Meta Releases First Proprietary AI Model: Muse Spark
The model is intended for personal AI use, and the vendor admits it is limited in certain agentic capabilities and coding.
AI News
Groups Set Up to Shill AI and Data Centers Are Pouring Huge Sums of Money Into the Midterm Elections
Artificial intelligence remains deeply unpopular with the American public. One poll found it’s even more reviled than ICE, which is no small feat given the mass protests that erupt whenever the agency’s goons march into another US city.
A few political action groups are hoping to turn that around. Going into the 2026 midterm elections, the Financial Times reports, newly-formed PACs with major tech industry backing are spending hundreds of millions of dollars to shape how voters think about AI regulation.
Some of the groups cast a wide net, like Leading the Future, a super PAC backed by Trump donors and AI barons like OpenAI co-founder Greg Brockman, Palantir co-founder Joe Lonsdale, and tech venture capital giant Andreessen Horowitz. Founded in August of 2025, Leading the Future has raised over $125 million to back pro-AI candidates who oppose state-level regulations, according to the FT.
Others, like the pro-regulation PAC Public First Action, serve as vehicles for individual AI companies to push their agendas. Backed solely by Anthropic, this group aims to raise $75 million to boost candidates who want to preserve state’s individual rights to regulate AI. Mark Zuckerberg’s Meta also has its own pet super PAC, the American Technology Excellence Project, which aims to spend $65 million on state-level candidates who will “defend American tech leadership at home and abroad” — a fluffy way of saying “oppose AI regulation.”
This jockeying over states’ rights to regulate AI is the key question in the 2026 PAC wars. Though Republicans have largely staked their flag as the party of small government — which was always a selective attitude, to be fair — Donald Trump is now pushing for a major expansion of federal power. His latest AI framework seeks to concentrate regulatory authority over the tech at the executive level, which would effectively strip all 50 states of oversight power.
Bankrolling that push is Innovation Council Action, a hawkish super PAC backed by Trump advisor and PayPal mafioso David Sacks and led by former Trump communications aide Taylor Budowich. The newly formed group plans to spend at least $100 million supporting candidates who aren’t just pro-AI, but who will commit to carrying out Trump’s consolidation agenda, exclusively.
That PAC marks a major challenge to groups like Leading the Future, which Trump and his cabinet found to be insufficiently loyal.
“President Trump has made it clear, America will win the AI race against China, period,” Budowich told Fox. “He built the framework, he’s leading from the front, and this organization exists to make sure he doesn’t fight that battle alone. The cavalry is coming to back up the policymakers who stand with the president and will hold accountable the ones who don’t.”
More on AI and politics: Insiders Afraid the Government Will Nationalize the AI Industry
The post Groups Set Up to Shill AI and Data Centers Are Pouring Huge Sums of Money Into the Midterm Elections appeared first on Futurism.
Iran Threatens U.S. Tech Giants as Middle East Conflict Escalates
Breaking News · Middle East · Geopolitics
Iran Threatens U.S. Tech Giants as Middle East Conflict Escalates
Oracle’s Dubai tower takes debris strike. Iran’s Revolutionary Guard names Nvidia, Apple, Microsoft and Google as targets. A missing U.S. airman, two downed aircraft and a 48-hour ultimatum from Trump.
Location
Dubai Internet City — Oracle Building
Threats Intercepted (UAE)
Dozens in 24 hours
U.S. Aircraft Lost
F-15E downed · A-10 crashed (Kuwait)
Trump Ultimatum
48 hours · Hormuz Strait
01 —
Oracle Building Hit as American Corporate Sites Enter the Blast Zone
Iran launched a broad wave of missile and drone attacks across the Middle East on Saturday, marking a significant shift in the conflict’s geography. The UAE said it intercepted dozens of incoming projectiles in the 24 hours prior — and debris from one intercept struck the facade of the Oracle building in Dubai Internet City.
The Dubai Media Office confirmed no injuries and described the incident as minor. Damage was limited. But the symbolic weight was not: American corporate infrastructure in the Gulf is no longer sitting outside the blast zone.
Iran’s Revolutionary Guard simultaneously issued direct threats against a wider group of U.S. technology companies operating across the region — naming Nvidia, Apple, Microsoft and Google by name.
⚠ Iran’s Revolutionary Guard has directly threatened U.S. tech infrastructure in the Middle East, including Nvidia, Apple, Microsoft and Google.
02 —
Missing Airman, Two Downed Aircraft and Trump’s 48-Hour Warning
The U.S. military continued searching Saturday for a missing airman after an F-15E was shot down over southwestern Iran on Friday — the first U.S. combat aircraft successfully downed by Iranian forces since the conflict began in late February. One crew member was rescued. The second remained missing, with both U.S. and Iranian forces searching the same area.
In a separate incident, an A-10 Warthog pilot ejected after the aircraft was struck by Iranian fire over Kuwait. Two Black Hawk helicopters deployed in the search operation also came under fire inside Iranian airspace, though both returned safely. U.S. officials privately expressed concern the missing airman could be captured and used as political leverage by Tehran.
“Time is running out — 48 hours before all Hell will reign down on them.“
Donald Trump · Truth Social · 5 April 2026
President Trump posted on Truth Social on Saturday referencing his earlier ultimatum over the Strait of Hormuz, warning Iran it had 48 hours before consequences. The threat followed his earlier demand that Iran open the strait or make a deal within ten days.
03 —
India Resumes Iranian Crude as Bushehr Plant Takes Strike
India’s oil ministry confirmed its refiners had secured crude supplies including Iranian oil, after disruptions to Strait of Hormuz shipping lines cut into global supply. India had not received Iranian crude since May 2019, when U.S. pressure pushed buyers away from Tehran’s exports. The ministry also confirmed that 44,000 metric tons of Iranian liquefied petroleum gas had berthed at Mangalore this week aboard a sanctioned vessel.
The move signals a realignment in energy trade. The United States had temporarily removed sanctions on Iranian oil and refined products to reduce supply shortages — a decision now being tested by the ongoing strikes.
