Domain, Once the Flagship of OpenAI’s ChatGPT, Now Becomes the New Face of Elon Musk’s X.AI in a Surprising Domain Transfer

Today: flips from ChatGPT to Elon Musk’s

Remember when started pointing you to ChatGPT, the project by OpenAI? Yeah, not anymore. Now, it seems Elon Musk, billionaire extraordinaire and busy bee, snatched that gem and rerouted it to, his yet-to-be-unveiled AI brainchild.

Now, this may not seem like a biggie, but consider this – domains with two letters, like, are like gold dust in the web world. They don’t come cheap. It’s been speculated that it could’ve gone for even more than the $3.8 million that sold for.

You might be wondering what is all about. Well, it’s still a mystery. Some bright minds are likely working on whatever wild idea Musk dreamed up. But since launching in July, they haven’t shown us anything to write home about.


IBM and NASA Open Source Largest Geospatial AI Foundation Model on Hugging Face

IBM, in partnership with NASA, has open-sourced its largest geospatial AI foundation model on Hugging Face. Built from NASA’s satellite data, this initiative will widen access to NASA’s earth science data, speed up climate-related discoveries, and overcome hurdles in analyzing large datasets.

In the fast-paced world of climate science, accessing the freshest data is tricky. While NASA’s data vault is expected to hold 250,000 terabytes by 2024, analyzing this massive pile isn’t easy. So earlier this year, IBM took up the challenge to build an AI foundation model for geospatial data under a Space Act Agreement with NASA. Now, by placing this model on Hugging Face—an open-source hub for AI models—they hope to democratize AI and spark new findings in climate and Earth science.

The model, trained by IBM and NASA using HLS satellite data and fine-tuned for flood and burn scar mapping, has shown a 15% performance boost over existing methods using half the labeled data. Further tweaking can gear the model for tasks like deforestation tracking, crop yield prediction, and greenhouse gas monitoring. Plus, there’s ongoing work with Clark University to adapt the model for other applications like time-series segmentation and similarity research.


Pinterest Reports Progress on Amazon Partnership and AI Investments

Pinterest spilled the beans on their Amazon deal and AI advances. Back in April, they announced a multi-year agreement with Amazon, marking their first dance with third-party ads. The good news? Things are moving along quicker than they’d planned, and their AI game is boosting engagement and the accuracy of ads across the board.

Pinterest isn’t just sitting pretty; they’ve bumped up their money-making supply by 30% year-over-year and keep users hooked with appealing content. This balancing act is made possible thanks to AI. By integrating “next-gen AI tech,” they’re serving up more personalized content, fine-tuning ad relevancy, and prompting users to act.

These AI wonders have been a real boon for Pinterest. They’ve seen an 8% global user growth, pushing the total to 465 million. Plus, with models 100 times larger than before, paired with their unique data, AI vision, and search tech, users are finding their content 10% more relevant compared to last year.


Twilio calls on OpenAI for generative AI

Twilio, the customer engagement platform, has announced its integration with generative AI leader OpenAI. The integration will see OpenAI’s GPT-4 model incorporated into Twilio’s Engage platform, a service that enables organizations to construct targeted and personalized marketing campaigns. The move is part of a larger initiative called Twilio CustomerAI, which was first previewed in June and aims to bring both generative and predictive AI to Twilio’s user community.

Twilio hosts over 10 million developers who create various types of customer engagement tools using the company’s APIs. This integration is an attempt to further leverage the generative capabilities of OpenAI’s GPT-4 to enhance the personalization and efficiency of customer engagement and marketing efforts on Twilio’s platform.

In addition to OpenAI, Twilio has announced a series of AI vendor partnerships, including Google, Frame AI, and AWS, to support the wider vision of Twilio CustomerAI.


GitHub Copilot can now tell developers when its suggestions match code in a public repository

GitHub has introduced a new feature to its AI-based coding tool, GitHub Copilot, which alerts developers when its code suggestions match with code from public repositories. The feature has been designed to offer developers more control and options when generating code, by presenting matching code snippets in a sidebar for developers to review and decide how to proceed.

Originally, Copilot had a blocking feature which automatically prevented suggestions of code that matched public code. According to GitHub, this feature was triggered less than 1% of the time, but it lacked the flexibility that some developers desired.

With the new code referencing feature, developers can opt to reject the code, use it directly (subject to licensing restrictions), or have Copilot rewrite the code so it doesn’t match the original anymore. GitHub is currently running a private beta of the new feature, which will eventually also come to Copilot Chat.


GoStudent adds another $95M to its war chest to go after VR and AI-enhanced tutoring

Edtech firm GoStudent has raised an additional $95 million in a strategic funding round, taking its total capital raised to $686.3 million. The latest investment included participation from Deutsche Bank, Left Lane Capital, DN Capital, Tencent, Prosus, DST, Coatue, and Softbank Vision Fund 2, and was a blend of equity and debt capital.

