Non-KYC debit card with tap to pay. Build this and I will give you so much money.

Untitled spark

xNostr

Fmr CIA officer here. Let me share how elections get stolen — and the role of intel agencies. We a...

How Elections Get Stolen: A Former CIA Officer's Warning A former CIA officer recently took to social media to lay out what he describes as a largely hidden truth about American elections — one that, he argues, both parties in Washington quietly understood for years but rarely discussed openly. The thread, posted by Bryan Dean Wright, draws on his intelligence background to connect the dots between classified capabilities and domestic electoral vulnerability. What the Intelligence Community Knows Wright's central claim is straightforward: no voting machine is truly "unhackable." According to him, the CIA and NSA maintain dedicated teams capable of penetrating virtually any electronic system — including voting machines stored in locked, air-gapped rooms with no internet connection. These operations, he says, rely on a combination of human intelligence (HUMINT), signals intelligence (SIGINT), and other technical disciplines, along with meticulous planning and what he calls "exquisite tradecraft." What makes this claim notable is his assertion that members of Congress on both sides of the aisle were regularly briefed on these capabilities and privately acknowledged the danger. Historical footage from Congressional hearings, he notes, shows both Democrats and Republicans expressing alarm at the idea of so-called "unhackable" systems — because they knew better. "They were briefed by NSA and CIA teams on how they 'hack the unhackable' every month without leaving a trace." The Taiwan Model — and Why America Ignores It Wright holds up Taiwan as a practical counterexample. Despite having a population of roughly 24 million people and operating under constant geopolitical pressure from China, Taiwan runs its elections with striking simplicity: Same-day voting only In-person participation with ID requirements Paper ballots A counting process open to public observation Results announced the same night Wright's argument is pointed: if Taiwan can pull this off reliably and transparently, there is no technical or logistical reason American states cannot do the same. The refusal to adopt similar safeguards, he contends, is not an oversight — it is a deliberate choice that creates exploitable vulnerabilities. He singles out California specifically, noting that the state's roughly 30-day vote-counting window is, in his view, an invitation for manipulation. The extended timeline and widespread use of mail-in and harvested ballots, he argues, provide cover for the kind of untraceable interference that intelligence professionals are trained to execute. The Playbook: Deny, Smear, Deflect Wright describes what he sees as a recognizable pattern whenever electoral irregularities surface. Like any well-run covert operation, the goal is to leave no fingerprints. But when something does leak, he says the response follows a familiar script: deny the allegation, attack the credibility of whoever raised it, and launch counter-accusations designed to muddy the waters and preserve the status quo. This, he argues, is not coincidence — it is operational discipline applied to domestic politics. The Bigger Picture Wright frames all of this around the Los Angeles mayoral race involving Spencer Pratt's challenge to incumbent Mayor Karen Bass, but he is explicit that his larger concern goes well beyond any single contest. He describes what he calls a "corrupt machine" — a structural system, particularly in Democrat-dominated cities and states, that he compares to the legendary political machines of old Chicago, New Jersey, and Louisiana under Huey Long. His conclusion is stark: candidates running against entrenched incumbents in these environments are not simply competing against a political opponent. They are competing against a system engineered, he believes, to produce predetermined outcomes. What Should Readers Make of This? It is worth noting that Wright's account is one perspective — an opinion thread, not a formal intelligence report or court filing. Claims about specific electoral fraud in any given race require evidence and due process to evaluate fairly. Reasonable people disagree sharply about the scale and nature of election integrity problems in the United States, and courts have repeatedly found insufficient evidence to overturn results in cases brought after recent elections. That said, the underlying technical point — that sophisticated state-level actors possess tools capable of compromising electronic voting infrastructure — is not fringe speculation. Cybersecurity experts and government agencies have long acknowledged vulnerabilities in voting systems, and the push for paper ballot backups and robust auditing has bipartisan support in many security circles. The debate Wright is entering is real, even where his specific conclusions remain contested. The question of how to build electoral systems resilient enough to earn broad public trust — and whether current systems meet that bar — is one that democratic societies cannot afford to stop asking.

xNostr

this is literally the best deal on the market rn for an ai coding plan > $10/month > kimi k2....

