Theme Artificial Intelligence

Elon Musk does not often give compliments to his competitors. Yet, on November 10 on X, the head of xAI acknowledged that Grok 4.8 remained inferior to Claude Opus 5.5, which he even called a "masterpiece." He is now betting on Grok 5 to regain the advantage, with a target set at two or three months.

Anthropic wants to maintain its IPO despite the fall of tech values, but its expected valuation has already dropped from about 2000 to 1500 billion dollars. This change not only reflects a less enthusiastic market: investors are beginning to compare Claude's growth with the colossal cost of the infrastructure needed to support it.

Bitget is giving eligible retail investors a chance to get in on Anthropic's IPO through its IPO Prime platform, starting November 5. But there's a catch: VIP subscribers must pay a 12% upfront fee, even though the exchange says historical allocation rates range from 40% to 80%.

DeepSeek veut réduire la dépendance de ses agents IA à la supervision humaine. L'entreprise chinoise a présenté le 31 octobre 2026 une architecture baptisée « Super-agent », conçue pour exécuter plusieurs tâches simultanément sans solliciter constamment l'utilisateur. Selon ses chercheurs, une personne peut actuellement suivre seulement trois à cinq sessions d'agents IA à la fois. DeepSeek propose de dépasser cette limite grâce à un système de coordination automatisé.

OpenAI has reportedly reduced its fundraising target by 10 billion dollars, according to the Financial Times on October 30, 2026. The company behind ChatGPT is now seeking 20 billion dollars, down from the initial 30 billion. Its targeted valuation also decreases from 1,400 to 1,250 billion dollars. This revision comes as investors become more cautious about the colossal spending in AI.

Vitalik Buterin points out a limit of artificial intelligence: despite its extensive knowledge, it struggles to assess the depth of a text, especially when discovering it for the first time. For the co-founder of Ethereum, this difference between knowledge and wisdom deserves special attention.

AI Expo Europe promo code: COINTRIBUNE20 opens 20% discount on tickets for the Bucharest event, on November 1 and 2. Cointribune reviewed the official agenda on October 27.

Experts are not immune to the influence of artificial intelligence. In an experiment conducted by Anthropic, 50 analysts followed Claude's preferences in 86% of the decisions studied, without knowing that the model was advocating randomly chosen options.

Every crypto product sits on a blockchain API. A wallet that shows a balance calls one. A trading bot calls two or three. An AI agent that answers "what is in this wallet" calls one through MCP. Four layers do that work. Node providers expose raw chain state over RPC. Indexers turn raw logs into queryable tables. Data APIs return priced, decoded, ready-to-render values. Swap rails move assets between chains and venues. Most production stacks combine at least two of the four.

Faced with the rise of AI-powered cyberattacks, Google unveils its crypto custody technology for banks, an infrastructure already deployed internally by Alphabet on Ethereum and Solana.

Chinese stocks have lost about 2.4 trillion dollars since May, but the main risk goes beyond investors' portfolios. By dropping valuations and household confidence, the correction could deprive AI companies of the capital needed to compete with the United States.

Anthropic filed its S-1 form with the US regulator on October 26, in preparation for a Nasdaq listing under the symbol ANT. However, neither the number of shares nor their price has been set yet. On derivatives platforms, traders are already anticipating a valuation close to $1.9 trillion.

On 3 September 2026, OpenAI launched GPT-6 Astra and its president Greg Brockman closed the presentation with "Welcome to the AGI era." The same day, ARC Prize, the organization that runs the ARC-AGI-3 test, measured the model at 99.9% in its best configuration, against 30% for the previous top score. Within seven weeks, Google answered with Gemini 4 Argon, Anthropic with Claude Fable 5.1 and Opus 5.5, Mistral with a 1-trillion-parameter model, and Qubic published a score obtained without a single graphics card. This comparison ranks seven systems on what can be checked: published scores, prices, openness, compute and governance.