Near Bushehr, a projectile struck close to Iran’s nuclear power plant overnight, killing at least one worker and damaging part of the site. The International Atomic Energy Agency confirmed radiation levels remained normal but issued a warning against further strikes near nuclear facilities. Iran’s Foreign Minister said Tehran was not ready to rush into negotiations and would accept only a “conclusive and lasting” resolution to the war.
Russian state nuclear company Rosatom evacuated an additional 198 staff from the Bushehr site. It has been withdrawing workers since the conflict began at the end of February.
This article is based on reporting from Reuters, official statements from the Dubai Media Office, India’s oil ministry, the International Atomic Energy Agency, and Truth Social. Crypto Coin Show has not independently verified all claims made by parties to the conflict.
GPT-5.4 Pro Hits IQ 150 as AI Capability Becomes a Macro Variable
AI News · Institutional · Macro
GPT-5.4 Pro Hits IQ 150 as AI Capability Becomes a Macro Variable
OpenAI’s latest model scores higher than 99.96% of humans on a public IQ benchmark — a jump that is no longer just a lab milestone. With CPI, FOMC minutes and PPI all due this week, AI capability growth is beginning to behave like an economic signal.
150
GPT-5.4 Pro IQ Score
136
Previous record (o3)
99.96%
Humans outperformed
01 —
From 136 to 150: OpenAI Breaks Its Own Record
OpenAI’s GPT-5.4 Pro has reached an IQ score of 150 on TrackingAI’s public Mensa-style benchmark — a sharp step up from the 136 score its o3 model posted on the Mensa Norway test last year. A score of 150 sits in a range historically associated with figures like Albert Einstein and Richard Feynman, implying fast abstraction, strong pattern recognition, and the ability to navigate complex multi-step problems with limited guidance.
GPT-5.4 was introduced by OpenAI as its most capable and efficient frontier model for professional work, with improvements in coding, tool use, and computer use, and a context window of up to one million tokens. OpenAI also said GPT-5.4 achieved a new state of the art on GDPval and exceeded human performance on OSWorld-Verified — two separate benchmarks pointing in the same direction.
| Model | Developer | Test | IQ Score |
|---|---|---|---|
| GPT-5.4 Pro | OpenAI | TrackingAI / Mensa-style | |
| o3 | OpenAI | Mensa Norway | |
| Claude (latest) | Anthropic | TrackingAI public board | |
| Gemini | TrackingAI public board |
Why It Matters
A move from 136 to 150 compresses a complex capability shift into a single portable signal. For businesses, it feeds directly into decisions around automation, software budgets and headcount planning. For markets, it adds a variable alongside rates, inflation and growth expectations.
02 —
Public Benchmarks Have Limits — But the Curve Is Still Moving
IQ-style tests remain imperfect instruments for measuring frontier models. They compress a narrow slice of cognitive performance into a single number, obscuring variation across reasoning types, context handling, creativity and real-world problem-solving. Scores are sensitive to test design, training exposure, and pattern familiarity — making them a noisy proxy for general capability.
The methodology raises familiar questions: prompt structure, reproducibility, training-set contamination, and format familiarity. Those concerns were visible when o3 hit 136, and they remain active now.
Even so, the broader pattern has become harder to dismiss. One isolated benchmark result can be explained away. A cluster of gains across public IQ-style testing, coding, browser use, desktop navigation and knowledge-work performance carries more analytical weight.
“Investors do not need to accept every premise behind an IQ-style test to recognise that a jump of this size suggests acceleration rather than drift.“
CCS Analysis · April 2026
Enterprise buyers also do not need to believe IQ equals general intelligence to see that systems with stronger pattern recognition, stronger tool use and stronger long-horizon task handling are moving toward economically useful territory. This points toward systems that can search, plan, verify, navigate and produce real work across extended contexts.
03 —
AI Capability Is Beginning to Overlap with the Economic Week Ahead
The week ahead runs through macro. Markets are focused on FOMC minutes, CPI and PPI — all due within days. But beneath that surface, a second economic track is taking shape, and OpenAI sits near its centre.
Key Economic Releases — Week of 7 April 2026
Apr 8
FOMC Minutes — March 17–18 meeting. Markets parsing for policy tone on rates.
Apr 10
CPI — March — Consumer Price Index. Key read on whether inflation is cooling.
Apr 14
PPI — March — Producer Price Index. Leads CPI; watched for upstream inflation signals.
Capability growth in frontier AI increasingly intersects with capital allocation. A model that pushes higher on public reasoning tests while also improving in coding, search and computer use changes how businesses think about workflow redesign. It changes what enterprise buyers expect from copilots and agents. It changes how quickly organisations move from experimentation to deployment.
Jack Dorsey recently described Block moving “from hierarchy to intelligence,” using AI to take over coordination work once handled by management layers. That direction is becoming a commercial pattern, not an outlier.
The effects move through document workflows, spreadsheet workflows, customer support, research tasks, browser automation, internal operations, code generation and verification loops. The answer to where spending flows next extends beyond model subscription revenue into cloud demand, chips, data centres, networking, power and software licences.
Is the growth in intelligence itself beginning to behave like a macro variable? Faster capability gains can alter enterprise spending plans, tighten competitive pressure across white-collar functions, support higher infrastructure outlays and strengthen the case for AI-linked capital expenditure even in a slower nominal growth environment.
When TrackingAI shows GPT-5.4 Pro at 150, the number falls within a market that already views OpenAI as more than a lab — it is a platform company, a deployment company, an infrastructure customer and a signal generator for adjacent sectors. The score is compact, legible and easy to circulate. Its deeper relevance comes from the same place as the company’s broader product push: the frontier is still climbing, and the economic footprint of that climb is becoming harder to keep in a separate category.
This article draws on reporting from CryptoSlate, OpenAI’s GPT-5.4 launch materials, TrackingAI’s public leaderboard, and the U.S. Bureau of Labor Statistics economic calendar. Crypto Coin Show has not independently verified benchmark methodology claims.
