Several key developments are shaping the AI landscape. Huawei is actively defending its Pangu Pro MoE AI model against accusations of copying Alibaba's Qwen-2.5 14B model, asserting independent development on Ascend chips and adherence to open-source licenses. Meanwhile, Meta's significant $14.8 billion investment in Scale AI, structured as a 'non-acquisition acquisition,' raises questions about the AI industry's trajectory, especially as competitors like Google and OpenAI distance themselves from Scale AI. In other applications, Google Labs is experimenting with Doppl, an AI app for virtual try-ons, showing mixed results with complex clothing designs. CarMax is leveraging AI to refine used car sales through pricing optimization, online information enhancement, and supply chain improvements, even using ChatGPT to organize car model data and the Rhodes tool to support employees. Concerns about AI's impact are also surfacing, with Robert Kiyosaki warning of massive job losses and advising investment in alternative assets. A university lecturer is adapting assessment methods to combat AI-driven cheating, while federal departments in Canada are testing ParlBrief, an AI tool powered by Cohere, OpenAI, and Mistral, to monitor Parliament and streamline report generation. However, the risks of AI are also evident, as an Atlanta lawyer faced penalties for using AI-generated fake legal cases in a divorce proceeding, highlighting the issue of AI "hallucinations."
Key Takeaways
- Huawei denies copying Alibaba's Qwen model for its Pangu Pro AI model, stating it was independently developed on Ascend chips.
- Meta is investing $14.8 billion in Scale AI, raising questions about the AI industry's future as competitors distance themselves.
- Google Labs' Doppl app allows virtual try-ons using AI, with varying success depending on outfit complexity.
- CarMax is using AI to improve used car sales through better pricing, online information, and supply chain management.
- Robert Kiyosaki warns AI will cause significant job losses and advises investing in assets like real estate and Bitcoin.
- A university lecturer is changing assessment methods to combat AI cheating by focusing on personal and creative assignments.
- Canadian federal departments are testing ParlBrief, an AI tool using Cohere, OpenAI, and Mistral models, to monitor Parliament and save staff time.
- An Atlanta lawyer was penalized for using AI to generate fake legal cases, highlighting the risk of AI "hallucinations."
Huawei denies copying Alibaba's AI model Pangu Pro
Huawei's Noah Ark Lab denies its Pangu Pro AI model copied Alibaba's Qwen model. HonestAGI claimed Huawei's model showed 'extraordinary correlation' with Alibaba's. Huawei stated Pangu Pro was independently developed and trained on Huawei's Ascend chips. They added that they followed open-source license requirements for any third-party code used. Huawei open-sourced its Pangu Pro Moe models on GitCode in late June.
Huawei defends its AI models amid copying allegations
Huawei is defending its Pangu Pro MoE AI model, saying it was developed in-house. Allegations arose after an account on GitHub, HonestAGI, claimed the model had a strong link to Alibaba's Qwen-2.5 14B model. Huawei's Noah's Ark Lab stated the model was trained on Huawei's Ascend AI chips and not based on other models. The lab said they used some open-source codes, following license rules. The controversy comes as Chinese tech firms compete to develop AI models.
Huawei's AI lab denies copying rival AI models
Huawei's AI lab is pushing back against claims it copied other companies' models for its Pangu platform. The lab says its Pangu Pro MoE model is the first of its kind trained on Ascend chips. Huawei admits it used open-source code, which is common, but followed all licensing rules. The accusations were posted on Github, claiming Pangu's code had uncredited material from rivals. Huawei's response is unusual for a company focused on using domestic technology.
Huawei's AI scandal raises investor concerns about competition
Huawei's Noah's Ark Lab addressed accusations that its Pangu Pro MoE AI model borrowed code without credit. A group on GitHub claimed to have found unacknowledged code. Huawei responded that it complied with open-source licenses and welcomed technical discussion. Investors are watching how Huawei competes with Alibaba and DeepSeek in AI. The situation highlights the importance of IP compliance and trust in the AI industry.
