smokin ape

[Victor @victormelillii]

https://t.co/L6kS7cvJ4s

[smokin ape @crypt0xbuddy]

$btc $hype $kntq $pod

[Eli5DeFi @Eli5defi]

I think, $POD from @dphnAI is one of the most interesting decentralized AI runners right now: uncensored open-source models, powered by DePIN.

Launched in 2023 by Eric Hartford and team, Dolphin has already shipped dozens of uncensored models on Hugging Face with millions of monthly downloads.

Now, post-rebrand, they’ve released the full peer-to-pool tokenomics.

Here’s the no-BS, 30s breakdown.

-

❶ What are they actually tackling?

- Censored AI models
- Expensive & centralized inference
- Wasted idle GPU supply
- Synthetic data bottleneck
- Decentralized verification complexities

This is achieved via two solutions.

→ Fine-tuning top open-source bases (Llama, Mistral, Qwen, etc.) into high-quality, truly uncensored LLMs that keep (or even boost) intelligence, creativity, tool-use, and roleplay ability.

→ Building the Dolphin Network – a distributed GPU network where anyone can run nodes with idle hardware (gaming PCs, consumer cards, whatever) to power cheap inference + massive synthetic data generation.

-

❷ Dolphin Network - Peer-to-Pool Design

Unlike most 1:1 rental DePINs, they use pools:

- Nodes running the same model get grouped together.
- Requests are randomly assigned (no direct buyer-seller link).
- Nodes earn purely based on tokens processed (paid from protocol treasury).
- Users buy credits straight from the protocol (POD, ETH, BTC, USDC, even Monero/Zcash accepted).

Dolphin Mistral 24b now is the default uncensored model for all @AskVenice user.

The roadmap is pragmatic, but ambitious:

(Now) Distributed inference beta

Upcoming (In Order):
→ Image / video / audio / transcription nodes
→ Synthetic data generation suite
→ Lighthouse auto-balancing
→ Sharded distributed inference
→ Distributed LoRA / SFT (supervised fine-tuning)
→ Distributed reinforcement learning
→ Distributed full-parameter SFT
→ Model creation & fine-tuning suite
→ Large-scale distributed pre-training

-

❸ POD

Utility & value accrual:

- Staking (xPOD vault) → auto-compounding dividends from buybacks + daily inference credits + perks.

- Node operators bond POD (slashable) for rewards. Curve-style boost multipliers the more you commit relative to earnings.

- Validators with bigger bonds help secure the network.

➠ Key Point: 100% of revenue auto-buys POD on the open market → constant, real buy pressure that can outpace emissions.

-

❹ Bottom Line

Usage = revenue = buybacks = value to stakers and bonded operators. no external equity means the POD token captures everything.

It’s built like a mix of ETH (slashing), CURVE (boosts), and XSUSHI (auto-compound) but tuned specifically for usage-based ai inference + synthetic data.

[Cobie @cobie]

if you scared to take a chance, how the fuck we gon get rich?

[Bull Theory @BullTheoryio]

SAM BANKMAN FRIED PICKED EVERY WINNER OF THE 2020s AND HIS LAWYERS SOLD THEM ALL AT THE BOTTOM.

If the FTX estate hadn't panic-sold its assets during bankruptcy, SBF would be sitting on a $114 billion empire today. Instead, he is watching the greatest trades of the decade from a prison cell.

The data is almost impossible to believe:

- Anthropic: $82.3 billion (165x) SBF bought an 8% stake for $500M. The estate sold it for $1.3B in 2024. Today, that stake would be worth over $80B.

- SpaceX:$15 billion (75x) A massive stake liquidated early to pay creditors.

- Solana: $5.1 billion (27x) SBF was an early backer at $8. The estate offloaded a massive chunk at $64.

- Robinhood: $4.9 billion (8x)

- Genesis Digital: $3.5 billion (3x)

The Latest "Missed" Fortune: CURSOR

In 2022, Alameda Research wrote a tiny $200,000 check for a 5% stake in the AI startup Cursor. In April 2023, the bankruptcy estate sold that entire stake back for exactly what they paid: $200,000.

Yesterday, SpaceX announced a deal to buy Cursor for $60 billion.

That "worthless" 5% stake would be worth $3 billion today. That is a 15,000x return that vanished because the lawyers wanted a quick exit.

