EricG
2.5K posts

EricG
@EricG_Crypto
Hunter of crypto gems across all chains | (Re)posts are not financial advice | No paid promotions, just on my grind | Join my tg channel 👇
Valhalla Katılım Haziran 2019
352 Takip Edilen1.3K Takipçiler

For years, bundling tech has become more and more advanced until the trenches are no longer recognisable.
Expose the bundles, save the trenches.
jazz@youngjazzeth
cleanse the space we once loved.
English

Join us tomorrow for the @BitAngels Pitchcast Deep Dive!
⏰ 8–9:30 AM PST
We won the pitch competition at Consensys, and now we're going to show everyone why.
#BitAngels #Web3 #PitchcastDeepDive
English

Most attractive men in web3
Also based on profitability, dick size, personality, trading style, height, alpha, networth and some other aspects
thoughts?

Boru@boru_crypto
women in web3 tier list based on profitability, personality, trading style, long term, alpha, networth and some other aspects thoughts?
English

AWS, Azure and Google Cloud are built to milk you for idle compute time.
@KovaNetwork is literally fixing every pain point in legacy cloud:
👉 Fractional GPUs
👉 True per-second billing (no minimums)
👉 Zero vendor lock-in
👉 ~10x cheaper on average
Corporate monopolies vs. Token-governed dynamic pools. The choice is getting pretty obvious. Keep an eye on $KOVA 👀

English

The cloud monopoly is over!
@KovaNetwork is disrupting the $1T+ industry led by AWS, Google & Microsoft. Built by @Tettehnetworks & @amaechicrypto (Ex-Chainlink/Binance), $KOVA is:
🔹 90% Cheaper than AWS/GCP
🔹 No KYC & No VCs
🔹 Setup in <2 mins
🔹 On-chain proof of usage
English

Been digging into AI infrastructure plays and @KovaNetwork is looking like a massive sleeper pick right now 🕵️♂️
They're fixing the cloud waste problem by building a decentralized compute layer for AI - fractional GPUs/CPUs with true per-second billing. Basically up to 90% cheaper than AWS because you don't pay for idle time.
The partner list is wild: NVIDIA, IBM, Fetch.ai, Chainlink and many more.
They're launching their own hardware soon (cutting out the middlemen entirely) alongside the $KOVA token and mainnet. Don't fade teams building actual utility in silence 🧠
English

@EricG_Crypto @SynqXYZ Ok thanks bro, what about the mobile app for this terminal?
English

I officially made the switch from Axiom to @SynqXYZ
I’ve been deep in their beta for a while and the setup is genuinely fire. It actually fixes the biggest friction points in trading right now.
- You load one single USDC balance and can trade seamlessly across SOL, BASE, BSC and Monad
- Zero bridging hell. When a random 2-week BASE or BNB szn kicks off, you just catch the hype immediately
- Trade memes, perps and options all from the exact same account
- Built-in multiwallet to keep your positions private
It’s hands down the best trading terminal out right now.
I secured a 40% cashback for you, which makes it the platform with the lowest trading fees in the market 👇
app.synq.xyz/r/Eric
Synq@SynqXYZ
Announcing Synq: A trading platform for the onchain era. One USDC balance. Every chain. Memes, perps, options one click away. No bridging, no juggling wallets. Just trade.
English

🚨 Dappit just open sourced a brain for all computers.
It’s called Bird’s Nest. A local AI workspace built for the next generation of models beyond transformers.
Right now, almost every local AI tool — Ollama, LM Studio, GPT4All — is built around transformers.
Bird’s Nest is built for what comes next.
Instead of only running attention models like GPT or LLaMA, Bird’s Nest runs RNN and state-space models designed for constant memory and long context.
→ RWKV
→ Mamba
→ xLSTM
→ StripedHyena
These models don’t scale with O(n²) attention.
They run at O(1) per token.
Which means something huge:
Your context window doesn’t explode your RAM.
Example:
A 7B transformer model typically needs 16GB+ RAM.
Bird’s Nest can run comparable non-transformer models on 8GB machines.
And because these architectures don’t rely on KV caches, they can support effectively infinite context.
But Bird’s Nest isn’t just a model runner.
It turns your local AI into a tool-using system.
25+ built-in tools your AI can call mid-conversation:
→ Web search
→ YouTube search
→ Image generation
→ Music generation
→ File search
→ Python execution
→ Local RAG
→ System monitoring
→ Weather
→ Translation
→ And more
All 100% local.
No APIs.
No telemetry.
No cloud dependencies.
Your AI downloads models directly from HuggingFace and runs everything on your machine.
It even supports on-device image generation with models like:
→ SDXL Lightning
→ FLUX.2 Klein
→ FLUX.1 Dev
→ Qwen Image 20B
And music generation with:
→ Stable Audio Open
→ Riffusion
Everything runs inside a local runtime powered by FastAPI + WebSockets, streaming tokens directly to your browser.
Open the UI.
Download a model.
Start chatting.
That’s it.
Bird’s Nest is essentially a full local AI workstation built for the next wave of architectures.
And it runs on a laptop.

English


In Conjunction with the amazing @Player1Taco Dappit is proud to enter the #Monolith hackathon hosted by @RadiantsDAO Here is our new mobile app for you guys to try out on your @solana Seeker phone.
github.com/Dappit-io/dapp…

English

PolyGun Completes Strategic Acquisition of Polymarket Analytics (@Poly_Data)
Read: binance.com/en/square/post…
English

