After years of promise, a massive shift is under way.
One day in late November, I cradled a red Samsung flip phone in my hands as if it was a ruby gemstone.
Tesla’s FSD V12 wide release has started, with plans to reach higher capability by next year, including the ability to drive through fog and potentially achieve true robot taxi capabilities.
Questions to inspire discussion.
The truck, based on an Aebi VT450 Transporter, featured a 320-horsepower electric drive system with a 90-kilowatt-hour battery.
An AI chatbot deployed by a car dealership went off the rails after mischievous users discovered a cheeky exploit, in some cases tricking the bot into offering them the deal of a lifetime: brand new cars for chump change. It’s an amusing but cautionary tale on relying on AIs for front-of-house interactions.
The dealership, Chevy of Watsonville in California, used the chatbot to handle customers’ online inquiries, a purpose it was expressly tailored for.
Chris White, a software engineer and musician, was one such customer. He innocently intended to shop around for cars at Watsonville Chevy — until he noticed an amusing detail about the site’s chat window.
BYD, the Chinese electric vehicle maker, is on the verge of surpassing Tesla as the global leader in EV sales.
China’s BYD is set to dethrone Tesla as the global EV sales leader, backed by affordable models and a robust charging infrastructure.
Tesla recorded $500M+ in gross profit from its Energy and Services (Supercharging) segments in Q3 2023. Elon Musk noted how strong energy gross margins were on the call, and insinuated strength in these businesses will continue. I think this is a super exciting development for Tesla investors as the company can smooth out cyclicality in it’s automotive business with consistent profits from its Energy and Services.
JAC Motors, a Volkswagen-backed Chinese automaker, is set to launch the first mass-produced electric vehicle (EV) with a sodium-ion battery through its new Yiwei brand. Although sodium-ion battery tech has a lower density (and is less mature) than lithium-ion, its lower costs, more abundant supplies and superior cold-weather performance could help accelerate mass EV adoption. CarNewsChina reports that the JAC Yiwei EV hatchback deliveries will begin in January.
Yiwei is a new brand in 2023 for JAC. Volkswagen has a 75 percent stake in (and management control of) JAC and owns 50 percent of JAC’s parent company, Anhui Jianghuai Automobile Group Holdings (JAG). The Chinese government owns the other half of JAG, making for one of the auto industry’s stranger pairings.
The Yiwei EV appears to be a rebranded version of the Sehol E10X hatchback (above), announced earlier this year. CarNewsChina describes the Sehol model as having a 252 km (157 miles) range with a 25 kWh capacity, 120 Wh / kg energy density, 3C to 4C charging, and a HiNa NaCR32140 cell. When JAC revealed the Yiwei brand in May, it said it would drop the Sehol label and rebrand all its vehicles to either JAC or Yiwei, leading us to this week’s EV reveal. JAC hasn’t yet said whether the Yiwei-branded model will keep the E10X moniker.
Tesla Giga Shanghai plans to release 2024 Model Y refresh as competition heats up in China.
People familiar with the matter told Bloomberg that Gigafactory Shanghai is already preparing to produce the 2024 refresh Model Y. Mass production on the new Model Y is expected to start by mid-2024.
The new Model Y units will be produced in the new phase of Gigafactory Shanghai. Production in the second phase of Tesla China’s factory will be suspended during New Year’s for upgrades. More updates to the Model Y assembly line will be made after the holidays.
The revolution in artificial intelligence may soon require more electricity than all electric vehicles combined.
Any activity that requires comprehension and production in one or more modalities is considered a multimodal task; these activities can be extremely varied and lengthy. It is challenging to scale previous multimodal systems because they rely heavily on gathering a large supervised training set and developing task-specific architecture, which must be repeated for every new task. In contrast, present multimodal models have not mastered people’s ability to learn new tasks in context, meaning that they can do so with minimal demonstrations or instructions. Generative pretrained language models have recently shown impressive skills in learning from context.
New research by researchers from Beijing Academy of Artificial Intelligence, Tsinghua University, and Peking University introduces Emu2, a 37-billion-parameter model, trained and evaluated on several multimodal tasks. Their findings show that when scaled up, a multimodal generative pretrained model can learn similarly in context and generalize well to new multimodal tasks. The objective of the predict-the-next-multimodal-element (textual tokens or visual embeddings) is the only one used during Emu2’s training. This unified generative pretraining technique trains models by utilizing large-scale multimodal sequences, such as text, image-text pairs, and interleaved image-text video.
The Emu2 model is generative and multimodal; it learns in a multimodal setting to predict the next element. Visual Encoder, Multimodal Modeling, and Visual Decoder are the three main parts of Emu2’s design. To prepare for autoregressive multimodal modeling, the Visual Encoder tokenizes all input images into continuous embeddings, subsequently interleaved with text tokens. The Visual Decoder turns the regressed visual embeddings into a movie or image.
