Good morning! This is your daily ☕️ Techpresso.
On this day, 81 years ago, researchers at the University of Pennsylvania secretly began the development of ENIAC, the world's first general-purpose fully electronic computer.
In today's Techpresso:
👀 OpenAI gives GPT-4 a major upgrade
💬 Quora's Poe now lets AI chatbot developers charge per message
🌐 Google updates and expands its open source Gemma AI model family
🔥 Intel unveils latest AI chip as Nvidia competition heats up
📱 WordPress parent acquires Beeper app which brought iMessage to Android
🤔 New bill would force AI companies to reveal use of copyrighted art
🎁 + 9 other news you might like
🔮 + 4 handpicked research papers and tools
👀 OpenAI gives GPT-4 a major upgradeLINK
OpenAI announced the general availability of GPT-4 Turbo with Vision to developers, enabling applications with both text and visual understanding.
This advanced model maintains the features of GPT-4 Turbo but introduces vision capabilities, allowing it to process and understand images.
Integrating vision with language, GPT-4 Turbo with Vision simplifies development processes for apps requiring multimodal inputs, fostering innovative uses like coding assistance, nutritional insights, and website creation from drawings.
💬 Quora's Poe now lets AI chatbot developers charge per messageLINK
Poe, a Quora-owned AI chatbot platform, introduced a new revenue model allowing creators to earn money by setting a price-per-message for their bots.
The revenue model aims to compensate creators for operational costs, fostering a diverse ecosystem of bots ranging from tutoring to storytelling.
This monetization strategy is initially available to U.S. creators, complemented by an enhanced analytics dashboard to track earnings and bot usage.
🌐 Google updates and expands its open source Gemma AI model familyLINK
Google has enhanced the Gemma AI model family with new code completion models and improvements for more efficient inference, along with more flexible terms of use. Three new versions of CodeGemma have been introduced, including a 7 billion parameter model for code generation and discussion, and a 2 billion parameter model optimized for fast code completion on local devices. Google also unveiled RecurrentGemma, a model leveraging recurrent neural networks for better memory efficiency and speed in text generation, indicating a shift towards optimizing AI performance on devices with limited resources.
🔥 Intel unveils latest AI chip as Nvidia competition heats upLINK
Intel introduced its latest artificial intelligence chip, Gaudi 3, highlighting its efficiency and speed advantages over Nvidia's H100 GPU and offering configurations that enhance AI model training and deployment.
The Gaudi 3 chip, which outperforms Nvidia in power efficiency and AI model processing speed, will be available in the third quarter, with Dell, Hewlett Packard Enterprise, and Supermicro among the companies integrating it into their systems.
Despite Nvidia's dominant position in the AI chip market, Intel is seeking to compete by emphasizing Gaudi 3's competitive pricing, open network architecture, and partnerships for open software development with companies like Google, Qualcomm, and Arm.
📱 WordPress parent acquires Beeper app which brought iMessage to AndroidLINK
Automattic, the owner of WordPress and Tumblr, has acquired Beeper, a startup known for its Beeper Mini app that attempted to challenge Apple's iMessage, for $125 million despite the app's quick defeat.
Beeper CEO Eric Migicovsky will oversee the merging of Beeper with Automattic's similar app Texts, aiming to create the best chat app, with the combined service expected to launch later this year.
The acquisition raises questions due to Beeper Mini's brief success and upcoming changes like Apple introducing RCS support to iPhones, but Automattic sees potential in Beeper's stance on open messaging standards and its established brand.
🤔 New bill would force AI companies to reveal use of copyrighted artLINK
A new bill introduced in the US Congress by Congressman Adam Schiff aims to make artificial intelligence companies disclose the copyrighted material used in their generative AI models.
The proposed Generative AI Copyright Disclosure Act would require AI companies to register copyrighted works in their training datasets with the Register of Copyrights before launching new AI systems.
The bill responds to concerns about AI firms potentially using copyrighted content without permission, amidst growing litigation and calls for more regulation from the entertainment industry and artists.
Other news you might like
Microsoft left internal passwords exposed in latest security blunder.LINK
Neuralink-rival Synchron prepares for groundbreaking brain chip clinical trial.LINK
Google partners with company behind Sports Illustrated's AI authorship scandal.LINK
GM’s Cruise robotaxis are back in Phoenix — but people are driving them.LINK
Kobo announces its first color e-readers.LINK
eBay introduces AI-powered 'shop the look' features.LINK
Grayscale CEO sees bitcoin ETF outflows reaching equilibrium.LINK
Apple Vision Pro users complain of black eyes.LINK
Dyson’s new AR feature shows where you have (and haven’t) vacuumed.LINK
Latest research and tools
go-mysql-server: a MySQL compatible database engine written in Go, offering a SQL engine and server that can query various data sources using the MySQL dialect, with Dolt as its primary production database implementation.LINK
XZ projects on GitHub: repositories for data compression tools are again accessible, maintained by Lasse Collin (Larhzu).LINK
Evaluating faithfulness and content selection of LLMs in book-length summaries: the paper assesses how well language models summarize long texts without losing important information.LINK
xz-utils: a software package that has had a backdoor removed.LINK
Want to get the latest news differently? Find us on:
See you tomorrow for a new dose of ☕️ Techpresso!
-min.png){kind=small}
.png){kind=small}
