ste

New podcast on vibe coding - A Return to Code. A Return to Coding 00:20 The Personal App Store 03:17 Vibe Coding Is a Video Game with Real-World Rewards 06:22 Pure Software Is Uninvestable 10:33 A Place for Each Model 14:22 AI Is Eager to Please 17:57 Why Math and Coding? 22:10 The Beginning of the End of Apple’s Dominance 24:17 Coding Agents As Customer Service Reps 27:55

Watching the news today feels a lot like watching a movie I’ve seen many times before. In 1971, I witnessed a major currency devaluation that I thought would be a disaster—but the markets actually went up. It was a lesson I only learned by looking back at 1933. When you study the historical "scripts," the current plot becomes much clearer.

No one seems to mention Scam Altman’s other company WorldCoin aka World which launched with a predatory low float crypto token $WLD that was on par with SBF / FTX companies. They have preyed on people from low income countries for biometric data by giving away small amounts of $WLD tokens. The tech was intended a way to prove you’re a human except it has resulted in a black market for verified accounts. The token supply inflates at unsustainable levels while insiders regularly OTC holdings.

Lium will become the first agent-first compute provider. We are leaning in heavily to the agent revolution. Here's how we make changes to lium now: We prompt our agents: "using this crypto account, create an account on lium, fund it, get an api key, rent some GPUs, and serve Kimi K2.6 with sglang (or another gpu-required task). When done, create a report for any issues you encountered with using lium and how it could have been faster or easier" Then we constantly improve the agent experience so that it's seamless to tell your agent to do any task, and it will just spin up lium gpus in seconds to solve it effortlessly and without any human intervention needed. We have recognized that SAAS companies need to pivot to agent-first software in order to keep up, and lium is the one leading the charge on that for cloud compute.

We're partnering with SpaceX to improve Composer. cursor.com/blog/spacex-mo…

@coinbureau lol, cmon man, wasn't mythos a top level security researcher??

🚨LATEST: Anthropic confirms breach of Claude Mythos model and has launched an investigation into the unauthorized access, per bloomberg. Attackers used contractor login credentials of a compromised third party vendor and guessed internal URLs to access multiple unreleased models. The group reportedly gained access on launch day despite the model being restricted to 40 firms due to cybersecurity risks Anthropic says “no evidence of impact to core systems has been found so far.”

SpaceXAI and @cursor_ai are now working closely together to create the world’s best coding and knowledge work AI. The combination of Cursor’s leading product and distribution to expert software engineers with SpaceX’s million H100 equivalent Colossus training supercomputer will allow us to build the world’s most useful models. Cursor has also given SpaceX the right to acquire Cursor later this year for $60 billion or pay $10 billion for our work together.

JUST IN: 🇺🇸 Fed Chair nominee Kevin Warsh says crypto is now part of the US financial system.

tl;dr: > Layerzero (LZ): "KelpDAO, please don't use 1/1 DVN" > KelpDAO: "no, can we please just use 1/1 DVN, and you operate the single DVN for us?" > LZ: "fine, if you insist. we will do it." > *LZ gets hacked*. > hacker minted rsETH for free and took out 200m+ of ETH on AAVE > LZ: "we told KelpDAO not to use the 1/1 setup. they didn't listen to us" but LZ operated the 1/1 DVN node! also: how did AAVE allow a 1/1 DVN asset into their protocol? I'm just catching up on this story but something tells me we can't just place all the blame on KelpDAO/rsETH.

An ASML Mafia is emerging. Fusion reactors. AI chips. Quantum computers. Autonomous robots. 24 alumni founders. 3 unicorns (12.5% conversion!). 8 countries. $13B combined value. Many of these founders stay and build around Veldhoven, a town of 45,000 people. Source: @dealroomco

In crypto and defi (ie in honest markets), when a component fails, those closest to the component—whether wildly negligent or innocent victim—suffer the loss, and are burdened with that responsibility. Unequal, but proper. In tradfi and banking (ie in coercively manipulated markets), when a component fails, the entire society is forced under the burden of its resolution. Costs are socialized. Equal, but improper. The former, with time, becomes self-correcting, self-improving, and crucially, retains vitality. The latter, regardless of time, becomes stagnant and soulless, and here everyone can wallow in an equivalent grey. Any man of agency should prefer the former, taking care over that to which he is proximate. It is from this that the virtue of markets emerges.