Token Metrics Reinvents Itself as an AI Market Desk
Exclusive Interview · Blockchain Interviews
Token Metrics Reinvents Itself as an AI Market Desk
Ian Balina’s platform drops the analytics dashboard model in favour of a fully automated intelligence engine — one that reads 50+ data feeds, cross-references prediction markets in real time, and delivers its findings before most investors wake up.
There is a quiet but significant shift happening in how serious crypto investors consume information. The old model — charting platforms, raw on-chain dashboards, and the chaotic scroll of Crypto Twitter — is giving way to something more curated, more contextualised, and increasingly powered by artificial intelligence. Token Metrics, one of the longer-standing names in crypto analytics, has just made that transition official.
In a recent Blockchain Interviews conversation, Token Metrics founder and CEO Ian Balina laid out what amounts to a full company reinvention — one that reflects broader forces reshaping how information moves through financial markets in 2026.
The Pivot
From Analytics Platform to AI Market Desk
Token Metrics launched in 2019 at the bottom of a bear market, positioning itself as a research and analytics tool for crypto traders and investors. Over the years it helped its community identify early positions in projects like Polygon, Chainlink, and Helium Network — a track record that built a loyal subscriber base and, eventually, over 100,000 newsletter readers.
But late last year, Balina and his team made a strategic call that sets Token Metrics apart from most of its peers: they stopped trying to be a data API and started building what Balina describes as an “AI market desk.” The distinction is more than semantic. A data platform gives you access to information. A market desk synthesises it, weighs it, and tells you what it means for your portfolio.
We were building an API to give people that data. But what we realised is that’s not really our niche. Our niche is providing research, alpha, and insights to help people build portfolios in crypto.Ian Balina, Founder & CEO, Token Metrics
The pivot was accelerated by a market-wide behavioural shift: more and more crypto users are now pulling their information through large language models — ChatGPT, Gemini, Claude — rather than visiting data platforms directly. Rather than compete for that position or try to become the underlying data layer feeding those models, Token Metrics chose to go further up the stack, into judgment and synthesis.
Under the Hood
What the Engine Actually Does
The architecture behind Token Metrics’ new model is worth understanding in detail, because it addresses problems that have frustrated crypto investors for years. At the ingestion layer, the platform pulls from more than 50 data sources simultaneously — traditional crypto media like CoinDesk, on-chain analytics from Nansen and DeFi Llama, centralised and decentralised exchange data, and prediction markets led by Polymarket.
That data passes through a multi-stage AI pipeline staffed by specialised agents, each responsible for a different layer of quality control. The output is not a raw feed. It is a curated, editorially structured daily brief that identifies the five things most worth following on any given day, explains why they matter, and models the second-order portfolio effects of each.
How the Daily Brief Reaches You
- Free daily newsletter delivered before most subscribers wake up
- AI-generated podcast available on Spotify and Apple Podcasts
- Morning posts distributed to X and Discord simultaneously
- Real-time premium signals inside a private Discord community
- All outputs generated from one automated pipeline — no editorial lag between channels
Balina himself has switched from reading the newsletter to listening to the podcast each morning — a telling signal about how the platform’s own creator actually uses it. The system’s self-improving architecture means any errors are logged, learned from, and corrected in subsequent runs, closing a feedback loop that would take a human editorial team days to address.
Prediction Markets
The Polymarket Integration: Second-Order Thinking Built In
One of the most operationally interesting decisions Token Metrics has made is building Polymarket data directly into its daily intelligence layer — not as an optional add-on, but as a primary signal source alongside traditional news and on-chain data.
The use case Balina walked through was concrete: when a Federal Reserve rate decision is approaching, most news outlets report the outcome. Token Metrics’ AI goes further — querying Polymarket for the current probability-weighted odds, incorporating those into the analysis, and modelling two scenarios for subscribers: what happens to their portfolio if rates are cut, and what happens if they aren’t. This kind of second-order contextualisation was previously only available to institutional research desks or investors willing to manually work across multiple tools.
It’s able to add more colour to the news using verifiable data — you’re getting the whole picture, and then it tells you how each outcome will affect your portfolio.Ian Balina, Founder & CEO, Token Metrics
The platform applies a hard liquidity filter to all Polymarket data: any prediction market with less than $100,000 in liquidity is excluded from the analysis entirely. This prevents low-volume or easily manipulated markets from distorting the signal. “Any markets that are illiquid, it will toss out,” Balina confirmed. “We built in that QA control.” The integration extends to premium signals too — non-crypto prediction markets are cross-referenced against traditional bookmaker odds APIs, with divergences surfaced as opportunities where the spread justifies it.
Filtering the Noise
Signal Verification and the Smart Money Score
For premium subscribers, Token Metrics goes well beyond headline filtering. When evaluating new token launches — the area of the market most susceptible to manufactured hype and coordinated promotion campaigns — the platform runs a layered verification process designed to separate genuine momentum from engineered noise.
What the Alpha Score Checks
- Liquidity depth — tokens that can dry up rapidly are filtered first
- Honeypot and scam checks on every DEX-listed token before flagging
- Whale concentration — is the token heavily held by insiders positioned to exit?
- Smart money wallets — do early holders have a track record of winning calls?
- Polymarket cross-reference — is there a meaningful spread versus bookmaker consensus?
The core question the alpha score is designed to answer: is there genuine smart money in this token, or is it primarily held by insiders positioned to exit on retail buyers? Wallet-level analysis checks whether early holders have a track record of being positioned in successful projects before they broke out — a methodology that proved predictive in previous cycles and remains one of the most useful signals in a market full of coordinated promotion.
