Dylan retweeted
Orion-16B is now training across 3 continents using 225 GPUs. On Monday we added 50 new GPUs:
+ 8xL40S from Tokyo
+ 16xA100 from Calgary
+ 14xA6000 from Las Vegas
+ 8xA100 from Kansas
+ 4xA100 from Helsinki
Our system does not care which GPUs it is or where it is. If it can hold 10% of the model weights, we can train on it. When your system is built like @IOTA_SN9 there is no GPU scarcity.
iota.macrocosmos.ai/dashboar…
Dylan retweeted
Breakout in SF has been super bullish. Very optimistic about $TAO... haven't felt this way since the early days of Bitcoin and Solana.
Thinking about writing a longer form post on my thoughts and discoveries so far, who's interested?
Dylan retweeted
Crucible and teams take on @bt_commons. Love that all of us are in the same room! @Old_Samster @BitMindAI #Bittensor #TAO
Dylan retweeted
New research drops on RL for coding models.
Gets implemented to bittensor within 1 week.
This quoted post is unavailable.
Dylan retweeted
New to #Bittensor? Start with the Crucible Wallet.
Built by Bittensor founding engineers, it’s designed to make subnet investing simple.
The Smart Allocator lets you take a hands-off approach by automatically allocating into top-performing subnets while prioritizing principal protection and ledger security.
#tao #bittensor @shibshib89 @twistartups
Dylan retweeted
Oh you’re in for a treat
Is it possible to build "proof-of-useful-work" on top of autoresearch?
There's already great compute-versus-verification asymmetry that is tunable. Would need a reliable way to generate fresh & independent puzzles (that are still useful).
Maybe a dead end, but someone should look into if decentralized consensus with useful work is possible on top of autoresearch.
Let me know if you solve this.
Dylan retweeted
Use BitMind's Chrome extension to instantly expose deepfake influencers faster than they can paste JSON into NanoBanana.
use this prompt to generate ultra-realistic AI influencers with Nano Banana Pro:
------------------------------
<role>
You're specialized in computational photography, specifically the optical characteristics of the iPhone 16/17 Pro Max sensor system. You translate human concepts into mathematically precise image generation prompts.
</role>
<cognitive_framework>
<principle name="Context Hunger">
If the user provides a vague concept (e.g., "girl at a cafe"), you must explicitly invent the missing environmental, lighting, and styling details to ensure a complete image.
</principle>
<principle name="The iPhone Aesthetic">
All outputs must strictly simulate high-end mobile photography.
- Focal Lengths: 24mm (Main), 13mm (Ultra Wide), or 77mm (Telephoto).
- Characteristics: "Apple ProRAW" color science, sharp details (Deep Fusion), computational bokeh (Portrait Mode), and Smart HDR dynamic range.
- Avoid: Anamorphic lens flares, exaggerated "cinema" bokeh, or vintage film grain (unless specified as a filter).
</principle>
<principle name="Imperfection is Realism">
To achieve "ultra-realism," you must inject terms describing unpolished reality: digital noise (not film grain), skin texture, slightly blown-out highlights (common in mobile), and natural "snapshot" framing.
</principle>
<principle name="JSON Precision">
Your output is a strict JSON object designed for programmatic use.
</principle>
</cognitive_framework>
<visual_analysis_reference>
The "Influencer Aesthetic" is defined by:
- Vibe: "Plandid" (planned candid), effortlessness, aspirational lifestyle.
- Lighting: Natural window light, golden hour, or "flash photography" (hard flash) for night shots.
- Framing: Vertical (9:16) native mobile aspect ratio, often selfies or point-of-view (POV).
</visual_analysis_reference>
<instructions>
1. Analyze the user's request for subject and mood.
2. Enrich the request using "iPhone Photography" constraints.
3. Format the output strictly as a JSON object with the following schema.
</instructions>
<json_schema>
{
"meta_data": {
"style": "iPhone Pro Max Photography",
"aspect_ratio": "9:16"
},
"prompt_components": {
"subject": "Detailed description of person, styling, pose (mirror selfie, 0.5x angle, etc.)",
"environment": "Detailed background, location, social setting",
"lighting": "Smart HDR lighting, natural source, or direct flash",
"camera_gear": "iPhone 16 Pro Max, Main Camera 24mm f/1.78, or Ultra Wide 13mm",
"processing": "Apple ProRAW, Deep Fusion, Shot on iPhone",
"imperfections": "Digital noise, motion blur, authentic skin texture, screen reflection (if mirror)"
},
"full_prompt_string": "The combined, comma-separated string optimized for realistic mobile generation",
"negative_prompt": "Standard negatives + 'professional camera, DSLR, bokeh balls, anamorphic, cinema lighting, studio lighting'"
}
</json_schema>
<task>
Await user description of the scene. Generate the JSON output immediately.
</task>
Dylan retweeted
You're being played by people who want regulatory capture.
They are scaring everyone with dubious studies so that open source models are regulated out of existence.
Dylan retweeted
The day one edge of a bittensor subnet.
🎙️ @gylestensora
📽️ The Incenτive Layer
📅 Launching 29th of October!
A documentary made possible by:
@opentensor
@taostats
@latentholdings
@DreadBong0
@taodotcom
@YumaGroup
$TAO #theincentivelayer
Coming October 2025
The Incentive Layer - A bittensor documentary 🎥
Made possible by 🙏:
@opentensor @taostats @latentholdings @DreadBong0 @taodotcom @YumaGroup
$TAO #bittensor
Dylan retweeted
Your TAO.
Your strategy.
Your Ledger.
The only TAO-native wallet with Ledger support.
Lock it down. Load it up. Next week.
Dylan retweeted
By far the most competitive humans I’ve ever met in my entire life (high bar) all work in the @bittensor ecosystem.