The $10/Month AI Coding Plan That's Quietly Outperforming Its Price Tag There's a quiet revolution happening at the budget end of the AI coding market. While most developers are paying $20 a month for tools from the big Western labs, a growing number are discovering that a $10 subscription can deliver more raw usage, more model variety, and surprisingly competitive performance — if you know where to look. The buzz right now is centered on one specific tier: a $10/month plan that bundles access to three powerful Chinese frontier models — Kimi K2.5, MiniMax M2.7, and GLM-5 — under a single subscription. Users who've tried it describe it as "way more usage than $20 plans with big labs" and "super underrated." That's a bold claim. But when you dig into the numbers, it holds up. What You're Actually Getting The plan in question is the Atlas Cloud Coding Plan Starter tier , which provides 800,000 credits per day across a lineup of models that would have been considered frontier-class just months ago. At that credit budget, a typical medium-sized coding request — say, 5,000 input tokens and 1,000 output tokens — gets you somewhere between ~50 and nearly 500 requests per day , depending on which model you use. Lean on the faster, lighter models for routine tasks, and the heavier ones for complex reasoning, and that daily budget stretches remarkably far. The three headline models are worth understanding individually: Kimi K2.5 — Built by Moonshot AI, this model shines on research-heavy coding tasks. It carries a massive 256K token context window (the largest of the three), scores impressively on math and scientific benchmarks (AIME 96.1%, GPQA 87.6%), and supports image input — making it a strong pick when you're wrangling long documents or need deep reasoning alongside code generation. MiniMax M2.7 — The speed champion of the group, running at up to 100 tokens per second on the high-speed tier and hitting a 78% SWE-Bench Verified score , which sits at or above Anthropic's Claude Opus 4.6 on the same benchmark. It's especially well-suited to production coding workflows where low error rates matter — one real-world 48-hour test on a Next.js/Convex project showed roughly 5 review fixes needed vs. ~50 for Kimi . Its pricing is also the cheapest of the three at around $0.30 per million tokens. GLM-5 — Developed by Zhipu AI and trained entirely on Huawei Ascend chips (no Nvidia hardware involved), GLM-5 is a 744B parameter model with 40B active parameters per token. It scores 77.8 on SWE-bench , rivaling Claude Opus 4.6, and is fully MIT-licensed — an unusual perk for a model at this capability level. How It Stacks Up Against the $20 Competition To put this in context, here's what the standard $20-per-month options offer: Claude Pro ($20/month) — Access to all Claude models plus Claude Code CLI, but on a shared token budget that can feel constraining for heavy daily users. ChatGPT Plus ($20/month) — GPT-4.5, some Codex access, and general-purpose tools. Solid for mixed workflows but not deep coding focus. Cursor Pro ($20/month) — Great IDE integration with unlimited tab completions and a frontier-model credit pool, but the credits run out fast on intensive sessions. GitHub Copilot Pro ($10/month) — The most obvious comparator at the same price point, offering inline completions and agent mode, but you're locked into GitHub's model choices with no flexibility. What the Atlas Cloud Starter plan offers that most of these don't is model optionality . Rather than being tied to one provider's decisions about which model you get, you can route different tasks to the model best suited for them — and all through an OpenAI-compatible API format that works with Claude Code, Codex, and any other tool that speaks that standard. The cost efficiency is striking when you zoom out. Running a million API calls per day through Claude Opus 4.6 would cost roughly $150,000 a month . The same workload through MiniMax M2.5 runs about $9,000 — a 94% reduction. Even at individual developer scale, those efficiency ratios matter when credits and rate limits hit. Who This Plan Makes Sense For This isn't necessarily the right plan for everyone. A few honest caveats apply. If your workflow is deeply integrated with GitHub or VS Code and you rarely think about which model is running underneath, GitHub Copilot Pro at the same $10 price is simpler and probably good enough. If you're a heavy Claude Code user and the quality difference matters on nuanced tasks, Claude Pro may still justify its $20 tag. But for developers who: Run multiple coding sessions daily and keep hitting rate limits on mainstream plans Want to experiment with different models for different task types Are comfortable working via API or tools like OpenCode, Cline, or Claude Code pointed at a custom endpoint Are curious about the rapidly improving Chinese frontier models …this plan is worth a serious look. One practical tip from developers who've tested these plans: don't abandon your existing subscription cold turkey. Start with a pay-as-you-go option to understand your actual daily usage patterns before committing to a subscription tier. The math only works in your favor if you're actually using the credits available. The Bigger Picture What's happening here reflects a broader shift in the AI model landscape. Chinese labs — Moonshot AI (Kimi), MiniMax, and Zhipu AI (GLM) — have spent the past year closing the gap with Western frontier models on coding benchmarks, and in some cases surpassing them. They've done it while pricing aggressively, sometimes at 5 to 17 times cheaper than equivalent Western models on a per-token basis. That's created an opening for API gateway services to bundle these models at prices the big labs simply can't match — at least not yet. The $10/month plan making waves right now may not be the flashiest tool in the shed. It doesn't have the brand recognition of Copilot or the marketing muscle of OpenAI. But on pure value-per-dollar for a developer who writes a lot of code, it's hard to argue with the math.

xNostr

Do Bitcoin Core Devs Lock Their Doors At Night?