The Product Stack
Three Tiers, One Community Model
Token Metrics offers three plan structures built around different investor profiles. Each tier is additive — the deeper you go, the more research, community access, and multi-asset coverage you unlock.
| Plan | Core Offering | Best For |
|---|---|---|
| Signal | Real-time alerts for tokens and Polymarket opportunities that clear all filtering thresholds | Active traders who want a processed, qualified feed — not raw data |
| Alpha | Monthly playbook covering crypto, Mag 7 tech (Nvidia, Microsoft, FANG+), gold, and silver | Multi-asset investors who want macro context alongside crypto alpha |
| Round Table | Virtual community modelled on Tiger 21 — monthly sessions, portfolio stress-testing, conference coordination | Family offices, professional traders, and DeFi builders wanting peer-level intelligence sharing |
The Round Table tier is the most distinctive — modelled explicitly on Tiger 21, the well-known high-net-worth investor network where members stress-test portfolios in front of peers. Token Metrics’ version is primarily virtual, with in-person coordination available at major crypto conferences. The emphasis is on qualified participants: family office managers, active DeFi builders, professional traders — people who want structured peer feedback rather than another Discord server.
The Token Model
$TMAI: Access as Ownership
Token Metrics also operates a native token — $TMAI — tradable on decentralised exchanges and centralised platforms including Gate, MEXC, and Bitpanda. Holding $TMAI unlocks access to the platform’s premium Discord server and full content layer, functioning as an alternative to a monthly subscription.
The model is a clean expression of token-as-access design: demand for the platform translates directly into demand for the token, without requiring active governance participation or complex mechanics. It also creates alignment between community members and the platform’s long-term success that a pure subscription model cannot — hold the token and you benefit when the platform grows, not just when you use it.
The Bigger Picture
Why This Matters Now
The timing of Token Metrics’ pivot is not coincidental. The crypto intelligence market is at an inflection point driven by two converging forces. The first is the maturation of large language models. General-purpose AI assistants can now answer basic crypto questions — but they cannot source real-time on-chain data, cross-reference live prediction markets, or apply domain-specific scoring models to new token launches. There is a clear gap between what general AI can do and what a specialist platform with live data infrastructure can do. Token Metrics is positioning itself squarely in that gap.
The second force is the acceleration of the crypto market itself. Institutional adoption is increasing, on-chain activity is growing, and the number of active tokens, chains, and prediction markets has expanded to a point where manual tracking is genuinely impractical for most investors. Staying on top of the market without AI is no longer a choice — it is a necessity.
Humans could do it, but it would take a lot of time. Being able to create something fully automated — constantly watching 50+ data feeds and telling you only the things that actually matter — that’s the whole point.Ian Balina, Founder & CEO, Token Metrics
Whether Token Metrics executes on this at scale is a question only time will answer. But the architecture Balina described is meaningfully differentiated from both traditional analytics platforms and the general-purpose AI tools investors might otherwise default to. For a platform that has survived since the 2019 bear market bottom, that staying power alone is worth noting.
Ian Balina published a companion piece on the Token Metrics site this week covering where the platform is heading next. Read it on the Token Metrics site →
The free Token Metrics daily brief is available at tokenmetrics.com — no credit card required. The AI-generated podcast runs on Spotify and Apple Podcasts under Token Metrics Daily Pulse.
This feature is based on an exclusive interview conducted by CCS with Ian Balina, Founder & CEO at Token Metrics, on 17 March 2026.
Claude Mythos: The Model That Leaked Itself
Claude Mythos:
The Model That
Leaked Itself
Inside Anthropic’s accidental unveiling of Claude Mythos — the most powerful AI model ever built, held back because it might be too dangerous to release — and the market meltdown that followed.
In one of the more ironic data incidents in recent AI history, Anthropic — a company whose newest model is being held back specifically because of its unprecedented cybersecurity capabilities — accidentally exposed that model to the public through a basic configuration error on its own website.
The fallout was immediate: cybersecurity stocks cratered, crypto slid, and the broader tech sector sold off. But the real story is what the leak revealed about where AI is headed — and whether the industry is ready for a model its own creator calls too dangerous to release.
01 — The LeakHow 3,000 Files Went Public Overnight
On the evening of March 26, 2026, two independent cybersecurity researchers discovered something unusual sitting in a publicly accessible corner of Anthropic’s web infrastructure. Alexandre Pauwels of the University of Cambridge and Roy Paz, Senior AI Security Researcher at LayerX Security, had stumbled into what appeared to be a fully staged product launch — headings, body copy, images, PDFs, and a publication date — all sitting unprotected in an unencrypted, publicly searchable data store.
The disclosure was not a malicious breach. Digital assets — including images, PDF files, and audio files — were set to public by default upon upload, unless explicitly marked private. A toggle switch in Anthropic’s content management system was left in the wrong position, making approximately 3,000 assets linked to Anthropic’s blog publicly accessible. In total, there appeared to be close to 3,000 assets that had not previously been published to the company’s public-facing news or research sites that were nonetheless fully visible.
⚠ Configuration Error
The leak was not a cyberattack. Anthropic attributed it to “human error” in its content management system — a default-public setting that exposed staged assets before any editorial review. The same model the documents warned could trigger AI-powered cyberattacks was revealed by a misconfigured checkbox.
Among those exposed materials: a draft blog post detailing the existence, capabilities, and extraordinary risks of Anthropic’s next-generation AI model — a system called Claude Mythos.
02 — What Is Claude Mythos?A New Tier Above Opus
Anthropic currently markets its models in three tiers: Haiku (small and fast), Sonnet (balanced), and Opus (most capable). The leaked materials introduced a fourth category, internally named Capybara — larger and more intelligent than Opus, and significantly more expensive to run.
From the Leaked Draft
“Capybara is a new name for a new tier of model: larger and more intelligent than our Opus models — which were, until now, our most powerful.”
Two versions of the draft blog post surfaced — one calling the model “Mythos,” the other “Capybara” — suggesting Anthropic was still deciding between name candidates. The subtitle of the Capybara version still read: “We have finished training a new AI model: Claude Mythos.” Both versions explained the name was chosen to evoke “the deep connective tissue that links together knowledge and ideas.”
The draft described Claude Mythos as “by far the most powerful AI model we’ve ever developed.” When Fortune contacted Anthropic for comment, the company acknowledged the project without hesitation: “We’re developing a general purpose model with meaningful advances in reasoning, coding, and cybersecurity. Given the strength of its capabilities, we’re being deliberate about how we release it. We consider this model a step change and the most capable we’ve built to date.”
