Content Fans

Category

AI News & Trends

1402 articles • Page 40 of 94

Nvidia acquires Kumo AI for $400M, expands enterprise software play

Nvidia acquires Kumo AI for $400M, expands enterprise software play

Nvidia has acquired Kumo AI, a five-year-old enterprise software startup, for at least $400 million, according to people familiar with the deal. The purchase may help Nvidia expand its software offerings, especially in making predictions for things like fraud detection and demand forecasting. Kumo's technology appears to quickly turn raw business data into predictions, and it claims to work with companies like Reddit and DoorDash. Some analysts say this move fits into Nvidia's recent pattern of buying AI software firms, suggesting a strategy to be more central in business workflows. However, performance claims mainly come from the vendor, and some details about the deal have not been made public.

CIOs Adopt New Playbook to Combat AI Memory Shortages Through 2030

CIOs Adopt New Playbook to Combat AI Memory Shortages Through 2030

Enterprise IT leaders may face memory shortages for AI through 2030, so CIOs are using new strategies to plan ahead. They can use better demand forecasting and work with multiple suppliers to avoid being caught off guard. Building flexible systems and treating memory as a resource that can be managed in tiers may help. Companies might use special tools and checklists to track memory needs and supplier reliability. With these steps, memory shortages appear to be a planning issue rather than a crisis.

Anthropic's Claude Opus 4.8 ships faster, cheaper AI model

Anthropic's Claude Opus 4.8 ships faster, cheaper AI model

Anthropic has released Claude Opus 4.8, which may be faster and cheaper than previous versions. Testing suggests it completes tasks about 2.5 times quicker and at about one-third the cost in fast mode. Early results and user feedback indicate better reliability for web tasks and possible improvements in spotting coding errors, though outside audits are still limited. Some benchmarks suggest Opus 4.8 leads in certain coding tasks but might lag behind OpenAI's Codex for command-line work. If more reviews support these findings, Opus 4.8 could be a good choice for developers, but some teams may still prefer other models for specific needs.

Microsoft Unveils Surface RTX Spark Dev Box for AI Agent Development

Microsoft Unveils Surface RTX Spark Dev Box for AI Agent Development

Microsoft announced the Surface RTX Spark Dev Box, which may help developers build and run AI agents locally on Windows computers. The device comes with powerful hardware and software tools, including Visual Studio Code and GitHub Copilot, and is designed for tasks like AI training and running large models. Microsoft suggests that its new approach connects hardware, multiple AI models, and security features so companies can use agents locally and then move tasks to the cloud if needed. Security tools such as Microsoft Execution Containers and Defender scanning aim to keep agent actions controlled and safe. Reports suggest that more companies are using AI agents, and Microsoft's new products may help support this trend by making agents easier to use and manage on employees' computers.

OpenAI uses Codex to migrate 600 petabytes in two months

OpenAI uses Codex to migrate 600 petabytes in two months

OpenAI used Codex, a large language model, to help move about 600 petabytes of data and rebuild 10,000 workflows in just two months. Codex generated scripts and checked data as it moved between cloud providers, which may have cut development time by about half. Engineers added approval steps and safety checks at risky points, suggesting that careful human review is still important. The results show Codex-style automation might soon be common in big data projects, but human oversight seems necessary for safety. Error rates were very low, and most problems were fixed quickly, which appears to match or beat usual manual methods.

Microsoft Unveils Four IQ Services for Enterprise AI Agents at Build 2026

Microsoft Unveils Four IQ Services for Enterprise AI Agents at Build 2026

Microsoft has introduced four new IQ services - Web IQ, Work IQ, Foundry IQ, and Fabric IQ - that may help enterprise AI agents work better by providing key information in separate layers. These services aim to make agent setup faster, safer, and simpler by handling different types of data such as live web content, organizational connections, business knowledge, and company metrics. Early feedback suggests these tools might speed up development, improve information quality for agents, and give IT teams better control. However, some details about how these services work and how easy they will be to use with other platforms are still unclear. Experts say that future adoption will likely depend on how well the new APIs perform and whether outside platforms can use them easily.

New Data Project Aims to Standardize AI Token Usage Metrics

New Data Project Aims to Standardize AI Token Usage Metrics

A new Data Project aims to help companies measure how many AI tokens are creating real economic value. The project suggests collecting anonymized data on token usage and product outcomes across different firms. This may allow leaders to compare their AI investments and spot trends in adoption and productivity. Early benchmark data suggests AI referrals might convert at higher rates than traditional search, but direct links between tokens and value are still unclear. The project hopes to create shared metrics that let companies track and compare their AI impact more accurately.

