AI issue is a news media that delivers AI news in an easy-to-read way, every day.
MoEngage, an Indian marketing automation company, has acquired a company with technology that assigns AI agents to individual customers in an all-cash transaction. Through this acquisition, MoEngage aims to integrate personalized AI agent technology into its platform and refine marketing execution granularity from group-level to individual-level targeting.
Moengage·Jun 24, 2026 05:23 UTCFlipkart, the major Indian e-commerce company backed by Walmart, announced that its ultra-fast delivery hubs called 'Micro-fulfillment Centers' have exceeded 1,000 locations. Amazon is also strengthening its quick commerce operations in the Indian market, intensifying competition between the two companies.
Flipkart·Jun 24, 2026 01:22 UTC
DeepSeek has published a research paper on SPCT, a novel technique for enhancing the scalability of general-purpose reward models in the inference phase. This work is considered part of the research effort toward developing the next-generation model R2, and has garnered significant industry attention in the field of inference scaling.
Deepseek·Jun 23, 2026 21:19 UTC
AI startup Krea has released its image generation model 'Krea 2' as open weights in two versions: 'Raw' and 'Turbo'. Krea 2 Turbo achieves generation in approximately 2 seconds, making it one of the fastest AI image generation models available. Companies with 50 or more employees are required to have a paid Enterprise contract, and license terms mandate that all users implement technical safeguards to prevent the generation of illegal content.
Krea·Jun 23, 2026 19:25 UTC
Chinese AI company Zhipu.AI has released its large language model 'GLM' series as open source after accelerating it up to 8 times faster than before. Simultaneously, the company launched a new service 'Z.ai' aimed at global expansion, marking a full-scale entry into overseas markets. These moves coincide with the period when the company is reportedly considering an IPO, and are viewed as a strategy targeting both penetration into the developer community and enhanced international brand recognition.
Zhipuai·Jun 23, 2026 17:21 UTC
SpaceX has signed a computing resource supply contract with a company that provides open-source AI models. This agreement demonstrates that SpaceX is expanding its business as a computing infrastructure provider and highlights the critical importance of securing computing resources for AI model companies.
Spacex·Jun 23, 2026 13:21 UTC
OpenAI has announced the cybersecurity-specialized model "GPT-5.5-Cyber" and an updated plugin "Codex Security," expanding its "Daybreak" initiative. The company has also established a partner network that includes over 25 security companies and multiple government agencies. This effort represents a significant shift in focus from vulnerability "detection" to automatic vulnerability "remediation."
OpenAI·Jun 23, 2026 11:28 UTC
Kwai, a Chinese video platform company, announced SRPO, a new reinforcement learning framework for large language models (LLMs). SRPO combines two-stage reinforcement learning with historical data reuse to reduce learning steps by approximately 90% compared to the existing GRPO method while achieving performance equivalent to DeepSeek-R1 in mathematics and code generation benchmarks.
In 2026, major tech companies are increasingly announcing large-scale workforce reductions citing AI adoption as the official reason. Against the backdrop of business automation and a shift in investment toward AI, the relationship between employment and AI implementation is being questioned across the technology industry.
OpenAI has launched a new initiative to support the discovery and remediation of security vulnerabilities in open source software. By leveraging artificial intelligence to detect bugs and facilitate fixes, the company addresses security challenges in open source software widely used throughout society.
OpenAI·Jun 23, 2026 01:20 UTC
DeepSeek has released an open-source AI model called "DeepSeek-Prover-V2" for automatically proving mathematical theorems. Supporting Lean 4 and combining training data generation using DeepSeek-V3 with reinforcement learning, the model achieved state-of-the-art results on the theorem proving benchmark "MiniF2F".
Deepseek·Jun 22, 2026 23:21 UTC