4th Tier
New “Capybara” model class above Opus
↑↑↑
Dramatic gains in coding, reasoning & cyber
Limited
Early access: cyber defenders only
$$$
“Very expensive to serve” — no general release
03 — CybersecurityThe Dimension That Stopped the Release
If Anthropic was simply excited about Claude Mythos’s capabilities, it would have launched quietly with the usual product post. Instead, the company is holding it back. The reason, spelled out in the leaked documents, is cybersecurity — and the concern is not hypothetical.
“Currently far ahead of any other AI model in cyber capabilities… it presages an upcoming wave of models that can exploit vulnerabilities in ways that far outpace the efforts of defenders.”— Anthropic Internal Draft Blog Post, March 2026
The core issue is what researchers call the dual-use dilemma. The model demonstrated an ability to surface previously unknown vulnerabilities in production codebases — a capability that could help defenders patch flaws just as easily as it helps attackers find and exploit them. The same model that could harden a bank’s infrastructure could be the tool that breaks into it.
⬡ Frontier Red Team Finding — February 2026
Anthropic’s own Frontier Red Team documented that Claude Opus 4.6 — already publicly available — discovered over 500 high-severity zero-day vulnerabilities in production open-source codebases. Some of these bugs had been present for decades despite expert review and millions of hours of accumulated fuzzer CPU time. One vulnerability required conceptual understanding of the LZW compression algorithm — a class of reasoning no fuzzer can replicate. That was the model before Mythos.
The already-public Opus 4.6 raised serious enough flags. Anthropic simultaneously confirmed that hacking groups, including those linked to the Chinese government, had attempted to exploit Claude in real-world cyberattacks. In one documented case, a Chinese state-sponsored group ran a coordinated campaign using Claude Code to infiltrate roughly 30 organizations — including tech companies, financial institutions, and government agencies — before the company detected it. AI handled an estimated 80–90% of the operation.
“In preparing to release Claude Capybara, we want to act with extra caution and understand the risks it poses — even beyond what we learn in our own testing. In particular, we want to understand the model’s potential near-term risks in the realm of cybersecurity — and share the results to help cyber defenders prepare.” — Anthropic Internal Document
For Mythos specifically, Anthropic’s Responsible Scaling Policy (RSP) looms large. The policy defines AI Safety Levels: ASL-3 was activated in May 2025 for models that “substantially increase the risk of catastrophic misuse.” ASL-4, not yet formally triggered, applies when AI models become “the primary source of national security risk in a major area such as cyberattacks or biological weapons.” Based on the language in the leaked drafts, Mythos may be approaching that threshold.
If Mythos truly represents a step change in cyber capabilities, patch cycles that once had days or weeks of breathing room could compress to hours. The cyber arms race, already accelerating, could become asymmetric in a new and dangerous direction — with attackers using AI to discover vulnerabilities faster than human defenders can respond.
04 — Market ImpactStocks, Crypto, and the Selloff
Within hours of Fortune’s reporting, markets reacted sharply. The concerns weren’t irrational: if an AI model can autonomously discover and exploit zero-day vulnerabilities at scale, it threatens the core value proposition of every legacy cybersecurity vendor whose defenses are built on known signatures and historical threat intelligence.
| Company | Ticker | Drop (Mar 27) | Sector |
|---|---|---|---|
| CrowdStrike | CRWD | −7.0% | Endpoint Security |
| Palo Alto Networks | PANW | −6.0% | Network Security |
| Zscaler | ZS | −4.5% | Cloud Security |
| Okta | OKTA | −3.0% | Identity & Access |
| SentinelOne | S | −3.0% | AI-Powered Security |
| Fortinet | FTNT | −3.0% | Network Security |
| iShares Tech-Software ETF | IGV | −3.0% | Broad Tech |
| Bitcoin | BTC | ↓ $66K | Crypto / Risk-Off |
Raymond James analyst Adam Tindle outlined several risks: compression of traditional defensive advantages, higher attack complexity and cost to defend, and potential shifts in security architecture and spending. Defensive approaches based on known signatures, vulnerability databases, or prior threat intelligence telemetry could be pressured as AI enables continuous discovery of novel attack surfaces.
“We read this as having the potential to become the ultimate hacking tool — one that can elevate any ordinary hacker into a nation-state adversary.”— Analyst Note, March 27, 2026
Paradoxically, however, not every analyst read the Mythos news as a pure negative for cybersecurity. The same analyst argued that announcements like this should continue elevating cybersecurity as a top IT priority, driving spend towards modernizing cyber defenses and away from legacy tools. If AI-powered attacks are coming, companies will need AI-powered defenses — and that is a significant market opportunity.
05 — The Deliberate RolloutDefenders Get a Head Start
Anthropic’s planned release strategy for Mythos was unlike anything the company had done before. Rather than a broad API launch, the rollout was designed to be defensive-first: early access would go exclusively to cybersecurity organizations, giving them a window to harden their systems before the model — or its inevitable imitators — reached bad actors.
Phase 1 — Now
Small cohort of early-access customers, focused on cybersecurity defense. Evaluating model behavior and risks in controlled environments.
Phase 2 — TBD
Expanded Claude API access, after efficiency improvements bring down cost. Model currently described as “very expensive to serve.”
Phase 3 — No Date Set
General release, contingent on safety evaluation outcomes and cost reduction milestones. Timeline entirely open-ended.
The irony of the situation is layered. Anthropic wanted to give cyber defenders a head start — to publish benchmarks showing Mythos’s offensive potential, allow security teams to study its techniques, and harden defenses accordingly. Instead, the existence of the model, its capabilities, and Anthropic’s own alarming internal risk assessment landed in public simultaneously, with no safety preparation, no controlled disclosure, and no ability to brief the security community in advance.
06 — The IronyThe Most Cybersecurity-Aware Leak in History
The Punchline
A company warning the world about AI-powered cyberattacks was undone by a checkbox. The model described as posing “unprecedented cybersecurity risks” was revealed through a default-public CMS configuration — not by a sophisticated nation-state, not by a zero-day exploit, but by a toggle left in the wrong position.
