Quantum computing applies the principle of superposition to process information far beyond the logic of classical systems. Although it still faces challenges with stability and temperature, it already hints at new ways to solve highly complex problems. Microblog by @antgrasso
RESEARCH
-
Sequential Decoding Behavior in Diffusion Language Models
By
–
interesting, if it is a DLM why does it seem to decode sequentially?
-

Valuing Wikipedia Data in LLM Training: Copyright and Markets
By
–
Wikipedia comprises ~1% of LLM training data, but what's its worth? Scholars are attempting to put a price tag on the URLs, books & data used by AI companies. This could aid copyright cases & create more accurate data markets. Follow the discussion here: https://t.co/rbgxicwtD0 pic.twitter.com/GIqRjZTqRM
— Stanford HAI (@StanfordHAI) 9 octobre 2025Wikipedia comprises ~1% of LLM training data, but what's its worth? Scholars are attempting to put a price tag on the URLs, books & data used by AI companies. This could aid copyright cases & create more accurate data markets. Follow the discussion here: https://
hai.stanford.edu/events/hoffman
-yee-symposium-2025
… -

2025’s AI shift: architecture over scale
By
–
everyone’s chasing bigger models. trillion tokens, billion-dollar data centers, rivers of compute. but the real AI story of 2025 isn’t about scale -> it’s about architecture. while openai burns oceans training gpt-5, samsung just dropped a 7-million-parameter reasoning model
-

Gemini 2.5 Deep Think Sets New Frontier Math Record
By
–
Gemini 2.5 Deep Think consigue un nuevo record en Frontier Math, evaluado por la buena gente de Epoch
-

Sonnet 4.5 vs GPT-5: METR benchmark autonomous work hours
By
–
Sonnet 4.5 aparece en la gráfica de METR colocándose en la línea de tendencia pero sin superar al punto más alto, ocupado por GPT-5, en este benchmark que mide el equivalente de horas de trabajo humano que un LLM puede resolver de forma autónoma.
-

GPT-5 Pro Achieves Top Results on ARC-AGI Benchmarks
By
–
Un par de actualizaciones a benchmarks interesantes que estábamos esperando. GPT-5 Pro en ARC-AGI 1 y 2 logra ser el LLM disponible que mejores resultados consigue en ambos benchmarks.
-

Microsoft delivers first NVIDIA GB300 cluster for OpenAI
By
–
Just announced: @Microsoft delivers the world's first at-scale NVIDIA GB300 NVL72 production cluster, providing the supercomputing engine needed for OpenAI to train multitrillion-parameter models in days, not weeks. Learn more: https://
nvda.ws/4ogROKu -
Small Fixed Data Can Poison Any Size LLM Model
By
–
Previous research suggested that attackers might need to poison a percentage of an AI model’s training data to produce a backdoor. Our results challenge this—we find that even a small, fixed number of documents can poison an LLM of any size. Read more:
-

Data-Poisoning Attacks Threaten LLMs Regardless Size
By
–
New research with the UK @AISecurityInst and the @turinginst
: We found that just a few malicious documents can produce vulnerabilities in an LLM—regardless of the size of the model or its training data. Data-poisoning attacks might be more practical than previously believed.