SBF was a genius at picking generational winners and a criminal at managing their money.

The lawyers recovered $18 billion for users. If they had just held, they would be sitting on $114 billion and the most valuable venture portfolio in history.

[Flood @ThinkingUSD]

If you're in Crypto and you're not bullish on Hyperliquid after reading this article you should really leave and go do something else.

https://t.co/BJlMIWhosG

[Hyperliquid Research Collective (HRC) @HyperliquidR]

Hyperliquid Annual Report 2025

Today, we're excited to release Hyperliquid's 2025 Annual Report.

What Hyperliquid, the core team, and hundreds of contributors have built this year is largely unprecedented in financial history. This report is our attempt to capture that extraordinary year with the rigor it deserves.

A few months ago, we created HRC because we felt a clear need to reduce information asymmetry and lower entry barriers for new participants through independent research. We hope this report helps do exactly that.

For us, it is a privilege to be part of this ecosystem and to have produced this work. Months of research, data work, debate, and collaboration across Four Pillars, GLC, and all our contributors went into every page. It means a lot to share it today, so please, let us know what you think and how we can improve the reporting going forward.

Enjoy the read. Some excerpts below, with the link to the full PDF.

Hyperliquid.

[smokin ape @crypt0xbuddy]

@0xOmnia @Kinetiq_xyz @Markets_xyz KNTQ fucking coded

[KoolKrypto @koolkrypto223]

If you want a good case study on how much easier it is to have actionable and profitable informational asymmetry in crypto as opposed to tradfi, look no further than $KNTQ.

In tradfi it's "Oh you weren't at the child sacrifice to Baal themed yacht party where we decided which billionaires sons' shell company was being awared the 10 figure government contract? NGMI lil bro"

Meanwhile @Kinetiq_xyz just casually dropped that they were changing tokenomics to intrinsically and programatically tie themselves to $HYPE valuation, probably double+ TVL, revenue and $kHYPE liquidity/peg stability, and increase token buybacks from 0% to 70%. Barrier to entry for this information was an elementary school reading level and reading the patch notes a week ago. Price still unchanged from then.

At $100m FDV, you're buying the first $HYPE beta the suits will look at for a catch up trade after they miss $HYPE to $100, with total market share domination in their vertical (I couldn't even tell you the next most popular $HYPE LST tbh). Now they'll direct millions of dollars of revenue to buying a $28m mcap token (39% of which is staked).

$KNTQ current valuation also currently prices in @Markets_xyz being an absolute zero, while it's quote asset, $USDH, just became the cleanest way to onboard new capital and users to @HyperliquidX.

$KNTQ is currently the only liquid way to gain exposure to a HIP3 deployer(s), and bet on $USDH proliferation. You're getting exposure to those for free at this valuation.

You only see prices diverge this aggressively from fundamentals at market pico-tops and pico-bottoms.

[Anthropic @AnthropicAI]

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software.

It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans.
https://t.co/NQ7IfEtYk7

[smokin ape @crypt0xbuddy]

@Kinetiq_xyz @Markets_xyz heavily coded.

[smartestmoney.hl @smartestmoney]

You will own nothing but @HyperliquidX and you will be happy.

[Ponyo @ponyo_fp]

https://t.co/vqvlm9xYmw

[AI Edge @aiedge_]

https://t.co/YPRrgBwqcU

[jeff.hl @chameleon_jeff]

Hyperliquid has quietly achieved an important milestone of becoming the most liquid venue for crypto price discovery in the world. See below for side by side comparison of BTC perps on Binance (left) and Hyperliquid (right).

With HIP-3 teams leading the way, Hyperliquid has also grown to become the most liquid venue for perps on tradfi assets. Thank you to everyone's hard work as we upgrade the financial system and house all of finance.

[smokin ape @crypt0xbuddy]

@lbexplorer Who the fuck is still building?

[smokin ape @crypt0xbuddy]

@Derivatives_Ape what do you say to the counter argument that “ico demand not equal post-ico demand” ?

[smokin ape @crypt0xbuddy]

@OnlyLJC No.

[smokin ape @crypt0xbuddy]

@SolumSursum @GwartyGwart @cobie @ledgerstatus How do you leave out @gametheorizing and steady lads

[thiccy @thiccyth0t]

https://t.co/eEzuNK85rh