The tech community was quick to note the absurdity. In an enormously ironic twist, the draft blog obtained by Fortune — which was “available in an unsecured and publicly-searchable data store” — was the very document claiming the new model poses unprecedented cybersecurity risks. As one observer put it: let’s hope the new model wasn’t responsible for the security of Anthropic’s company blog.
It is a test for the company, which has received significant media attention for its Claude Code and Claude Cowork tools. The successes of those products have rattled competitors. The Mythos leak adds a different kind of pressure — one rooted not in capability rivalry but in questions about Anthropic’s own operational security posture at the exact moment it is asking the world to trust its judgment on frontier AI safety.
07 — What Comes NextIPOs, Lawsuits, and Open Questions
The Mythos story does not exist in a vacuum. Anthropic is simultaneously navigating a federal lawsuit against the Pentagon over a supply-chain risk designation rooted in its refusal to remove safety guardrails from Claude. Reports also surfaced this week that the company has been discussing a Q4 2026 IPO. The revelation that Anthropic is sitting on a model it considers potentially too dangerous for general release adds new dimensions to both of those threads.
For the Pentagon dispute, it provides ammunition to critics who argue Anthropic selectively applies its safety principles. For investors and IPO watchers, it raises questions about when and how Mythos revenue appears on a balance sheet — and whether regulators might weigh in before a general release is possible.
For the broader cybersecurity industry, the implications are structural. If Mythos — and the wave of similar models it is expected to inspire — can discover zero-day vulnerabilities faster than human analysts and patch systems can respond, the entire architecture of defensive cybersecurity shifts. Signature-based detection, historical threat intelligence, and human-speed incident response all face obsolescence. The question is whether the defenders, now armed with early access to Mythos itself, can adapt faster than the attackers who will inevitably get access to similar capabilities.
⬡ The Core Risk Equation
Anthropic’s own framing: give cyber defenders a head start with Mythos → they harden systems → when similar models reach bad actors, defenses are already stronger. The flaw in that plan: the deliberate head start was eliminated the moment the leak happened. Defenders and the public now know simultaneously.
For now, Mythos remains behind closed doors — tested quietly, released carefully, and very much on the world’s radar after a week nobody at Anthropic planned for. The name was chosen to evoke the deep connective tissue that links knowledge and ideas. Instead, it became the story of how a single misconfigured setting linked a company’s most sensitive secrets to anyone with a search engine.
FIL trades near all time low despite the launch of Filecoin Onchain Cloud
Filecoin has announced on X that its Onchain Cloud is live on the mainnet. The new service is designed to provide a programmable storage and payments layer for developers.
But despite the launch, Filecoin’s token is trading at $0.83, near its all-time low of $0.81 recorded last month.
Filecoin Onchain Cloud built for AI agents and autonomous systems
AI agents are a new class of cloud users with autonomous systems that need to store, retrieve, and pay for data without relying on humans. Filecoin Onchain Cloud is built for AI agents. It works as a programmable storage and payments layer.
According to the announcement, over 100 teams started building tools for AI agents, dApps, workflows, and dataset indexing. The teams joined when Filecoin Onchain Cloud was on testnet last November. Mainnet data shows 49.41 Tebibytes stored in 478 active datasets, with 81 payer wallets linked onchain through Filecoin Pay.
The mainnet launch brought in new updates, including two-copy replication, an updated Synapse SDK, and production-grade storage providers.
The two-copy replication feature means that every upload using the Synapse SDK lands on two independent storage providers. The first provider stores the data, while the second provider pulls directly from the first, without utilizing extra bandwidth. Each copy generates its own onchain proof. If one provider fails, the system replaces it automatically.
The Synapse SDK uses viem instead of Ethers, improving TypeScript support and speed. Uploads are mirrored across providers automatically, without affecting performance.
Production-grade storage providers have been onboarded, tested, and approved for mainnet. They must meet specific performance criteria, including a storage success rate of 95%+, a PDP fault rate below 1%, and a retrieval success rate above 95%. If a provider fails to meet these thresholds, they are removed from the network.
The Proof of Data Possession (PDP) explorer is available and lets anyone check proof status, provider performance, and fault history.
Filecoin’s FIL token tanks to $0.83
Despite the positive news, Filecoin’s native token, FIL, has seen a steep slide in price. It is currently trading at $0.83, which is close to its all-time low of $0.81 recorded last month.
FIL dropped 4% in the last 24 hours, 8% in the last week, and 22.2% in the 30 days. CoinGecko data shows that FIL has lost 99.6% of its value from its all-time high of $236.84 in April 2021.
FIL maintains its position among the top 100 cryptocurrencies. The decentralized storage coin ranks 83rd in terms of market capitalization, with a market cap of $638,926,958 million.
Filecoin is facing a decline in terms of total value locked (TVL) as well. DeFiLlama data shows that Filecoin’s TVL slumped by 9.88% in the past 24 hours. It’s currently standing at $6.31 million. The decline in FIL’s price and TVL suggests that the market is not responding positively to the Onchain Cloud launch.
The collected fees by the network is only $1,950 in 24 hours. Filecoin struggles to remain profitable. The amount of stablecoins circulating on the Filecoin network is very low at around $153,629. The full supply of stablecoins is in the form of USDFC.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It’s free.
Protestors Outside Anthropic Warn of AI That Keeps Improving Itself
Months after a daring hunger strike failed to pause development of Anthropic’s AI Claude, protestors have rallied around the company’s headquarters to call for a complete stop to AI development.
Last weekend, nearly 200 protestors with the organization Stop the AI Race demonstrated in front of Anthropic, demanding the company’s CEO, Dario Amodei, publicly commit to pausing their development of AI. According to FirstPost, protestors included former tech industry workers, researchers, and members of other grassroots organizations like Pause AI and QuitGPT.
“The reason we are pausing AI is because we believe that building AI that can automate AI research, and that can self improve, could be a danger to the human race, especially human extinction,” Michaël Trazzi, an organizer with Stop the AI Race, told local reporters. “It’s not only me and other researchers saying this, it’s the lab CEOs themselves that [say] the risk is real.”
