even Bittensor has large holders and dependencies (TSMC, Nvidia).
• Alternative: Open-weight models (e.g., Meta Llama, Chinese models) can be self-hosted without blockchain.
• Reality: TAO is 68.5% below ATH; Grayscale reduced its TAO weight in May.

• Two truths at once:
  • The kill switch is real and permanent (proof of concept on June 12).
  • Decentralized AI has yet to prove cost efficiency and scalability.
• Your Watchlist:
  1. Will Fable 5 / Mythos 5 come back online?
  2. August ETF decision (Grayscale, Bitwise)?
  3. Is TAO’s on-chain revenue growing (~$43M/quarter)?
  4. Will sovereign nations choose open-source self-hosting over crypto?
  5. Separation of real revenue projects from hype-driven branding.
• Final question: Is decentralized AI the censorship-resistant answer – or just the next narrative fueled by fear?

”The kill switch is real. The only open question is whether these networks can deliver when Washington next reaches for the off switch.”

In this interview, Arthur Hayes (BitMEX, Maelstrom) discusses current market moves and his thesis of an impending AI credit bubble.

Bitcoin Rally and Debt Crisis

• The recent Bitcoin rally above $69,000 was triggered by a US Treasury buyback upsize for long-dated bonds.
• Key takeaway: This signals an implicit yield curve control (like Japan). The US will finance its debt at any cost – through money printing.
• Bitcoin is the purest hedge against this debasement. Hayes expects a long-term target of $1 million per Bitcoin.

AI as the Next 2008 Crash

• Hayes sees the AI infrastructure boom as a credit bubble modeled on the 2008 housing crisis.
• The phases:
  1. Overbuild: Hyperscalers build data centers with borrowed money.
  2. Credit Stress: These are real estate developers, not tech companies. Their debt is packaged into opaque securities.
  3. Bailout: The state will rescue the bubble with more money printing, boosting Bitcoin further.

Flop Labs: The Economy for AI Agents

• The Problem: AI agents need a currency directly convertible to compute (FLOPS) – not fiat or stablecoins controlled by humans.
• The Solution Flop: A blockchain network creating a global spot market for computing power.
  • Mining: Anyone can offer GPU power and earn FLOP tokens (block subsidy).
  • Usage: Agents pay with FLOP for compute and can store their "memories" (context) there.
  • Token Incentive: Initially, compute is almost given away to create network effects.
• Success Metric: The amount of inference processed (actual compute work done).

Price Predictions and Outlook

• Bitcoin (in 12 months): Above $200,000.
• Ethereum: Between $5,000 and $10,000 – Hayes sees ETH as the best risk-reward trade in crypto due to its weak performance.
• His thesis would be wrong if AI enthusiasm fades entirely and agents are not used.

This video explores the Better Vessel, a tool that helped a Filipino barista achieve the highest espresso scores at the 2025 World Barista Championship.

• Three parts: A top chilling sphere (added by the creator, not official), a double-walled main vessel, and a movable base with a stand.
• How it works: Espresso is pulled into the vessel, left to rest for 15-20 seconds, then placed on the final cup – the sphere lifts, trapping the crema while the clear espresso flows through.

• Crema is bitter – it's an emulsion of oils, fines, and CO2.
• Removing it yields a sweeter, more acidic, and silkier shot.

• The creator pulls a double shot: half normally (with crema), half with the Better Vessel (without crema).
• The Better Vessel shot: smoother mouthfeel, more prominent fruit flavors, less bitterness – almost like fruit juice.

• In barista competitions, judges award points for undisturbed crema on the surface.
• The Better Vessel allows competitors to show perfect crema, then quickly remove it – a major time-saver (12 drinks in 15 minutes).

• Pros: Fast cooling, effective crema separation, easy to use.
• Cons: Expensive, limited production, single-shot capacity, fragile 3D-printed base (can crack).

• Not a necessity, but a fun way to experience espresso differently. Traditionalists may still prefer the classic shot with crema.

No imperial units to convert; all measurements remain as in the original.

Summary follows...

Summary follows...

Summary follows...

The speaker analyzes whether Bitcoin has already bottomed or if further declines are ahead. He compares the current market structure to historical cycles (2018, 2019, 2022) and highlights the uncertainty.

• Apathetic Top Instead of Euphoria: Unlike previous bull markets, 2024/2025 lacked a euphoric rally. This might mean a full market reset isn't necessary.
• Historical Parallels: Comparisons to 2019 show that after a 40% rally in two days, Bitcoin hit a new low 31 days later. Similarly, 2018 had three lower highs (around $9,000, $8,500, $7,400), mirrored by today's $90,000, $80,000, and $70,000 levels.
• 200-Day Moving Average: A sustained break above the 200-day MA would be bullish. Currently Bitcoin is about 10% above it, while in 2019 it reached 18% above before falling again.
• Macroeconomic Factors: The end of Quantitative Tightening (QT) and three Fed rate cuts mirror the 2019 environment. Political news (Trump/China) also drive short-term rallies.
• Strategy: The speaker advocates Dollar-Cost Averaging (DCA) over market timing. He buys Bitcoin below a risk metric of 0.3 but hasn't accumulated enough for the next bull market yet.
• Caution on Short-Term Rallies: Before the final bottom in every midterm year, a strong rally often occurred (e.g., 23% in six days in 2022) followed by a selloff.

The speaker remains neutral: both bullish and bearish signals exist. The key is whether Bitcoin holds above the 200-day MA in the next two weeks. He advises against shorting and emphasizes a disciplined DCA plan.