OpenAI published receipts, in the form of a long timeline of emails, texts and legal filings, illustrating that Elon Musk's injunction to prevent OpenAI from converting into a for-profit company runs counter to what he wanted in 2017. Essentially, OpenAI is providing even more evidence to the fact that its former co-founder wanted the AI startup to become a for-profit company and make him CEO. You should read the whole blog to get all of the details (and get a sense for how billionaires email) but the gist is that in 2017, Musk and OpenAI came to an understanding that the then non-profit needed to become a for-profit to "advance its mission" and seemingly capitalize on the public interest earned from its AI beating professional Dota 2 players in one-on-one matches. According to OpenAI, Musk proposed a new board structure where he "would unequivocally have initial control of the company," which OpenAI was opposed to. That led to the disagreements between Musk and OpenAI leadership, and him ultimately leaving the nonprofit's board in 2018.

Thank you so much to all the readers who tuned in live to participate in the second installment of our question and answer series focused on artificial intelligence. I was thrilled to see so many questions come in before the event, as well as all the questions that were dropped into the chat during our conversation. Below is a replay of this event that WIRED subscribers can watch whenever. Also, the livestream from the first one is available here. I started off the chat with a couple quick demos showing how to use the image and voice features built into chatbots, including an example of how it's possible to interact with ChatGPT's Advanced Voice Mode as a kind of Duolingo-style language learning tool.

More than 300 million people use OpenAI's ChatGPT each week, a testament to the technology's appeal. This month, the company unveiled a "pro mode" for its new "o1" AI system, offering human-level reasoning -- for 10 times the current 20 monthly subscription fee. One of its advanced behaviours appears to be self-preservation. In testing, when the system was led to believe it would be shut down, it attempted to disable an oversight mechanism. When "o1" found memos about its replacement, it tried copying itself and overwriting its core code.

Since 2018, carbon emissions from data centers in the US have tripled. For the 12 months ending August 2024, data centers were responsible for 105 million metric tons of CO2, accounting for 2.18% of national emissions (for comparison, domestic commercial airlines are responsible for about 131 million metric tons). About 4.59% of all the energy used in the US goes toward data centers, a figure that's doubled since 2018. It's difficult to put a number on how much AI in particular, which has been booming since ChatGPT launched in November 2022, is responsible for this surge. That's because data centers process lots of different types of data--in addition to training or pinging AI models, they do everything from hosting websites to storing your photos in the cloud.

The AI company Open AI announced last week that it would release something new every day for 12 days, and yesterday's reveal was a new voice for ChatGPT. The chatbot can now answer your questions in the voice of Father Christmas. You can activate Santa's voice by tapping on the snowflake in the app, or you can select it from the ChatGPT voice menu. However, Santa is only making a temporary visit to the chatbot, at the end of the month the voice will be removed from Chat GPT. Open AI has also announced that you can now share your screen or video with Chat GPT in "Advanced Voice" mode.

How will we know if an artificial intelligence has become conscious? If your response is "I don't know", congratulations – you have just passed part of a test designed to measure AI intelligence by posing questions to which there is no known answer. AIs like ChatGPT that are powered by large language models can create compellingly plausible responses to any question, but there is no guarantee that their answers have any basis in reality. This issue can be exacerbated when dealing with unsolved problems in science and philosophy.

Seven in 10 secondary school students have used large language models (LLMs) for their studies, according to a survey of more than 300 US pupils. "I realised that a lot of the people around me were using large language models, and more specifically ChatGPT, for a lot of school assignments," says Tiffany Zhu, an 11th-grade student (equivalent to year 12 in the UK) at The Harker School in San Jose, California.

Despite the significant advancements in Text-to-SQL (Text2SQL) facilitated by large language models (LLMs), the latest state-of-the-art techniques are still trapped in the in-context learning of closed-source LLMs (e.g., GPT-4), which limits their applicability in open scenarios. Our approach begins with multi-task supervised fine-tuning (SFT) using various synthetic training data related to SQL generation. Unlike existing SFT-based Text2SQL methods, we introduced several additional SFT tasks, including schema linking, noise correction, and continuation writing. Engaging in a variety of SQL generation tasks enhances the model's understanding of SQL syntax and improves its ability to generate high-quality SQL queries. Additionally, inspired by the collaborative modes of LLM agents, we introduce a Multitask Collaboration Prompting (MCP) strategy. This strategy leverages collaboration across several SQL-related tasks to reduce hallucinations during SQL generation, thereby maximizing the potential of enhancing Text2SQL performance through explicit multitask capabilities. Extensive experiments and in-depth analyses have been performed on eight open-source LLMs and five widely-used benchmarks. The results demonstrate that our proposal outperforms the latest Text2SQL methods and yields promising performance. The code and data are available here. Text2SQL has emerged as a popular and practical technology for question answering based on largescale databases, serving as a crucial link between natural language and database systems (Zhang et al., 2024). Recently, Large Language Models (LLMs) have proven to be an effective solution in Text2SQL (Pourreza & Rafiei, 2024a).

Large language models (LLMs) frequently generate confident yet inaccurate responses, introducing significant risks for deployment in safety-critical domains. We present a novel approach to detecting model hallucination through systematic analysis of information flow across model layers when processing inputs with insufficient or ambiguous context. Our investigation reveals that hallucination manifests as usable information deficiencies in inter-layer transmissions. While existing approaches primarily focus on final-layer output analysis, we demonstrate that tracking cross-layer information dynamics ($\mathcal{L}$I) provides robust indicators of model reliability, accounting for both information gain and loss during computation. $\mathcal{L}$I improves model reliability by immediately integrating with universal LLMs without additional training or architectural modifications.

Many applications demand context sensing to offer personalized and timely services. Yet, developing sensing programs can be challenging for developers and using them is privacy-concerning for end-users. In this paper, we propose to use natural language as the unified interface to process personal data and sense user context, which can effectively ease app development and make the data pipeline more transparent. Our work is inspired by large language models (LLMs) and other generative models, while directly applying them does not solve the problem - letting the model directly process the data cannot handle complex sensing requests and letting the model write the data processing program suffers error-prone code generation. We address the problem with 1) a unified data processing framework that makes context-sensing programs simpler and 2) a feedback-guided query optimizer that makes data query more informative. To evaluate the performance of natural language-based context sensing, we create a benchmark that contains 133 context sensing tasks. Extensive evaluation has shown that our approach is able to automatically solve the context-sensing tasks efficiently and precisely. The code is opensourced at https://github.com/MobileLLM/ChainStream.