Trying Google's new AI dressing room app Doppl
Google Labs launched Doppl, an AI app that lets users virtually try on clothes. Users upload a photo of themselves and an outfit, and the app creates a digital image of them wearing it. The reviewer tested Doppl with various outfits, with mixed results. Simpler outfits worked best, but the AI struggled with complex designs. Doppl is still an experiment, but it offers a glimpse into the future of online shopping.
Meta's Scale AI deal signals AI industry trouble
Meta's $14.8 billion investment in Scale AI suggests the AI industry may be declining. The deal is structured as a 'non-acquisition acquisition' to avoid regulatory review. Meta gains control of Scale AI's data-labeling pipeline and hires key personnel. Competitors like Google and OpenAI are cutting ties with Scale AI due to conflict-of-interest concerns. The deal highlights Meta's struggle to catch up in the AI race and questions the progress of AI development.
Robert Kiyosaki warns AI will cause massive job losses
Robert Kiyosaki, author of "Rich Dad Poor Dad," warns that AI will cause massive unemployment. He advises people to take proactive action and not become victims. Kiyosaki suggests investing in entrepreneurship, real estate, gold, silver, and Bitcoin. He criticizes traditional financial advice and the unstable financial system. Kiyosaki encourages people to prepare for a future where job security is uncertain.
CarMax uses AI to reinvent used car sales
CarMax is using AI to change how people buy and sell used cars. Shamim Mohammad, CarMax's tech chief, learned from Blockbuster's failure to adapt to streaming. CarMax uses AI for pricing, online information, and supply chain management. They were early adopters of ChatGPT to organize car model data. CarMax also uses AI to summarize customer reviews and provide support to employees through the Rhodes tool.
US AI chipmaker Groq opens data center in Finland
Error Something went wrong here. Please log out, log in and try again. If error persists, we're ready to help at**[email protected]**...
Lecturer says AI cheating increases workload, changes assessments
A university lecturer, Risa Morimoto, is concerned about students using AI to cheat. She says AI makes cheating harder to detect, increasing her workload. Morimoto plans to change her assessment methods to discourage AI use. She will ask students to summarize class learning and create a blog. The goal is to make assignments more personal and creative, relying less on AI-generated content.
Federal departments test AI tool to monitor Parliament
Federal departments in Canada are testing ParlBrief, an AI tool to track Parliamentary committees. The tool transcribes meetings, summarizes discussions, and analyzes sentiment. It aims to cut the time to produce reports by one hour, saving staff time. Four departments are testing ParlBrief, powered by AI models from Cohere, OpenAI, and Mistral. The government hopes AI will improve public service productivity.
Lawyer penalized for using fake AI cases in divorce
An Atlanta lawyer, Diana Lynch, was penalized for citing fake legal cases in a divorce case. Judges believe she used AI to create her arguments. Lynch was fined $2,500 for the error. The judges noted the use of fictitious cases deprived the ex-wife of a fair response. The case highlights concerns about AI "hallucinations" in legal filings.
Sources
- Huawei's AI lab denies that one of its Pangu models copied Alibaba's Qwen
- Huawei defends AI models as home-grown after whistle-blowers raise red flags
- Huawei’s AI Lab Fends Off Accusations It Copied Rival Models
- Huawei's AI Scandal Just Exploded--And Investors Should Be Paying Attention
- Exclusive | I tried Google’s new AI dressing room — here’s what it got...
- Meta's $14.8B Scale AI deal signals AI industry decline
- Robert Kiyosaki Calls AI The 'Biggest Change In Modern History.' He Warns You Need To Take Proactive Action To Avoid Becoming A 'Victim'
- CarMax's top tech exec shares his keys to reinventing a legacy retailer in the age of AI
- US AI chipmaker Groq opens first European data center in Finland
- I'm a university lecturer concerned that students are using AI to cheat. It's made my workload skyrocket, and I've had to make drastic changes.
- Federal departments test AI tool to help monitor what Parliament is doing
- Judges slam Atlanta divorce lawyer’s apparent use of AI | Chattanooga Times Free Press