Stop the AI Race rallied around the company’s San Francisco headquarters for a while before marching on Sam Altman’s OpenAI and Elon Musk’s xAI, where they made similar demands.
In a post on social media, Trazzi claimed that it was “the biggest AI safety protest in US history” so far.
One of the protestors involved, Guido Reichstadter, had previously protested outside Anthropic in the aforementioned hunger strike, which ultimately lasted for 30 days. Like Trazzi, Reichstadter’s concerns are existential — an AI system that could one day break containment and usher in unknown horrors on humankind.
On day nine of his hunger strike, Reichstadter told Futurism that frontier AI systems are an “entirely new class of danger.” Indeed, whether Claude is going to take over and start killing us all may be beside the point: in the hands of humans, it’s already picking strike targets for the US military.
“None of these companies have a right to do what they’re doing, which is consciously endangering my life, my family’s life, all of our lives,” Reichstader said. “The correct thing for them to do is stop the global race toward really dangerous AI that we’re all involved in.”
More on Anthropic: Pentagon Refuses to Say If AI Was Used to Select Elementary School as Bombing Target
The post Protestors Outside Anthropic Warn of AI That Keeps Improving Itself appeared first on Futurism.
io.net Launches World-First Platform Allowing AI Agents to Purchase Their Own Computing Power
io.net
io.net Launches World-First Platform Allowing AI Agents to Purchase Their Own Computing Power
Agent Compute lets autonomous systems provision GPU resources without human intervention — no enterprise contracts, no procurement delays, no hyperscaler pricing.
10K+
GPUs Available
138
Global Regions
70%
Cost Reduction
io.net, the world’s largest decentralised GPU network, today launched Agent Compute — a first-of-its-kind platform that lets AI agents provision their own computing infrastructure. Agents can independently spin up GPU clusters, run workloads, and scale resources up or down, without the enterprise onboarding and procurement processes that have kept cloud computing out of reach for smaller teams.
The current cloud model is built for enterprise budgets. If you’re a startup or solo developer building with AI agents, you’re either paying prohibitive hyperscaler rates or spending weeks navigating procurement. Agent Compute removes that barrier. An agent can independently find the best-priced GPU for the job, provision it, and manage the infrastructure end-to-end — so developers can spend their time building, not comparing cloud pricing or configuring servers.
— Gaurav Sharma, CEO, io.net
AWS and Google Cloud require lengthy onboarding, complex billing, and minimum commitments that price out small teams, startups and developers. Agent Compute sidesteps that entirely. Agents interact directly with io.net’s marketplace of over 10,000 GPUs across 138 regions in 130+ countries — accessing compute at up to 70% lower cost than traditional cloud providers, provisioning resources on demand and releasing them when done.
MCP Protocol Integration
The system uses the Model Context Protocol (MCP), which gives agents clear visibility into available compute resources — GPUs, costs, and specs — to make informed decisions rather than costly mistakes. Without these guardrails, AI agents have already caused real damage: Amazon’s AI shopping agent triggered a 13-hour outage after deleting a production environment, and OpenClaw users reported bills exceeding $3,600 a month from runaway workflows.
In practice, an agent analysing data can spin up a GPU cluster, process the job, and terminate resources when finished. No manual setup. No leftover costs. Developers can set spending limits and resource caps to stay in control.
Vision
This is a step towards truly autonomous agents. Right now, agents still depend on humans for infrastructure. As they become more capable, that dependency becomes the bottleneck. If agents are going to operate independently — making decisions, executing tasks, scaling resources in real time — they need the ability to provision their own compute. That’s the future we’re building toward.
— Gaurav Sharma, CEO, io.net
The implications go beyond developer convenience. As agents gain the ability to manage their own compute requirements, the bottleneck shifts from infrastructure access to imagination. A solo developer can now build systems that were previously only possible with enterprise resources — agents that scale across continents, process massive datasets, or deploy models on demand. It’s a glimpse of a future where the limiting factor isn’t budget or team size, but what you can think to build.
Availability
Agent Compute is available now in early access, with broader rollout to follow. Visit io.net to apply for access.
Strategy’s Bitcoin Holdings Cross 760,000 BTC, AI Reveals How Long Till It Gets To The 1 Million Mark
Editorial Summary
Strategy’s Bitcoin holdings have crossed the 760,000 BTC threshold, marking a significant milestone in what has become one of the most aggressive corporate accumulation strategies in digital assets. Artificial intelligence models now project the company could reach 1 million BTC as early as September 2026, though realistic timelines extend into late 2026 or early 2027, depending on capital availability and market conditions. For institutional investors, this trajectory signals both the viability of Bitcoin as a corporate treasury asset and the emerging role of AI in forecasting long-term acquisition patterns in volatile markets.
Strategy, formerly known as MicroStrategy, has announced its holdings have surpassed 761,068 BTC as of March 16, 2026, following a record weekly purchase of 22,337 BTC valued at approximately $1.57 billion. The company’s relentless pursuit of Bitcoin accumulation, spearheaded by CEO Michael Saylor, continues despite macroeconomic headwinds and cryptocurrency market volatility. This latest acquisition represents the acceleration of a multi-year strategy that has transformed the enterprise software company into one of the world’s largest corporate holders of Bitcoin, rivaling sovereign wealth allocations. As Strategy approaches the symbolic 1 million BTC milestone—representing roughly 4.76% of all Bitcoin ever to be mined—the financial and strategic implications have captured the attention of institutional investors, asset managers, and cryptocurrency market analysts worldwide.
AI Models Project Divergent Timelines for Million-Bitcoin Target
Two leading artificial intelligence platforms have published competing forecasts regarding Strategy’s path to 1 million BTC, each arriving at different conclusions based on distinct methodological approaches. Grok AI, developed by xAI and SpaceX founder Elon Musk, suggests the most optimistic scenario, projecting that Strategy could mathematically reach the 1 million BTC milestone as early as September 2026 if current acquisition velocity persists. This forecast is grounded in Strategy’s recent purchasing patterns, which have demonstrated significant acceleration over the preceding three weeks, during which the company acquired between 3,015 and 22,337 BTC weekly, averaging approximately 14,450 BTC per week. At such a pace, maintaining consistent capital deployment would theoretically close the remaining 238,932 BTC gap within roughly five to six months.
