epoch.ai
PulseAugur coverage of epoch.ai — every cluster mentioning epoch.ai across labs, papers, and developer communities, ranked by signal.
- instance of AI chips 90%
- developed MirrorCode 90%
- competes with GPT-4 70%
- instance of MirrorCode 70%
- used by MirrorCode 70%
- affiliated with Less Wrong 60%
- affiliated with GPT-5 60%
- instance of GLM-5.2 50%
- competes with GPT-3.5 50%
- affiliated with Hacker News 50%
- invested by Andreessen Horowitz 50%
- affiliated with Redwood Research 50%
12 day(s) with sentiment data
-
Anthropic's Claude AI now leads 26% of company's R&D efforts
Anthropic has reported that its AI model, Claude, now leads 26% of the company's AI research and development efforts. This figure represents a significant increase from under 1% in February, indicating a substantial shi…
-
Together AI outlines strategy for migrating to open-source models
Together AI's blog post outlines a strategy for migrating from closed-source to open-source AI models, emphasizing that such migrations can be faster and less complex than traditional ones, especially when utilizing man…
-
Open-weight AI models now rival frontier providers, shifting focus to deployment infrastructure
The landscape of open-weight AI models has significantly advanced, with top-tier open models now closely rivaling frontier closed models in performance, according to Epoch AI. This development suggests that the primary …
-
AGI arrival seen as gradual emergence, not single event · 1 source tracked
A review of 59 podcast conversations reveals that the arrival of Artificial General Intelligence (AGI) is not expected to be a singular event, but rather a gradual emergence of capabilities and societal impacts. Experts…
-
Proposal suggests 5M token limit for AI context windows to slow growth and enable monitoring
A proposal suggests implementing a temporary, legally enforced maximum of 5 million tokens for AI context windows. This measure aims to slow AI capability growth and ensure that AIs generate external memory artifacts, s…
-
Epoch AI: Claude shows lower latency than GPT with long contexts
Epoch AI conducted a comparison of long-context model latency, revealing notable differences between GPT and Claude models. The findings indicate that Claude generally exhibits faster response times compared to GPT when…
-
GPT-6 Astra cracks final FrontierMath Tier 4 math problem · 1 source tracked
GPT-6 Astra has successfully solved the final remaining problem in the FrontierMath Tier 4 benchmark, a set of research-level mathematical problems designed to challenge advanced AI models. This achievement marks a sign…
-
GPU compute leasing costs cut by optimizing data transfer efficiency
This article discusses optimizing data transfer efficiency in GPU compute leasing, a critical factor for reducing costs and improving performance in AI workloads. It highlights that GPU compute is often billed by the ho…
-
AI data centers are doubling in size every 10 months
AI data centers are experiencing exponential growth in size, with the IT power capacity of the largest facilities doubling approximately every 10 months since mid-2024. This rapid expansion raises questions about the su…
-
AI vulnerability discovery to favor defense, but expect bumpy transition
AI models are becoming increasingly adept at discovering software vulnerabilities, potentially shifting the balance towards defense in cybersecurity. While current AI tools like Mythos Preview have shown significant cap…
-
US urged to consider military action against China's AGI race
A former White House official has advised the United States to consider extreme measures, including military strikes, to prevent China from achieving artificial general intelligence (AGI) first. Jacob Stokes of the Cent…
-
LLM inference costs plummet, yet user bills soar due to increased usage
Despite a dramatic decrease in LLM inference prices, many users are seeing their bills increase due to the adoption of larger models and always-on agent infrastructure. While the cost per token has plummeted by as much …
-
AI Newsletters Get Narration Podcasts for Wider Accessibility
A new service is offering narration podcasts for several AI-focused newsletters, including those from Redwood Research, Epoch AI, and Zvi. The podcasts aim to provide audio versions of research and content from these or…
-
AI capability costs plummet 1000x, but frontier models remain expensive
The cost of achieving a specific AI capability has decreased dramatically, by approximately 1,000 times since 2021, largely due to the development of smaller, more efficient models. However, the price of accessing the a…
-
OpenAI, Anthropic dominate AI compute growth, driving centralization · 1 source tracked
Dylan Patel, speaking on the Dwarkesh Podcast, stated that OpenAI and Anthropic have significantly increased their AI compute consumption, absorbing approximately 30% of new compute this year and securing 40-50% of next…
-
MaaS Unit Economics: Discounts and Latency Erode Net Revenue
The true cost of Model-as-a-Service (MaaS) is determined not by list prices but by net revenue after discounts and inefficiencies, with discount chains and storage latency being key factors. Storage latency, particularl…
-
Big Tech's $3.1 Trillion AI Funding Pipeline Strains Cash Flow
Big Tech companies are funding their AI expansion through a substantial $3.1 trillion in off-balance-sheet commitments, a figure that significantly surpasses their annual capital expenditures. This financial strategy, w…
-
AI training data supply chain emerges as public text nears finite limits
The era of simply scraping public web data for AI training is ending as the finite supply of high-quality, unique text is becoming apparent. Epoch AI estimates this supply at around 300 trillion tokens, with frontier de…
-
Survey: 20% of US workers delegate tasks to AI
A recent survey by Epoch AI indicates that 20% of US workers are now delegating at least one task to AI that was previously handled by a human colleague. The study also found that these workers generally accept AI-gener…
-
Google consolidates AI leadership in California, eyes $1.5B acquisition
Google is consolidating its AI leadership and talent in California to accelerate the development and deployment of its Gemini models. This strategic shift involves relocating key personnel from London to the Silicon Val…