This is the moment NVIDIA should be seriously worried. In the next couple of weeks DeepSeek V4 will be launched. It’s a direct attack on the entire AI stack that American companies have spent years locking down. Full “de-NVIDIA-ization”, a complete shift away from CUDA into Huawei’s CANN ecosystem, running on Huawei Ascend chips. That means one thing, breaking the dependency that made NVIDIA untouchable. 35x faster inference vs early versions. Nearly 3x the performance of NVIDIA’s H20 on a single card. 40% less energy consumption. Over 95% CUDA compatibility with migration times collapsing from months to hours. Even Jensen Huang has already admitted it. If this works at scale, it’s a “terrifying outcome” for US companies. Because here’s the real problem, this isn’t happening in isolation. Chinese tech giants like Alibaba, ByteDance, and Tencent are already ordering hundreds of thousands of Ascend chips. Market share is shifting fast, domestic chips now at 41%, NVIDIA slipping to 55% in China’s AI server market. Additionally DeepSeek V4 is reportedly offering API costs at a fraction of US competitors. $300 for massive workloads that would cost $2,500+ on OpenAI models, or even $5,000 on Anthropic. So this isn’t just about one model. It’s about China building a fully independent AI stack, chips, frameworks, models, and applications. Completely outside of US control. NVIDIA doesn’t just lose sales. It loses its grip on the global AI standard.

They should change the name to "Strait of Schrödinger". It's both open and closed at the same time.

$TAO dropped 30% ... And most people still don’t get it. Everyone watched the price. Nobody watched this. While $TAO dropped from $378 → $235: → PwC France signed with $SN44 @webuildscore → $SN64 @chutes_ai hit a revenue ATH → @const_reborn told the community to rent machines from $SN4 Targon to fix a protocol issue Price was the distraction. The infrastructure kept building. $TAO #Bittensor

Sam Altman owns no equity in OpenAI. His salary was $66,000 in 2024. He says he does it because he "loves it." Here's what he actually does: He asked OpenAI to invest $500 million into Helion, a nuclear fusion company where he personally has $375 million invested. OpenAI employees were so uncomfortable they avoided the internal Slack channel discussing the deal, afraid their messages might end up in court. He pushed OpenAI to acquire Stoke Space, a rocket company his husband invested in through their family office. Some board members didn't even know about the talks. He got SoftBank's CEO to personally invest in Helion while SoftBank was in the middle of negotiating a $40 billion deal with OpenAI. Helion promised a breakthrough in 2024. Missed the deadline. Won't share details. Doesn't publish research. Outside experts can't evaluate the technology. Meanwhile at OpenAI: his top product executive is on medical leave, the company's AI lead over competitors is slipping, the video tool he championed got rolled back, and shareholders are privately floating his replacement. A man with no equity in the company he runs, hundreds of personal investments that intersect with its business, and a board that was fired once already for asking too many questions about exactly this. The $850 billion IPO is still on track. Of course it is.

We've started working with some Bittensor subnets. New Venice Uncensored 1.2 model was tuned on Targon/subnet4

Look at these headlines from past 24hrs… Goldman Sachs launching btc-related ETF. Morgan Stanley’s most successful ETF launch ever is btc ETF. Charles Schwab rolling out direct spot crypto trading. New York Stock Exchange all in on crypto. Wall Street has literally arrived.

When Claude Opus solves a math problem and "thinks" for 30 seconds before answering, what are you paying for? Reasoning tokens, billed at full output rates. Opus 4.6 charges $25 per million output tokens. A serious reasoning task can burn 5,000 to 10,000 reasoning tokens before producing an answer. That works out to $0.13 to $0.25 per query. GLM 5.1 TEE on Chutes runs $0.95 input, $3.15 output. The same query at the same depth of reasoning costs $0.016 to $0.032. About 8x cheaper per reasoning query, with hardware-enforced privacy included. GLM 5.1 scores 51 on the Artificial Analysis Intelligence Index, two points behind Claude Opus 4.6 and leading among open-weights reasoning models. 200K context, full chain-of-thought, JSON mode, structured outputs. It also holds the state-of-the-art on SWE-Bench Pro at 58.4, ahead of Opus 4.6, GPT-5.4, and Gemini 3.1 Pro. If you're building anything that needs the model to think before answering (math tutoring, scientific analysis, deep code review, multi-step research agents), this is the one to benchmark against your current default. What kind of reasoning workload are you running? Curious if Opus is overkill for it. 🔗 chutes.ai/app/chute/b048…

The world doesn’t realize this yet but AI will finally make philosophy a highly marketable skill

900B in yearly remittances volume with high fees charged to the poorest people. Now Tether just released a Wallet designed for them, self-custodial, super-easy to use, that allows to move digital dollars, gold and bitcoin with fees as low as $0.00001. People are already saying to their friends: "Tether me some money". And now they have an app for it. tether.me