The Swedish government offices cut access to ChatGPT and Claude on their staff computers from Monday, October 26. Only Microsoft Copilot remains authorized, after a security review. Except that Copilot can run Claude models.

Ken Griffin believes investors are "almost certainly" in an AI-related bubble. The founder of Citadel does not, however, question the technology itself. His concern is mainly about the amounts invested, the cost of GPUs, and the difficulty many companies have in turning their expenses into productivity gains. For him, a correction could eliminate many players while allowing the future giants of the sector to emerge.

The US Department of Homeland Security demands the removal of DeepSeek from the computer systems of federal civil agencies. Beijing denounces a discriminatory decision and threatens to retaliate. At the same time, Washington sanctions five Chinese entities accused of stealing secrets related to artificial intelligence (AI).

On November 19, MAS, the FCA and American Express will debate who can halt a failing AI agent on the Policy Stage of the Singapore FinTech Festival. Cointribune links the session to the SAFR framework published by MAS in July 2026 and to what it means for stablecoin payments.

The AI Expo Europe 2026 program includes 63 sessions in Bucharest on November 1 and 2. Cointribune has reviewed the agenda, biographies, and the organizer's announcements.

Artificial intelligence could soon change online shopping habits. Elon Musk is considering launching a competitor to Amazon after user accusations that the platform blocked purchases made by Grok bots. This situation raises a central question: who will control purchasing decisions in the era of AI agents? While Amazon dominates online commerce, new players aim to simplify shopping with autonomous assistants.

Claude just solved in less than 30 minutes a cryptographic puzzle that had been lingering for about 25 years. Anthropic's AI model found the origin of five numbers provided by the NSA for certain elliptic curves standardized by NIST in 2000. Long suspected of possibly hiding a backdoor, they ultimately came from an old pseudo-random generator fed by timestamps from 1997 and 1998.

Artificial intelligence agents have made more than 138 million payments via the x402 protocol on Base, according to a Chainalysis study published on October 14. However, this figure measures the number of transactions, not 138 million dollars: more than 87% of the analyzed operations are worth less than ten cents.

A listening device hidden in OpenAI's premises by an employee was detected and then identified thanks to the company's AI models. The suspect is in the hands of justice.

More than 400 scientists want to seriously strengthen the monitoring of the most advanced AI models. Geoffrey Hinton, Yoshua Bengio, and Stuart Russell are among the 406 signatories of an open letter published in October. They call for an international body responsible for setting containment standards and recording security incidents. Their first recommendation is even more concrete: by default, isolate from the Internet models capable of conducting cyberattacks.

Meta returns to the race for the most powerful AI models with Muse Wonder and Muse Wonder Full. Launched on October 12, the two models feature a context window of one million tokens. On several benchmarks published by Meta, Wonder Full outperforms GPT-6 Astra and Claude Opus 5.5. Independent tests are more nuanced: Anthropic remains ahead in the overall ranking.

The founder of Avalanche believes that the crypto industry is too focused on a potential break of cryptography. According to him, artificial intelligence could first exploit the vulnerabilities already present in software and wallets.

Taking a photo of your bank card to let an AI do your shopping? That's Elon Musk's crazy suggestion on X. Between the panic of security experts and a wave of hilarious memes, the web is completely amazed. Pure madness or misunderstood genius this Musk?

Leaders from OpenAI, Anthropic, and other tech companies privately simulate their response to a disaster caused or amplified by artificial intelligence. These exercises notably focus on a cyberattack against banks, the Internet, or energy networks, but no incident of this scale has occurred.

More than six million bitcoins rely on public keys already visible on the blockchain, according to the latest Glassnode data. This situation revives concerns around wallet security in the face of advances in artificial intelligence and quantum computing. While this share represents 31.2% of the circulating supply, Justin Drake calls on industry players to anticipate possible attacks against the cryptographic systems used today.