Good morning! This is your daily ☕️ Techpresso.
In today's Techpresso:
🤖 First artwork by robot sells for $1M
💰 Amazon wants Anthropic to switch from Nvidia to Amazon chips
🛡️ Anthropic and Palantir bring AI model Claude to US defense sector
💬 Apple finally lets green bubbles send message reactions
❌ TSMC is cutting China off from its advanced AI chip
🎁 + 8 other news you might like
🔮 + 1 handpicked research papers and tools
🤖 First artwork by robot sells for $1MLINK
A portrait created by the humanoid robot artist Ai-Da was sold for $1.08 million at Sotheby’s auction, surpassing estimates of $120,000 to $180,000.
Ai-Da, the world's first ultra-realistic robotic artist, used her advanced artificial intelligence to conceptualize and paint the portrait of British mathematician Alan Turing.
The artwork is intended to spark discussions about the implications of artificial intelligence, as it reflects Turing's historical concerns about the ethical use of technology.
💰 Amazon wants Anthropic to switch from Nvidia to Amazon chipsLINK
Amazon is negotiating a multi-billion-dollar investment with Anthropic, aiming to replicate a similar deal from the previous year.
The investment comes with a condition that Anthropic must increase the use of Amazon's Trainium chips for AI model training instead of relying on Nvidia chips.
This proposed transition could pose both technical challenges and limit Anthropic's ability to partner with other cloud service providers or manage its own data centers.
🛡️ Anthropic and Palantir bring AI model Claude to US defense sectorLINK
Anthropic, Amazon Web Services, and Palantir have formed a partnership that provides US intelligence agencies access to Anthropic's Claude AI models, enhancing the connection between AI companies and the defense sector.
The collaboration operates through Palantir's AI Platform, utilizing AWS infrastructure, and since early November, these AI models have been accessible in Palantir's highly secure Impact Level 6 environment."
This cooperation is part of a larger industry move, with companies like Meta and OpenAI pursuing closer links with defense partners, and Anthropic negotiating a new funding round possibly valuing it at $40 billion.
💬 Apple finally lets green bubbles send message reactionsLINK
Apple has updated its system so that Android users' emoji reactions appear in line with iOS message bubbles, enhancing cross-platform communication.
The improved reaction feature was confirmed by testing messages between iPhones with iOS 18.1 and various Android devices, though it's unclear when and how the change was implemented.
Apple's adoption of the RCS standard, including emoji reaction support, comes after external pressures, but some features like message editing still remain non-functional on iOS.
❌ TSMC is cutting China off from its advanced AI chipLINK
TSMC will halt production of advanced artificial intelligence chips, specifically 7 nm or smaller, for Chinese customers starting Monday, as impacted by US policies.
Future supplies of these semiconductors to Chinese customers will require an approval process, potentially involving examination by Washington authorities.
The decision may significantly impede the AI ambitions of Chinese companies like Alibaba and Baidu, as well as emerging AI start-ups relying on TSMC for chip manufacturing.
Other news you might like
Apple botched the Apple Intelligence launch, but its long-term strategy is sound.LINK
Intel says it's bringing back free office coffee to boost morale after a rough year.LINK
Mistral launches a moderation API.LINK
OpenAI scores key legal victory as judge throws out copyright case brought by news websites.LINK
Man sick of crashes sues Intel for allegedly hiding CPU defects.LINK
Police Freak Out at iPhones Mysteriously Rebooting Themselves, Locking Cops Out.LINK
Hacker says they banned ‘thousands’ of Call of Duty gamers by abusing anti-cheat flaw.LINK
Detroit residents will be able to pay taxes with crypto from mid-2025.LINK
Latest research and tools
LoRA vs. Full Fine-Tuning: An Illusion of Equivalence: this paper reveals that while Locally Linear Rule-based Approximations (LoRA) and full fine-tuning methods may seem similar in performance for language models, they actually function differently and have distinct advantages and limitations.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}