GoStudent, currently valued at around €3 billion, has used previous funds to acquire complementary businesses, such as the traditional tutoring company, Studienkreis. The new capital will be channeled into improving hybrid learning solutions in the DACH region (Germany, Austria, Switzerland), expanding the use of GoVR, GoStudent’s recently launched virtual reality language learning platform, and developing AI-driven tools.

One of these AI-driven initiatives is an “AI lesson plan generator”, which is trained on the local curriculum. The platform will aim to save tutors around 15 minutes per lesson, enhancing the efficiency of the tutoring process. The company currently has about 23,000 tutors on its platform.


Hailo unveils new edge AI accelerators, targeting entry-level and high-capacity applications

AI startup Hailo, based in Tel Aviv, is expanding its Hailo-8 product line with two new devices: Hailo-8L and Hailo-8 Century. These gadgets, which speed up AI processes on local machines, cater to a wide range of AI uses in areas like security, transport, retail, industry, and cars.

The Hailo-8L is for entry-level needs, and can do up to 13 trillion operations per second. It can also manage multiple real-time streams or several models and AI tasks at once. The Hailo-8 Century lineup, on the other hand, brings the Hailo-8 AI speed-up as cards that fit in a PCIe slot, providing 52 to 208 trillion operations per second. This version is perfect for heavy-duty jobs like smart vision systems or platforms that handle lots of video streams in real time.

Hailo also has the Hailo-15 vision processor in its lineup, a specialized chip that can be built into a camera to perform heavy-duty image processing while saving power. This market of built-in AI is predicted to grow from $9.4 billion in 2023 to $18.0 billion by 2028.


London Stock Exchange Group teams up with Microsoft to develop AI models

The London Stock Exchange Group (LSEG) is partnering with Microsoft and several banks to create tailored generative artificial intelligence (AI) models. This move reflects how the financial services industry is leveraging AI while ensuring their proprietary data stays secure.

Microsoft acquired a 4% stake in LSEG in December last year and also secured a board seat as part of a 10-year strategic partnership. They also invested $10 billion in OpenAI. Banks are keen on creating their own generative AI models to ensure their data doesn’t inform other language models.

LSEG, which has seen a shortage of listings in London, sees AI-related products as a potential new line of business, especially after its $27 billion acquisition of Refinitiv, a financial market data provider. 


Surgeons successfully restore touch and movement in quadriplegic man using AI brain implants

A quadriplegic man in the U.S., Keith Thomas, has regained his touch and movement abilities due to successful brain implant surgery, a world first. Surgeons implanted microchips into his brain, which, coupled with artificial intelligence, interpret and translate his thoughts into action.

Thomas had broken his neck in an accident and became paralysed from his chest down. He then joined a clinical trial at the Feinstein Institutes for Medical Research in New York, where his brain was first mapped using MRIs to locate the areas responsible for arm movement and hand touch sensation. Following this, he underwent a 15-hour open-brain surgery.

Dr Ashesh Mehta, the surgeon who performed the operation, described the procedure as a “double neural bypass”. They bypassed the broken pathways where electrical signals are sent between the brain, body, and spinal cord. A computer reads Thomas’ thoughts, translating them into hand movements. Moreover, this bypass works both ways, enabling Thomas to feel sensations through tiny electrodes instead of the usual neurons responsible for feeling in his fingertips.


From smart stethoscopes to predicting bed demand: how AI can support healthcare

A collection of studies published by the UK’s National Institute for Health and Care Research (NIHR) reveals the potential benefits of applying artificial intelligence (AI) in healthcare.

1. Heart Disease: AI technology is assisting in diagnosing heart failure. A smart stethoscope that employs AI can accurately detect heart failure in primary care settings nine out of ten times. In addition, AI combined with routine clinical data and blood tests can determine whether patients visiting A&E have experienced a heart attack.

2. Lung Cancer: AI applications can accurately identify cancerous growths on CT scans. These AI tools perform better than the standard Brock score, potentially allowing for earlier detection of lung cancer and saving lives.

3. Disease Progression: AI can predict whether patients with wet age-related macular degeneration (AMD) in one eye will develop the condition in the other eye. AI has also been used to develop a tool that predicts the risk of flare-ups in people with ulcerative colitis, potentially speeding up assessments and providing accurate prognosis information.

4. Personalised Treatment: AI can help determine the best drug combinations for cancer patients within 12 hours, potentially improving outcomes. An AI tool can predict the risk of death in the month after surgery for people who have COVID-19 at the time of their operation.

5. Hospital Pressures: AI can accurately predict which individuals do not need to attend emergency departments and help managers forecast the demand for beds. The predictions outperformed the hospital’s own emergency admissions planning.

These studies exemplify how AI can address important health challenges. However, Dr. Jemma Kwint, a senior research fellow at the NIHR, stresses the need for further research to fully understand how these AI tools could work in routine clinical practice, their long-term impact on patient outcomes, and their overall cost-effectiveness.