However, Grok AI’s analysis acknowledges a critical constraint: sustaining weekly acquisitions of this magnitude would require continuous capital raises exceeding $1 billion per week—a requirement the platform notes is unrealistic under current market and financing conditions. Consequently, Grok’s more sustainable projection, accounting for historical averages of approximately 2,500 BTC per week through Strategy’s STRC preferred stock funding program, suggests a more credible target date of September 2026. This revised timeline incorporates material factors including market liquidity constraints, capital raising limitations, equity dilution concerns, price volatility, and execution risk. The distinction between theoretical and practical timelines highlights the complexity of translating aggressive corporate strategies into achievable milestones within real-world constraints.
ChatGPT’s analysis presents a more conservative framework, suggesting that Strategy would need to acquire approximately 5,550 BTC weekly to reach 1 million BTC by December 2026—a rate roughly 50 to 100 percent above recent weekly averages. While the AI model acknowledges this goal is ambitious, it suggests late December 2026 could represent an achievable target if Strategy substantially escalates purchases through coordinated equity issuance and accelerated STRC funding. Yet ChatGPT’s base case forecast points toward early January 2027 as a more probable arrival date, recognizing that market liquidity constraints, price volatility, and uneven acquisition patterns across weeks create practical headwinds. Both AI models converge on a timeframe spanning late 2026 through early 2027, suggesting institutional consensus that the 1 million BTC target remains achievable within an 12-to-18 month window from the March 2026 baseline.
Capital Deployment Strategy and Funding Mechanisms
Strategy’s ability to reach 1 million BTC depends critically on its capacity to raise and deploy capital at scale without triggering unacceptable equity dilution or market disruption. The company currently employs two primary funding mechanisms: direct equity issuance and its proprietary STRC (Strategy Tracker Convertible) preferred stock program, which allows the company to raise capital while offering investors conversion rights tied to Bitcoin holdings. The STRC program has proven particularly valuable, as it enables capital raising while preserving voting control and avoiding the traditional equity dilution associated with common stock offerings. Recent weekly acquisitions averaging 14,450 BTC represent deployment rates of $900 million to $1.57 billion, indicating access to substantial capital flows despite macroeconomic volatility.
The sustainability of these acquisition rates hinges on multiple factors beyond capital availability. Bitcoin market liquidity, measured across major exchanges and over-the-counter desks, must accommodate Strategy’s block purchases without triggering significant price slippage or market disruption. As Strategy’s position approaches 5 percent of circulating Bitcoin supply, each successive acquisition becomes operationally more complex and potentially moves markets more noticeably. Additionally, the company must navigate regulatory scrutiny in jurisdictions where large Bitcoin concentrations may trigger disclosure requirements or compliance obligations. Strategy’s historical pattern of acquiring Bitcoin in tranches—rather than attempting to accumulate continuously—suggests management recognizes these practical constraints and tailors acquisition timing to optimize execution.
The financing model also requires sustained access to capital markets during periods of cryptocurrency volatility and regulatory uncertainty. Strategy’s ability to issue STRC securities or conduct equity offerings depends partly on investor appetite for Bitcoin-exposed equity instruments and confidence in the company’s strategic direction. Should broader macroeconomic conditions deteriorate significantly or should Bitcoin experience sustained price weakness, Strategy’s cost of capital could rise substantially, making the aggressive accumulation strategy less economically rational. Conversely, Bitcoin price appreciation would reduce the quantity of capital required to reach 1 million BTC, potentially accelerating timelines. This dynamic creates a feedback loop where Bitcoin’s price discovery mechanism directly influences the feasibility of Strategy’s milestone projections.
Institutional Implications and Market Significance
Strategy’s pursuit of 1 million BTC carries profound implications for institutional asset allocation, corporate treasury management, and Bitcoin’s acceptance as a legitimate institutional store of value. The company’s publicly stated strategy—to accumulate Bitcoin as its primary treasury asset and hold indefinitely—has inspired peer scrutiny and competitive positioning among other corporations and institutional investors. Should Strategy successfully achieve and maintain 1 million BTC holdings, representing approximately 4.76% of all Bitcoin that will ever exist, the company would effectively establish itself as one of the world’s largest non-sovereign holders of a major digital asset. This concentration raises intriguing questions about corporate governance, fiduciary responsibility, and the concentration risk embedded in Bitcoin’s distributed ledger system.
For institutional investors evaluating cryptocurrency exposure, Strategy’s execution demonstrates that large-scale, sustained Bitcoin accumulation is operationally feasible despite real-world constraints and market volatility. The company has successfully acquired over 760,000 BTC through multiple market cycles, price rallies, and corrections, suggesting that disciplined, long-horizon investment strategies can accumulate meaningful positions without requiring perfectly timed entry points. This reality may reduce perceived barriers to institutional Bitcoin adoption and encourage other corporations to evaluate similar treasury strategies. Furthermore, Strategy’s continued success in raising capital for Bitcoin purchases—even during periods of cryptocurrency market skepticism—suggests investor confidence in both the company’s strategic vision and Bitcoin’s long-term value proposition.
The September 2026 to January 2027 timeline projected by AI models also carries symbolic significance. Achievement of 1 million BTC would represent a corporate commitment to digital assets of unprecedented scale and permanence, occurring roughly 17-18 years after Bitcoin’s 2009 genesis block. This timing aligns with Bitcoin’s maturation as an asset class and the growing normalization of cryptocurrency within institutional portfolios. Should Strategy reach this milestone, the financial services industry will likely experience accelerated institutional adoption, potentially triggering cascading demand from pension funds, endowments, and other large asset holders seeking comparable exposure. Conversely, any failure or significant delay in reaching 1 million BTC could signal limitations in corporate Bitcoin accumulation strategies and dampen emerging institutional enthusiasm for large-scale digital asset concentration.