Inferix GPU @InferixGPU

Big warnings

Jacob Coxon @hilbertspaess

a month ago

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

21K 167K 801K 175.3M 273K

AI agents will reshape the financial landscape

Animoca Brands @animocabrands

2 months ago

AI agents will reshape the financial landscape, but how? Our executive chairman and co-founder @ysiu joined @miaferrariii on @MadSocietyTV to discuss why autonomous agents need programmable money instead of traditional banking. It is inevitable that digital assets will fuel the

8 3 24 9K 4

The @iotex_io team is relentless. We’re accelerating our real-world AI mission, starting with local AI for everyone at rapidmlx.com. Computing intelligence used to move away into institutions and clouds. Now, the arc is bending back. We're building intelligence that lives, works, and grows with you—until the gap between mind and machine completely disappears.

IoTeX @iotex_io

2 months ago

The next phase of AI is personal. Start local for context, privacy, and speed. Reach the best models when needed. @rapidmlx is building the local intelligence layer for Personal AI. 🐆

78 53 268 48K 9

Crypto is currently facing its most profound existential crisis since genesis. Past headwinds—like government crackdowns, macro tightening, or brutal bear markets—were purely cyclical. But AI’s aggressive squeeze on 🧠 mindshare, 👨‍💻 top-tier talent, and 💰 liquidity feels structural and permanent. But here is my contrarian take: We are still miles away from true AGI. The current wave of AI is a technological marvel, but it will eventually hit a plateau. Why? Because achieving AGI isn't just a tech or compute-scaling problem. To truly integrate AGI, we need a massive restructuring of human society, economics, and organizational frameworks. You can't just throw GPUs+HBM at societal bottlenecks. When this current AI cycle inevitably peaks and reality sets in, we will witness a massive rotation. The holy trinity of tech growth (Attention, Talent, and Capital) will aggressively flow right back into crypto.

Cointelegraph @Cointelegraph

2 months ago

🚨 NEW: Bitcoin has never been below Fire Sale for this long.

433 457 4K 801K 312

DeepSeek-V4-Flash-0731 literally wrote its own PR, added itself to Rapid-MLX v0.11.9, and optimized its own performance for Apple Silicon! 🤯🤖 Running locally on an M3 Ultra (256GB), the speed is absolutely unreal: ⚡️ ~30 tok/s single-request decode 🧠 284B total / ~13B active MoE (MXFP4) 🚀 DSpark decoding & Prefix-caching for blazing-fast multi-turn workflows! 🛠️ Native tool calling + Codex ready Want to run this beast locally with Codex? Just 2 simple steps: 1️⃣ Spin up the server: rapid-mlx serve /path/to/DeepSeek-V4-Flash-0731-MLX --speculative-config '{"method":"dspark","num_speculative_tokens":5}' --enable-prefix-cache --enable-auto-tool-choice 2️⃣ Update your ~/.codex/config.toml: model = "deepseek-v4-flash-0731" model_provider = "rapidmlx" model_context_window = 1048576 model_max_output_tokens = 32768 [model_providers.rapidmlx] name = "Rapid-MLX (local)" base_url = "http://127.0.0.1:8000/v1" Try it out here! github.com/raullenchai/Ra…

DeepSeek @deepseek_ai

3 months ago

🚀 DeepSeek-V4-Flash Official API is now LIVE in public beta! 🔷 We’ve massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! 👇 🔷 The official V4-Flash now natively supports the

2K 3K 30K 9.6M 5K

Weird but true!

Haseeb Qureshi ＞|＜ @hosseeb

3 months ago

Weird but true: without crypto, the current AI boom would not have been possible. Today's data center buildout would've been impossible without Bitcoin miners spending a decade building the knowhow around power, sites, and scale in remote locations. Nvidia could not moved

119 97 1K 68K 174

Pay-Play-Proof. Congrats @syfuofficial

SyFu | Real-World Spending × GameFi @syfuofficial

3 months ago

◤ SyFu Mainnet β is LIVE ◢ Turn everyday spending into a Web3 game. SyFu turns payment data into digital assets— and spending history into a foundation for future credit. Pay → Play → Proof. The first SyFu loop starts now. Built on @BNBCHAIN. syfu.io/product

125 71 194 60K 5

Behind every block on IoTeX, there's a builder. iotex-core, the Go engine powering Real-World AI, is 100% open source. The community keeps showing up: -120+ open-source contributors -3,500+ pull requests -100+ releases The infrastructure for Real-World AI, built in the open👇