OpenAI Codex tops 5M weekly users, expands into knowledge work

OpenAI Codex tops 5M weekly users, expands into knowledge work

OpenAI says Codex now has more than 5 million weekly users, a big increase since February. About 20 percent of users may be knowledge workers, not just developers. Codex is being used for more than just coding, as it is now part of many business tools and workflows inside ChatGPT. Studies suggest it might boost productivity for both developers and business users. It appears that more companies are starting to use Codex regularly, but it is not certain if the high growth will continue.

AI Hallucinations Spur Slopsquatting Threat for Software Supply Chains

AI Hallucinations Spur Slopsquatting Threat for Software Supply Chains

Researchers say AI tools sometimes suggest fake software package names that sound real, and criminals may register these names to spread malware through automated build systems. This new risk, called slopsquatting, appears to be growing in 2025, but there have not yet been large confirmed attacks. Some experts warn that the threat is credible because automated systems might trust these fake packages without review. Security advice suggests using stricter controls like lockfiles, human approval, and monitoring for suspicious new packages. The risk is mostly theoretical right now, but researchers recommend acting early to prevent future problems.

Veracode: 45% of AI-generated code snippets contain security flaws

Veracode: 45% of AI-generated code snippets contain security flaws

Recent research suggests that 45% of AI-generated code snippets may contain security flaws. While AI tools can help teams create features quickly, proving that the code is safe and reliable often takes much longer. Some studies show that AI assistance leads to more completed tasks without lowering code quality, but others report more bugs and duplicated code. Experts recommend treating all AI-generated code as untrusted and using careful review and testing to find problems. Many teams are now adding security checks early, having experts review risky changes, and being careful about where and how they use AI in critical systems.

New AI Metric Tracks Tokens Per Shipped Product

New AI Metric Tracks Tokens Per Shipped Product

A new project proposes collecting data on how many AI tokens are used per shipped product across different companies, which may help measure AI efficiency better than just counting tokens. Experts suggest tracking tokens per completed task or feature, since raw token numbers may not show real business impact. The project would use privacy-safe, aggregated data and focus on metrics like tokens used, tasks finished, and defect rates. Current studies suggest that many AI projects do not show clear financial gains, so linking token use to shipped results might help spot effective practices. If done carefully, this approach could let teams and leaders compare AI efficiency more accurately.

Anthropic unveils Claude Opus 4.8: Faster, cheaper, safer AI

Anthropic unveils Claude Opus 4.8: Faster, cheaper, safer AI

Anthropic has released Claude Opus 4.8, which may be faster, cheaper, and safer than its earlier version. The company says Opus 4.8 costs about one-third as much and gives responses around 2.5 times quicker than Opus 4.7 in fast mode. Early tests suggest it performs well at fixing complex code and avoids misaligned behavior more often, but some engineers might still prefer other tools like Codex for certain quick tasks. Anthropic's own checks indicate the new model flags uncertainty more and avoids unsupported claims, though these results may need confirmation from outside labs. Pricing stays similar to before, and actual savings might depend on how teams use it.

AWS Trainium Cuts AI Costs Up to 50% for Anthropic, Uber

AWS Trainium Cuts AI Costs Up to 50% for Anthropic, Uber

Amazon, Google, and Meta are now competing in AI by focusing on things like computer chips, data systems, and how companies make money from AI. AWS's custom Trainium chips reportedly help companies like Anthropic and Uber cut AI costs by up to 50 percent, though the use of these chips is still selective. Microsoft is focusing on controlling data access and security, which might make it harder for companies to switch away from their products. Reports suggest most AI startup revenue is now going to just two companies, OpenAI and Anthropic, which together may control about 89 percent of the market. These trends suggest that control over chips, data, and APIs may decide which companies lead in AI, but it is not yet clear who will win.

AI Code is 86% Flawed in Benchmarks, CSET Study Finds

AI Code is 86% Flawed in Benchmarks, CSET Study Finds

A recent study suggests that AI-generated code has many flaws, with failure rates as high as 86% for some security issues. Researchers found that nearly half of AI code samples show problems like SQL injection and improper authentication. The main challenge may be checking and fixing AI code, not just creating it, so teams are adding more human reviews and strict checks. Productivity gains are mixed; AI might speed up some tasks, but could also slow down experienced programmers and increase maintenance work. Experts recommend careful oversight and collecting data to ensure code quality and safety.

AI hallucinations now enable 'slopsquatting' attacks, say security experts

AI hallucinations now enable 'slopsquatting' attacks, say security experts

Security experts say AI tools sometimes invent fake package names, which attackers might quickly register to trick automated coding systems. These so-called "slopsquatting" attacks may allow bad actors to silently add malicious software when teams use AI to pick code dependencies. Tests in 2025 and 2026 show that many fake names created by AI were actually downloaded or used widely, even though they did not include harmful code. Experts suggest multiple security steps may reduce, but not fully stop, these risks. It remains unclear how often these attacks lead to real harm, since there is little direct evidence so far.