Bitcoin's Spam Wars: A Satirical Take on a Very Real Debate Matthew Kratter of Bitcoin University recently published a video that, on its surface, asks a ridiculous question: do Bitcoin Core developers lock their doors at night? The answer, of course, is almost certainly yes — but that's precisely the point. The video is a sharp satirical parody, using home security as a mirror to reflect what Kratter sees as flawed reasoning from Bitcoin Core developers in a long-running technical and philosophical dispute about spam on the Bitcoin network. If you've been following Bitcoin closely, you'll recognize the debate immediately. If you haven't, buckle up — because it touches on some of the most fundamental questions about what Bitcoin is actually for . The OP_RETURN Debate, Explained Simply At the heart of Kratter's parody is a real controversy that has simmered — and sometimes boiled — within the Bitcoin developer community for well over a year. It centers on a technical feature called OP_RETURN , a small piece of code that allows users to embed arbitrary data directly into Bitcoin transactions. Originally introduced as a controlled way to attach metadata to the blockchain, OP_RETURN has increasingly been used to store things that have nothing to do with moving money: images, text, links, and data associated with projects like Ordinals and BRC-20 tokens . Critics call this spam — transactions that clog the network, inflate fees, and serve no monetary purpose. Defenders argue that if someone pays the transaction fee, their use of the network is legitimate by definition. The Bitcoin Core development team has generally leaned toward the latter position, and in 2025 made moves to expand the default OP_RETURN data limit rather than restrict it — a decision Kratter skewers with his joke about making the "designated theft basket" 1,250 times larger. The Satirical Argument, Unpacked Kratter's video works by taking the arguments he attributes to Bitcoin Core devs and translating them into home security advice: "Locking your front door won't stop all burglaries, so why bother?" → Mirroring the argument that spam filters won't eliminate all spam, so they're pointless. "Who are you to define crime?" → Reflecting the claim that "spam" is subjective, and one person's spam is another's legitimate transaction. "If you don't give burglars somewhere to rob, they'll just hit bigger targets." → A nod to the argument that if Bitcoin restricts certain data uses, those users will migrate to other chains, which would be worse for the ecosystem. "Locking your door is a slippery slope to gluing it shut." → Satirizing the concern that any restrictions on Bitcoin's use cases could lead to creeping censorship. The punchline, of course, is that no reasonable person would accept this logic when applied to their own home — yet Kratter argues it's being applied seriously to Bitcoin's architecture. Why This Debate Actually Matters Beneath the jokes lies a genuinely important question about Bitcoin's identity and long-term health. Bitcoin as a Monetary Network The original vision, outlined in Satoshi Nakamoto's 2008 white paper , was a peer-to-peer electronic cash system . Every byte of block space is finite. When that space is consumed by non-monetary data, it competes directly with ordinary financial transactions, potentially driving up fees for everyday users. The "Use-Case Agnostic" Counterargument Bitcoin Core's position is often framed around neutrality : the protocol shouldn't be in the business of deciding which transactions are worthy. If the fee is paid, the transaction is valid. Proponents of this view argue that trying to police content on the blockchain is both technically futile and philosophically dangerous — a form of censorship. The Fee Market Question There's also an economic dimension. Some Bitcoin supporters welcome increased transaction demand, even from non-monetary uses, because it generates fees that will eventually need to sustain the network's security budget as block rewards diminish over time. Critics counter that artificially inflated demand from spam is a poor and unreliable substitute for genuine monetary adoption. The Tone and the Takeaway It's worth noting that Kratter is careful — both in his video description and throughout the piece — to clarify that this is parody , not a personal attack. He explicitly wishes Bitcoin Core developers well and assumes they do, in fact, lock their doors. But parody has always been one of the sharpest tools in public debate, and what the video captures effectively is a sense of frustration felt by a segment of the Bitcoin community: that sophisticated-sounding technical arguments are being used to justify what they see as a straightforward failure to protect the network's core purpose. Whether you side with Kratter or with the Bitcoin Core position, the underlying tension is real and unresolved: What is Bitcoin for? Who gets to decide? And what happens when the answer to both questions is "no one in particular"? These aren't just technical questions. They're questions about governance, values, and the future of the most widely recognized cryptocurrency in the world. The spam debate may sound like an arcane developer squabble — but it's really a proxy war over Bitcoin's soul. This article is for informational and educational purposes only and does not constitute financial or investment advice.

xNostr

Nick tells meme coin traders who lost money on his token to get out of the trenches and go get a job...

dddd

xNostr