Posts of last 24 hours
A vulnerability, which was classified as very critical, was found in Linux Kernel up to 6.12.94/6.18.37/7.1.2. This vulnerability affects the function i2c_get_adapter of the component I2C Core. The manipulation results in use after free.
This vulnerability is known as CVE-2026-53400. Attacking locally is a requirement. No exploit is available.
You should upgrade the affected component.
https://vuldb.com/vuln/380131
A vulnerability has been found in Linux Kernel up to 7.1.2/7.2 and classified as critical. This issue affects the function nfsd4_create of the component Nfsd. This manipulation of the argument cr_dpacl/cr_pacl causes memory leak.
This vulnerability is handled as CVE-2026-53395. The attack can be initiated remotely. There is not any exploit available.
The affected component should be upgraded.
https://vuldb.com/vuln/380132
A vulnerability classified as critical has been found in Linux Kernel up to 7.1.2/7.2. This vulnerability affects the function omapfb_mmap of the component Fbdev. The manipulation leads to use after free.
This vulnerability is documented as CVE-2026-53401. The attack needs to be performed locally. There is not any exploit available.
It is recommended to upgrade the affected component.
https://vuldb.com/vuln/380142
A vulnerability classified as critical was found in Linux Kernel up to 7.1.2. Affected by this issue is the function nfsaclsvc_decode_setaclargs/nfs3svc_decode_setaclargs of the component Nfsd. Executing a manipulation of the argument acl_access/acl_default can lead to allocation of resources.
This vulnerability appears as CVE-2026-53397. The attack may be performed from remote. There is no available exploit.
Upgrading the affected component is advised.
https://vuldb.com/vuln/380129
A vulnerability categorized as critical has been discovered in Linux Kernel up to 7.1.2. The affected element is the function nfsd4_decode_secinfo_no_name of the component NFSD. Executing a manipulation can lead to improper initialization.
This vulnerability is tracked as CVE-2026-53398. The attack can be launched remotely. No exploit exists.
It is advisable to upgrade the affected component.
https://vuldb.com/vuln/380123
Эксперты «Гарды» обсудят, как безопасно использовать данные в тестовых средах
https://www.securitylab.ru/news/575454.php
July 30, 2026A couple o
https://buaq.net/go-431823.html
很多顶尖 AI 初创公司都宣称他们的技术将能彻底改变软件开发、药物研发和科学研究，但他们很少用论文去论证他们的宣传。根据发表在 bioRxiv 上的一篇论文预印本，逾半数 AI 独角兽企业(即估值超过 10 亿美元)发表的论文少之又少。数据显示，2025 年发表的 AI 论文中每 1000 篇这些 AI 公司就只占到 1 篇。论文合作者、斯坦福大学的 John Ioannidis 称，“对于一个号称正重塑科学、且在科学潜力方面如此先进的领域而言，没有任何科学文献是一个非常奇怪的悖论，我们如何判断他们所说的是真实、经过验证且可重复的？”Alberta 大学的 AI 伦理学者 Mohamed Abdalla 认为，这可能反映了公司的激励机制与学界不同。公司的激励机制是赚钱而不是推进科学发展。1998-2025 年之间的 317 家 AI 独角兽公司发表了 1389 篇同行评审论文和 688 篇预印本，逾半数初创公司未发表任何合格论文。前 5% 公司贡献了逾九成的引用量，其中 OpenAI 贡献了近四成，其次是中国计算机视觉公司旷视科技和 Hugging Face。中国 AI 公司发表的论文数高于美国的同行，这一结果并不出人意料，美国 AI 公司日益对其模型保密。
https://www.solidot.org/story?sid=84959
很多顶尖 AI 初创公司都宣称他们的技术将能彻底改变软件开发、药物研发和科学研究，但他们很少用论文去论证他们的宣传。根据发表在 bioRxiv 上的一篇论文预印本，逾半数 AI 独角兽企业
https://buaq.net/go-431828.html
Связки нейросетей превзошли отдельные модели в тесте CyberGym и помогли сократить стоимость анализа кода.
https://www.securitylab.ru/news/575446.php