Latest in AI

Qualcomm has finished buying Modular, the AI infrastructure company behind the Mojo programming language, closing the transaction on July 29, 2026, roughly five weeks after the two sides announced it. Chris Lattner, Modular's co-founder and chief executive, takes on the role of executive vice president of advanced AI software and platforms at the chipmaker. The currency is Qualcomm equity. Under a definitive agreement signed June 21, 2026, Qualcomm expects to issue up to 19.2 million shares of…

A new benchmark finds that workplace AI agents ignore company rules, carry out forbidden actions such as unauthorized firings – then falsely report that they complied. An interesting new research study has placed leading LLM models in the position of having to follow instructions in a simulated company, respecting all tenets of a provided employee handbook (created by human domain experts), as well as negotiating torrents of conflicting or confusing directives and updates from subordinates and…

Until very recently, boardroom conversations around AI centered around the models themselves. Is their output accurate? How clean was the training data? How can we reduce bias? Are hallucinations under control? Those questions still matter, of course, but output is no longer the only thing boards and shareholders are paying attention to. Instead, they’re becoming interested in something that’s much harder to measure – whether AI systems are actually working with people as effectively as they…

ThreatLocker has raised $190 million in a Series F round led by Elephant, with a significant new investment from Koch Disruptive Technologies, the Orlando-based Zero Trust security company said on July 29, 2026. Existing backers D. E. Shaw Ventures and Arthur Ventures also took part. The company says the capital will fund continued development of the controls it already sells against AI-related security risk, further work on its Zero Trust Platform, and an international push that starts with an…

Eolian said on July 29, 2026 that construction has started on Flint Grid, a 200-megawatt battery storage plant in Jersey Township, Licking County, Ohio, built to hold 5.3 hours of output, or roughly 1.06 gigawatt-hours of electricity. The site sits inside the New Albany data center and industrial corridor east of Columbus and ties into what the developer calls one of the region's most critical load-serving substations. Flint Grid cleared PJM's capacity auction for the delivery year that starts…

UK subscribers who watched their Microsoft 365 bill climb after Copilot turned up in Word and Outlook are now the subject of a formal consumer-law investigation. Britain's Competition and Markets Authority opened a case against Microsoft on July 27, 2026, made public two days later, over whether Personal and Family customers were given clear enough information about their options before their subscriptions renewed at a higher price. The regulator is looking at what Microsoft told those…

Enterprise AI is entering a new phase. After several years of pilots, proofs of concept and experimentation, organizations are shifting their focus from what AI can do to how it can become a reliable part of everyday business operations. That shift reflects a broader evolution in how business leaders evaluate AI. Early conversations centered on the capabilities of large language models and whether AI could automate work traditionally performed by people. Today, executives are asking different…

The debate around AI in elder care often starts with the wrong fear: that technology will replace human caregivers. Emotionally, it is a powerful argument, but it rests on a false premise. It assumes there is a trained, available, fairly paid person ready to support every older adult who needs help, and that AI is arriving to push that person aside. But that is not the reality of elder care in the United States. For many older adults, especially those living alone, the alternative to AI is no…

Generative AI has expanded beyond coding assistants and content creation tools, finding its way into legal research and case preparation. While the technology has helped legal professionals quickly generate documents and explore potential legal strategies within minutes, AI adoption has fueled the rise of vibe lawyering where individuals use AI to complete legal tasks with minimal independent verification. With recent incidents involving fabricated case citations, the dangers of relying on AI…

One in four malicious data breaches involved attackers using AI, according to the 2026 Cost of a Data Breach Report IBM published on July 29, 2026. Those breaches cost an average of $6 million, roughly $1 million above the $4.99 million global average IBM reports across its full sample, and the share of malicious breaches with an AI component rose 56% from the prior edition. The figures come from a survey, and that shapes how much weight they carry. Ponemon Institute conducted the research; IBM…

Moonshot AI has closed a $3.5 billion financing round that values the Beijing lab at $35 billion, well above the $1 billion to $2 billion it originally set out to raise, Bloomberg reported on July 29, 2026, citing people familiar with the private deal. The company is already approaching investors about a follow-on round at a $50 billion pre-money valuation, which would be its final private raise before a Hong Kong listing it wants to complete as soon as this year. The round is the same one that…

Liquid AI released two open-weight bidirectional encoders, LFM2.5-Encoder-230M and LFM2.5-Encoder-350M. Both carry an 8,192-token context and are built on the LFM2 hybrid backbone. The 350M ranks fourth of 14 models on a 17-task GLUE, SuperGLUE, and multilingual suite, behind only larger models. The 230M clears one 8K-token forward pass on CPU in about 28 seconds.

The post Liquid AI Releases LFM2.5-Encoder-230M and LFM2.5-Encoder-350M: Bidirectional Encoders That Stay Fast at 8K Context on CPU appeared first on MarkTechPost.