Just landed on QuickSilver Pro: Two new frontier models. ⚡️ 🔹 OpenAI GPT-5.6 (Luna / Terra / Sol) 🔹 Anthropic Claude Fable 5 — Mythos-class flagship Unlocking: ✓ Top-tier reasoning & long-horizon agentic coding ✓ 1M context window ✓ Just ONE OpenAI-compatible key Start building: quicksilverpro.io

AI demand is outpacing GPU supply. At the same time, most deployed GPUs sit idle. @RenderNetwork's Dispersed aims to bridge that gap through a decentralized marketplace for AI and general-purpose compute. > A distributed GPU marketplace connecting developers and enterprises with underutilized GPUs worldwide. > On-demand compute for AI inference, model training, scientific research, autonomous agents, and more. > Node Operators earn RENDER by contributing idle GPU capacity and completing workloads. > A Burn-Mint Equilibrium model where 95% of compute payments are used to burn RENDER, creating direct token value accrual. > Early deployments powering autonomous AI agents, scientific reproducibility research, generative art, and sovereign AI memory infrastructure. Our latest report explores how Dispersed extends the Render ecosystem beyond rendering and whether decentralized GPU marketplaces can become a viable alternative to traditional cloud infrastructure. Learn more here 👇

Trio-Lumen turns plain English into a live perception rule in under a minute: 1️⃣ Language compiles into a structured detection rule 2️⃣ Real-time tracking (~30ms) + VLM semantic judgment per subject 3️⃣ Every alert fires as a structured event — frame, zone, track ID Zero labeling. Zero training. Zero CV pipeline. For Developers: Perception becomes a programmable primitive — any camera turns into an event stream you can build on: automation, safety, agent workflows grounded in real-world observation. This is world-model-powered physical understanding, in production. 🎥👇

🚨 JUST IN: The IoTeX Mainnet v2.4.0 upgrade has successfully activated at block height 48,985,561 $IOTX

26 protocols so far: - Ethena - Ether_fi - trondao - CurveFinance - BitGo - WrappedBTC - River - Pengu - Morpho - Agora - F(x) protocol - Apecoin - Euler - Katana - Orderly - Matrixdock (wtf is this ??) - meth protocol - Solv protocol - Mocacoin - Re - Avant - Beefy - Flare - USDT0 - Lombard - InfiniFi That's crazy!!

Fishy Catfish @CatfishFishy

6 months ago

I'm dropping a thread of all the protocols that had to freeze their interop because of LayerZero being compromised. Let's go:

130 308 2K 338K 576

📣 April 6, 2026 #IoTeX Weekly Report 📣 1️⃣ IoTeX launches IoTeX Hub, deeply integrating governance, staking, data, and ecosystem dashboards, aiming to become the ultimate gateway to next-generation smart infrastructure. 2️⃣ The x402 Foundation incorporates Quicksilver into its ecosystem, enabling a real-world payment loop using USDC on-chain. 3️⃣ ioTube launches a compensation claim portal, providing 100% full compensation to users affected by the February 21st cross-chain bridge incident. 4️⃣ IoTeX plans to upgrade its network to full quantum security; related plans will be announced soon. 5️⃣ IoTeX L1 staking ratio: 38.78%. 6️⃣ Quicksilver AI data requests: 94,839. 7️⃣ #ioID has been integrated by 29 projects across 99,772 devices.

💥BREAKING: New York Times article, published today, concludes that Adam Back, a 55-year-old British computer scientist and cryptographer, is Satoshi Nakamoto.

keyboard_arrow_left Previous keyboard_arrow_right Next

Crowdsourced physical data will perhaps be the primary way that physical AI and robots will develop a shared understanding of the world. A key question is whether users should be compensated for this data and how? On a global basis the easiest way is DePIN.

NewsForce @Newsforce

7 months ago

POKÉMON GO PLAYERS TRAINED 30 BILLION IMAGE AI MAP Niantic says photos and scans collected through Pokémon Go and its AR apps have produced a massive dataset of more than 30 billion real-world images. The company is now using that data to power visual navigation for delivery

732 3K 22K 23.7M 8K

TODAY 🚨: The Commission issued an interpretation that clarifies the application of federal securities laws to crypto assets. This is a major step to provide greater clarity regarding the Commission’s treatment of crypto assets. Read the release here: ow.ly/XhhV50YvxvO

Every year, crypto projects publish roadmaps. Neat timelines. Color-coded phases. Quarterly milestones with checkmarks that nobody checks. We're not doing it this year. Instead, here's our Anti-Roadmap for 2026 👇

DePIN is growing up at @consensus_hk on Feb 11. Catch @larrypang (IoTeX) with @Austin_Federa (DoubleZero) + @fmong (Helium) on the Frontier Stage, during the AI & Robotics Summit. consensus-hongkong.coindesk.com/agenda/event/-…

x.com/i/article/2016…