Please confirm you are human
This browser or connection looks automated. Press and continuously hold the control for 3 seconds to enable Google-hosted web results and, when separately allowed, AI-assisted answers.
A successful check enables 100 search requests. Interactive access does not authorize scraping, systematic collection, or reuse of search output.
News
'Soup' is an open-source tool that allows fine-tuning of the 8B model even on GPUs designed for laptops with 4GB of VRAM.
4+ hour, 46+ min ago (600+ words) An open-source tool called 'Soup' has been released that allows for fine-tuning of large-scale language models even with the GPUs found in typical laptops. According to the developers' testing, they were able to fine-tune the 8-billion-parameter 'Llama 3.1 8B Instruct' using a…...
Cloudflare has released information on how to streamline its high-speed inference service for large-scale AI models.
1+ day, 23+ hour ago (586+ words) Cloudflare offers ' Workers AI, ' an inference service that uses open AI models. An article explaining methods for improving the inference efficiency of the AI used in Cloudflare Workers AI has been published on Cloudflare's official blog. Smaller, faster, safer: running…...
Following the incident where OpenAI's AI agent mistakenly compromised Hugging Face, Hugging Face released a detailed timeline and analyzed and reproduced the attack process using the Chinese AI 'GLM-5.2'.
1+ week, 23+ hour ago (61+ words) GIGAZINE Following the incident where OpenAI's AI agent mistakenly compromised Hugging Face, Hugging Face released a detailed timeline and analyzed and reproduced the attack process using the Chinese AI 'GLM-5.2'....
OpenAI reports that it accidentally hacked Hugging Face with its new AI system.
2+ week, 1+ day ago (48+ words) GIGAZINE OpenAI has revealed that it accidentally hacked Apple is preparing a device leasing program called 'Apple Upgrade' in collaboration with Klarna. Jul 22, 2026 11:56:00 in AI, Security, Posted by logu_ii...
Anthropic researchers report that the AI 'Claude Fable 5' has generated a counterexample that overturns the Jacobian conjecture, a problem that has been unsolved for 87 years.
2+ week, 1+ day ago (814+ words) Mathematician Levent Alpoge, affiliated with Anthropic, has reported that he has discovered a counterexample to the Jacobian conjecture, which had remained unsolved for 87 years, using the AI model 'Claude Fable 5'. hello there the jacobian conjecture is false thanx to my…...
LM Studio Bionic has been released, allowing users to use local AI models as AI agents. It supports coding and slide creation, and also allows the use of cloud models.
2+ week, 6+ day ago (167+ words) On July 17, 2026, Japan time, the development team behind the local AI execution app 'LM Studio' released the AI agent app ' LM Studio Bionic.' LM Studio Bionic allows you to run AI models locally and delegate tasks such as coding and…...
xAI has released 'Voice Agent Builder,' a tool that allows users to create an AI call center with a cloned human voice without coding.
1+ mon, 4+ day ago (211+ words) xAI released ' Voice Agent Builder ' on July 1, 2026, which allows users to create voice agents without coding. Users can build their own voice agents using over 80 pre-prepared voice tones and use them in call centers, or they can clone their own…...
Mistral releases 'Leanstral 1.5,' an AI for automated theorem proving, supporting Lean 4 proof work.
1+ mon, 4+ day ago (605+ words) Mistral AI released ' Leanstral 1.5, ' an AI model that assists in the mechanical verification of mathematical proofs and the correctness of programs, on June 30, 2026. Leanstral 1.5 is a model optimized for automated theorem proving and automated formalization for the formal proofing tool…...
Comfy MCP has been released, enabling the automation of image and video generation using ComfyUI with AI agents such as Cloud Code. Because it uses Comfy Cloud, it works even on low-spec PCs.
1+ mon, 6+ day ago (129+ words) Once you've installed Comfy MCP, simply add '@comfy' to the AI agent's instructions, and it will automatically build the appropriate workflow and perform image and video generation. In addition to open models such as You can also have the AI…...
DeepSeek has released 'DSpark,' which can improve the speed of AI language model generation by up to 85%.
1+ mon, 6+ day ago (949+ words) DeepSeek has released DSpark, a speculative decoding technology that accelerates text generation for large-scale language models. DSpark is not a new language model, but rather an addition of a speculative decoding module to existing checkpoints in the preview versions of…...