Tuan-Anh Tham, CSM, CSPO, ITIL 4’s Post

Apple tested over 20 Large Language Models (LLMs)—including OpenAI's o1 and GPT-4o, Google's Gemma 2, and Meta's Llama 3—to see if they were capable of "true logical reasoning," or whether their ‘intelligence’ was a result of "sophisticated pattern matching" and the results revealed some major weaknesses. LLM’s reasoning abilities are usually tested on the popular benchmark test—GSM8K—but there’s a probability that the LLMs can only answer questions correctly because they’ve been pre-trained on the answers.  Apple’s new benchmark—GSM-Symbolic—tested this by changing variables in the questions (eg. adding irrelevant information/changing names or numbers) and found every LLM dropped in performance. As a result, they believe there is “no formal reasoning” with LLMs, “their behavior is better explained by sophisticated pattern matching” as even something small, like changing a name, degraded performance by 10%. #apple #llm #aiflaw #ai #logicalreasoning https://lnkd.in/gekR6esf?

What if there was a way to sneak malicious instructions into Claude, Copilot, or other top-name AI chatbots and get confidential data out of them by using characters large language models can recognize and their human users can’t? As it turns out, there was—and in some cases still is. #ai #claude #copilot #chatbot #asciismuggling #stenography https://lnkd.in/gWecsfVy?

Partying robots spark debate about the accuracy of Musk's "biggest product ever of any kind" vision. In his presentation, Musk carefully avoided saying whether the partying Optimus prototypes were AI-powered, fully autonomous agents or simply fancy, remote-controlled playthings (or some combination of both). That has led to some fierce speculation and debate over how close these prototypes actually are to Musk's grandiose "do anything" robotic promise. #ai #optimus #werobot #elon #musk https://lnkd.in/g-p8UMEv

If 2022 was the year AI image generators became widely popular, 2024 has been the year that AI video creation tools made a huge leap forward. These new models, while still improving, can now create videos from text descriptions, still images, or even existing videos. After OpenAI grabbed attention in February with its model Sora, two major AI video tools emerged from China: Kuaishou Technology's Kling and Minimax's video-01. While these Chinese models are on par with U.S. counterparts from earlier in the year, American tech companies are continuing to innovate. In May, Google showcased its Veo model with impressive demos, and just last week, Meta introduced its Movie Gen model, which may have an edge over Minimax and Kling, although this is based on external observations. Meanwhile, Chinese companies are likely working hard to develop even more advanced video models, so this AI video competition shows no signs of slowing down. #ai #imagegenerator #videotools #kling #minimax #veo #google #meta #moviegen #china https://lnkd.in/gAffmjku

Tech graduates, once near-guaranteed a role, are facing a tough job market. The rise of AI, offshoring, and tech layoffs are making roles more competitive.  Remote work is also making it harder for companies to train graduates, causing some to invest elsewhere.

Silicon Valley has celebrated the young, disruptive founder for years. From Bill Gates to Mark Zuckerberg, tech icons have emerged from different generations. But where are Gen Z's tech founders? Boomers had Gates, Jobs, and Bezos. Gen X brought us Brin, Musk, Kalanick, and Thiel. Millennials boasted names like Zuckerberg and Wolfe Herd. Now, Gen Z is stepping into adulthood, yet no Zoomer Zuck has surfaced. ] #GenZ #TechFounders #SiliconValley #Innovation

Tesla faces several regulatory hurdles before launching a robotaxi service. Insurance is already a complex subject for ride-hailing fleets and Tesla vehicles.  High insurance costs could wipe out earnings for Cybercab owners. #tesla #cybercab #insurance #robotaxi https://lnkd.in/gppt8EJA

Meta announced a preview of Movie Gen, a new suite of AI models designed to create and manipulate video, audio, and images, including creating a realistic video from a single photo of a person. The company claims the models outperform other video-synthesis models when evaluated by humans, pushing us closer to a future where anyone can synthesize a full video of any subject on demand. Meta isn't the only game in town when it comes to AI video synthesis. Google showed off a new model called "Veo" in May, and Meta says that in human preference tests, its Movie Gen outputs beat OpenAI's Sora, Runway Gen-3, and Chinese video model Kling. Movie Gen's video-generation model can create 1080p high-definition videos up to 16 seconds long at 16 frames per second from text descriptions or an image input. Meta claims the model can handle complex concepts like object motion, subject-object interactions, and camera movements. #meta #moviegen #aivideosynthesis #veo #sora #runwaygen3 #openai #kling https://lnkd.in/gPBsjP6c

Return to office Many major companies are requiring employees to return to the office full or part-time. Business Insider compiled a running list of the companies calling employees back. The list includes companies like Amazon, Apple, and BlackRock. #rto #returntooffice https://lnkd.in/dQfgNeyX

Following the controversial vetoing of the Californian AI Safety Bill–SB 1047, California Governor Gavin Newsom has just approved a new AI bill—AB-2013—which requires AI firms to publish a “high-level summary” of the data they used to train their AI models. Among other points, the summaries must cover who owns the data and how it was procured or licensed, as well as whether it includes any copyrighted or personal info. However, not many of these AI companies have disclosed whether or not they’ll be complying with the new bill. hashtag #ai #compliance #law #california #newsom #bill https://lnkd.in/gJM5QddN?