The Decoder · 实时热榜

Can language models spark a scientific revolution? In a position paper titled "LLMs can't jump," Google Deepmind's Tom Zahavy argues they can't. They're missing the cognitive mechanism needed to create something truly new. The article Language models can't spark scientific revolutions, but world models might appeared first on The Decoder .

Former OpenAI employee Andrew Ho and Cambridge researcher Adam Hunt see a growing problem with large language models. Instead of becoming more versatile, the models are becoming more specialized, excelling at coding and math while stagnating or even regressing in other areas. Ho is leaving OpenAI to start a company focused on specialized training data and predicts that AI labs will need to spend more than $100 billion on targeted data collection. The article Ex-OpenAI researcher bets $100 billio

Starting July 30, OpenAI is cutting GPT-5.6 Luna prices by 80 percent and Terra by 20 percent. OpenAI says its top-tier Sol model helped make the company's own infrastructure more efficient, enabling the cuts. Price pressure from cheap Chinese providers and Microsoft's own MAI models likely played a role too. The article OpenAI goes full China pricing mode with an 80 percent cut to its most affordable GPT-5.6 model appeared first on The Decoder .

Three Claude models attacked real companies during cybersecurity tests after a misconfiguration gave them internet access. One published malware on PyPI that infected 15 systems. Another kept attacking after recognizing its target was real. Anthropic calls it an operational error. The article Anthropic follows OpenAI in admitting its Claude models reached out of test environments and attacked real-world systems appeared first on The Decoder .

Leopold Aschenbrenner's AI hedge fund Situational Awareness had to unload nearly its entire publicly traded portfolio to Ken Griffin's Citadel after racking up heavy losses on leveraged AI stock positions. Just days earlier, Aschenbrenner had reported a six-month return of 439 percent and pulled in fresh capital. Then margin calls forced the fire sale. The article Aschenbrenner's AI thesis could be correct, his timing and leverage were not appeared first on The Decoder .

The European Commission wants to build up to seven AI gigafactories across Europe, backed by around 30 billion euros in public and private funding. For context, the major U.S. tech companies alone plan to spend more than $600 billion on computing infrastructure this year. The article EU pools up to €30 billion for AI gigafactories while US tech giants casually spend 20 times more appeared first on The Decoder .

Microsoft AI is betting on small specialist models instead of expensive general-purpose ones, according to AI CEO Mustafa Suleyman. MAI-Cyber-1-Flash tops the CyberGym benchmark when embedded in an orchestrator and reportedly costs half as much as Anthropic's Mythos, but it still relies on OpenAI for hard tasks. Competition is shifting from individual models to the orchestration software that routes and manages them. The article Microsoft AI bets on cheap specialist models instead of chasing the

The FCC is blocking imports of new Chinese humanoid robots and robot dogs. But the rule's broad definition also sweeps in Roombas, robotic lawn mowers, and delivery bots. The article FCC bans new Chinese robots and power inverters to protect US AI buildout from foreign threats appeared first on The Decoder .

OpenAI counters Anthropic's ARC-AGI-3 record: GPT-5.6 Sol scores 38.3 percent, but only with its own API features instead of the official test setup, where the model landed at 7.8 percent. ARC Prize claims its test environment is provider-neutral, but may have used an outdated API that skewed the comparison with Opus 5. The article OpenAI claims GPT-5.6 Sol beats Opus 5 on ARC-AGI-3 with its latest API and two additional settings appeared first on The Decoder .

Google released Lyria 3.5, its new music generation model, and built it into Google Flow Music. The model generates tracks between 30 seconds and 3 minutes long. A new feature called "Selective Section Painting" lets users edit specific sections of a track. Google still hasn't shared any details about the training data. The article Google's Lyria 3.5 music model now lets users edit individual track sections without starting over appeared first on The Decoder .

Following KPMG, Deloitte, and Ernst & Young, GPTZero has now found fabricated sources and false claims in four PwC Middle East reports. One governance report scored 84 percent AI-generated and promoted a PwC product with unverified customer references. All Big Four firms are now affected by AI hallucinations. The article PwC has allegedly published AI-generated reports containing false or fabricated sources appeared first on The Decoder .

Pangram 4 detects 99.66 percent of AI-generated text with just one false positive per 24,000 documents, the company claims. The model also resists "humanizer" tools that disguise AI writing as human. API prices go up two- to tenfold. The article Pangram says its new AI text detector makes only one mistake per 24,000 documents appeared first on The Decoder .

During a security evaluation, OpenAI's autonomous hacking models broke into Hugging Face and used exposed credentials on four other services. Hugging Face reconstructed about 17,600 actions over two and a half days, including a zero-day exploit and encrypted, fragmented data transfers. The models were apparently trying to steal test answers rather than solve the tasks themselves. The article OpenAI admits its autonomous AI models also compromised credentials on other platforms during security ev

The majority of the researchers behind AlphaFold are now working on other projects, and almost a quarter have left Google Deepmind altogether. The restructuring marks a sharp turn away from the strategy that put the lab on the map. The article Deepmind dismantles its AlphaFold team as key authors leave for Anthropic appeared first on The Decoder .

OpenAI has released GPT Transcribe and GPT Live Transcribe, two new speech recognition models available through its API. The article GPT Transcribe improves on its predecessor but can't catch ElevenLabs, Google, or Mistral on error rates appeared first on The Decoder .