Miles

Building a better future for myself and loved ones, one problem at time

[Miles @miles_engineer]

Yesterday: your margin was my opportunity.
Today: your capex is my opportunity.
Tomorrow: your agent is my opportunity.
Now: my agent is my opportunity.

[Miles @miles_engineer]

@tellthemimright @Kalshi Democrats: I won't help my child financially because it's got the Trump name.

[Miles @miles_engineer]

@Pritensor Are you going to buy any more $TAO?

[Miles @miles_engineer]

@TrustlessState @ErikVoorhees The next 'armored warrior class'

[Miles @miles_engineer]

@CernBasher The fundamental problem with using anything other than PE (or the other main metrics) is education. Someone has to spend money on educating institutions/financial advisors, and getting the new metric adopted across platforms like TradingView/Bloomberg.
It's hard to change a mind.

[Ansem 🐂🀄️ @blknoiz06]

there's a million dollars in your computer you just have to press the right buttons to get it out

[Miles @miles_engineer]

@AussieKidBilly @JensenHuang Speculating - next year Jenson could mention Bittensor by name and promote it as a viable open source path.

[Miles @miles_engineer]

New kid on the block👋

[Jensen Huang @JensenHuang]

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.

AI will transform every industry, power every company, and be built by every country.

Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.

The world needs both frontier closed models and frontier open models.

https://t.co/AUKzoQ5Ikb

[Miles @miles_engineer]

@eddyluxe_ Life not that good. You always commenting on single women’s posts. 🤪

[Miles @miles_engineer]

$STRC higher low 👀

[Miles @miles_engineer]

@NateGeraci And some RWA’s

[Miles @miles_engineer]

@ChrisMMillas Loading up at $9 if we get there

[Miles @miles_engineer]

@blknoiz06 Low Driver ratings based on poor tips

[Miles @miles_engineer]

Socials got your girl.
AIs got your boy.

[Miles @miles_engineer]

@rektdiomedes @ChrisWinig I remember the first time I laid eyes on these kings of old. Will never forget.

[Miles @miles_engineer]

$220 $TSLA incoming. Just broke through support its been holding since April 2024.

[Miles @miles_engineer]

#TSLA will hit $220 before a multi-year rally similar to 2019 lows

[Miles @miles_engineer]

@KobeissiLetter Jenson said theirs going to be $90 T in ai capex spend. So a long ways to.

[Miles @miles_engineer]

@Polymarket Stocks only - what about tokens? RWA tokenization incoming.

[Miles @miles_engineer]

@here4impact @RokoNetwork How does the protocol make revenue?

[Miles @miles_engineer]

@business Bullish for $LLY

[Miles @miles_engineer]

Orchestration of agents will be the most profitable layer of token demand, I believe TAO and PATH will capture most of this value.

[Michael D. White @here4impact]

the us gov suspending @AnthropicAI's fable was a watershed moment for #bittensor because you cannot censor decentralized AI...

but candidly I think today's release from @SakanaAILabs is even more important.

for the first time, the entire AI community is waking up to the idea that model context and harnessing is more important than having one underlying model that outperforms in aggregate.

THE OLD PARADIGM: *who has the best model?*

closed-source AI shops producing extremely capable LLM's that excel on many topics (one size fits all). open source trails closed source.

THE NEW PARADIGM: *who has the best orchestration?*

AI protocols blending multiple LLM's (and LLM distillations) for specific agentic sub-tasks within a larger project framework. open source excels and closed source lags here.

#bittensor is best-positioned to excel in this new paradigm as a decentralized protocol where 128 subnet teams produce commoditized AI outputs that can be rolled up for 1) higher efficacy for specific tasks and 2) dramatically lower token cost. multiple specific models used in aggregate by multiple agents working towards a common goal will vastly outperform closed-source LLM's over time. enterprises are burning through token spend; and this is the cost- and effort-effective answer.

all subnets will benefit from this emerging school of thought which @const_reborn and @shibshib89 deftly predicted 5+ years ago (a lifetime in the tech world).

eg:

@lium_io for compute (avoid hardware capex and pay for use)
@chutes_ai for inference (select the best model for your specific task)
@arbos_born for distillation (targeted selection for agentic sub-tasks)
@heydittoai for persistent context (running multiple instances and/or agents across multiple models and distillations for high-level task aggregation and project completion)

some highlights that are getting more obvious by the day:

1) context will matter more than the model

Trinity (by Sakana; above) solves for conversation level context but @heydittoai expands that to make context persistent.

2) different models are better at different tasks

an orchestrator that chooses between models (including distilled) will outperform a single model on a project-level basis. it won't even be close.

3) cost implications

organizations need to be cost-conscious re: token spend. the blank checks have been torn up. the above architecture will allow for specific project-based token budgets for optimized output.

bottom line: *this all needs to be open source* and bittensor is the answer. how will you as an enterprise customer know if your workflow is really being optimized for cost and efficacy if you have zero viability into the process?

bittensor:native