Good morning! This is your daily ☕️ Techpresso.
In today's Techpresso:
🔮 Elon Musk launches the world's largest Nvidia supercomputer
💸 Canva says its AI features are worth the 300 percent price increase
⚖️ Brazil's supreme court panel unanimously upholds decision to suspend X
👀 Huawei's AI chips still lag far behind Nvidia's
🚗 Tesla releases A.S.S. feature for its S.3.X.Y electric cars
🎁 + 8 other news you might like
🔮 + 2 handpicked research papers and tools
🔮 Elon Musk launches the world's largest Nvidia supercomputerLINK
Elon Musk's company, xAI, has brought an AI training cluster named Colossus online, claiming it is the most powerful AI training system in the world.
Colossus, built using 100,000 Nvidia H100 GPUs, aims to help Musk catch up to Mark Zuckerberg's Meta in AI technology advancements.
Musk revealed that the cluster, established in Memphis, was completed in 122 days and will double in size within a few months as more GPUs are added.
💸 Canva says its AI features are worth the 300 percent price increaseLINK
Canva is significantly increasing the price for Canva Teams subscriptions by over 300 percent next year, citing the addition of generative AI features as the reason.
In the US, Canva Teams users will see their annual subscription costs rise from $120 to $500, but a discount will reduce it to $300 for the first year; Australian users will experience a similar steep increase in fees.
These new prices make Canva less affordable compared to its original position as a cost-effective alternative to Adobe, leading some users to plan on canceling their subscriptions in favor of Adobe applications.
⚖️ Brazil's supreme court panel unanimously upholds decision to suspend XLINK
A panel of five Supreme Court justices in Brazil has unanimously upheld the decision to block Elon Musk's social media platform X in the country.
The platform has been suspended since Saturday for failing to adhere to local regulations, according to the court's ruling.
Brazil's telecoms regulator noted that only Musk's Starlink has refused to comply with the suspension order imposed by the court.
👀 Huawei's AI chips still lag far behind Nvidia'sLINK
Customers of Huawei's AI chips have reported multiple performance issues and bugs, making the software difficult to work with compared to Nvidia's products.
Huawei has been sending engineers to assist companies in transitioning from Nvidia’s software to its own, but stability problems and crashes have hindered these efforts.
The company has faced significant challenges due to US export controls, impacting supply and leading to increased prices for its AI chips.
🚗 Tesla releases A.S.S. feature for its S.3.X.Y electric carsLINK
Tesla has released a new feature called Actually Smart Summon, abbreviated as A.S.S., which enhances the existing Smart Summon functionality in its S3XY line of electric cars.
This upgraded feature allows owners to remotely command their Tesla vehicles to drive to their location or another specified area using the Tesla app, and relies on Tesla Vision rather than ultrasonic sensors.
Actually Smart Summon addresses the shortcomings of the original Smart Summon feature, which was known for issues like failing to detect obstacles, and offers a live camera feed to help guide the vehicle if needed.
Other news you might like
Starlink in Brazil's crosshairs as spat over Elon Musk's X escalates.LINK
Bluesky added 2 million new users in a week following Brazil's X ban.LINK
Tesla’s plans revealed for a new variant of the Model Y EV.LINK
iPhone SE 4 tipped to complete Apple's adoption of OLED technology across all iPhones.LINK
Meta Quest 3S Update Suggests Affordable Headset Is Coming Soon.LINK
A third of all generative AI projects will be abandoned, says Gartner.LINK
Huawei wants to steal Apple's thunder with a tri-folding phone.LINK
Clearview AI fined by Dutch authorities for ‘illegal’ facial recognition database.LINK
Latest research and tools
AFRIM: a modern input method engine designed to facilitate typing in various African languages, ensuring the preservation of native dialects by using a universal phonetic-based approach, and offering support for desktop and web platforms.LINK
Smaller, Weaker, yet Better: Training LLM Reasoners via Compute-Optimal Sampling: the paper discusses how using compute-optimal sampling methods can lead to more efficient training of large language models with reduced computational resources.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}
