Google has unveiled Gemini 2.0, the next generation of its AI model family. Gemini 2.0 Flash, the smallest model of the family in terms of parameter count, launches today through Google's developer platforms. It can generate text, images, and speech while processing multiple types of input including text, images, audio, and video, but its image generation and text-to-speech features will remain limited to early access partners until January 2025. Google plans to integrate the technology into products like Android Studio, Chrome DevTools, and Firebase.
The visionOS 2.2 update upgrades Mac screen mirroring on the Apple Vision Pro with a curved and sharper display. There are now three screen options - Standard, Wide, and Ultrawide. The Ultrawide mode can display resolutions of up to 10240 x 2880. The update makes the Vision Pro more than a personal movie theater.
Self-driving cars are now ready for prime time, with robotaxis leading the charge. There is evidence that the industry is making progress on overlapping technologies that make adoption possible. This article looks at the adoption of self-driving cars, robotaxis, public perceptions of the technology, and the future of self-driving cars. The regulatory environment will determine how long it takes for markets to adopt self-driving technology. It will still likely be a couple of decades before consumers are buying self-driving cars in meaningful numbers.
Google Quantum's claims should be approached with caution, particularly those of an extraordinary nature. Its supremacy claims in the past have been refuted by several groups. This article looks at Google's 2019 experiment to review its claims of quantum supremacy. The methodology used in the experiment was flawed and the competition in the space is a serious additional reason to not take Google's assertions seriously.
The circuit that lets junior developers grow into senior developers is broken, or at least damaged. Many tasks can now be automated, so companies don't need as many junior contributors. Developers who don't adopt chat-based programming as their primary modality will be left behind.
ucacher is a CLI tool that tracks which files are accessed by commands, skips unnecessary executions, and restores cached outputs when possible. It brings faster feedback loops, simplified caching and skipping, enhanced precision, resource efficiency, scalability for complex workflows, and more. Using ucacher can result in significantly reduced execution times.
People contributing to open source are under no obligation to stay in a project. The best thing to do is to make sure projects can continue to operate smoothly even after you leave. The license maintainers choose will influence how their project is utilized and how it develops, so maintainers should fully understand the implications of each license before making a decision. It's okay to step away from projects if they don't make you happy.
The Pixel 9a is expected to be released in March next year for $499. It will be powered by a Tensor G4 and use the Titan M2 security chip. The 6.3-inch display, which will use Gorilla Glass 3, will have a peak brightness of 1,700 nits (or 2,700 nits for HDR content) with a 120Hz refresh rate. The Pixel 9a will have Android 15 out of the box and seven years of updates.
Apple and Broadcom are working on a server chip designed especially for artificial intelligence internally code-named Baltra - it is expected to be ready for mass production by 2026.
Some of Microsoft's biggest rivals say that its deal with OpenAI harms competition by saddling rivals with extra costs and blocking them from hosting OpenAI's latest models themselves.
Former PlayStation Studios boss Shawn Layden doesn't think there will ever be a generational improvement in console performance matching that of the PS1 to PS2 ever again.
A federal district judge in California has ordered Matt Mullenweg and his companies to stop interfering in WPEngine's contacts with its customers and to restore its access to the wordpress.com platform.
Amazon Autos is an e-commerce business that lets customers find, order, and buy new cars, trucks, and SUVs from dealerships. It launched on Tuesday with Hyundai in 48 US cities. Amazon Autos will allow customers to secure financing and e-sign paperwork. It will also handle trade-ins. Customers can only buy new cars. Amazon will add more cities, additional auto manufacturers, leasing, and expanded financing options in 2025.
Cruise CEO Marc Whitten announced that GM would no longer fund the company in a Slack message on Tuesday afternoon. Cruise will be absorbed into GM and combined with the automaker's own efforts to develop driver assistance features and eventually fully autonomous personal vehicles. It is unclear whether employees' jobs are safe. Several employees reported feeling blindsided by the decision - they had been following a roadmap to launch a driverless service in Houston in 2025.
Google Willow is a new 105-qubit superconducting chip. It has been used to demonstrate an error-corrected surface code qubit as well as a new, bigger quantum supremacy experiment based on Random Circuit Sampling. The development is a real milestone for the field. This article discusses Google Willow and its implications for the field of quantum computing.
The universe is expanding about 8% faster than initially assumed possible. New measurements from the James Webb Space Telescope suggest that the long-established discrepancy between the observed expansion rate of the universe and the rate we thought possible based on our knowledge of the cosmos is potentially due to the influence of dark matter and dark energy. The discrepancy suggests that something is missing in our understanding of the early universe. The new data confirms that the universe is expanding faster today than it did during its first few billion years.
Fragmented processes and point automation solutions often hide inefficiencies, restrict scalability, and block data-driven insights. Process orchestration helps businesses transform this complexity into opportunity. This guide by Camunda [9] covers actionable steps, best practices, and tips for modern process orchestration + access to bonus
JSON is famously simple, but its formal definition is quite complicated. The biggest problem with JSON is the incredible diversity of parser behavior and non-conformity across and within language ecosystems. Everything is implementation-defined, which can result in data loss and other issues. This article looks at how JSON is designed, how it is used and misused, and how its 'human readability' features can cause headaches for users.
Limbo is an in-process OLTP database management system. It features an in-process OLTP database engine library, asynchronous I/O support with io_uring, SQLite compatibility, and JavaScript/WebAssembly bindings. Limbo is still a work in progress.
WhatsApp contains an ecosystem of services, features, and connections that make many businesses possible. It is the world's most widely-used messaging app, with 2 billion daily users sending more than 100 billion messages a day in 60 languages across 180 countries. WhatsApp achieved its global dominance in large part by enabling cheap, private, and reliable messaging on almost any phone almost anywhere in the world. The app has transformed into a sort of 'everything app' in the decade since Meta acquired it. This article takes a look at how the app is used in different places and the possibilities it has enabled.
The creator of Redis is returning to the project after a four-and-a-half-year break. He will be working with Redis' new CEO to tune the future of the company's relationships with the community and the code base direction. He will help be a bridge between the company and the community and also help produce programming demos, invent and describe new patterns, and create documentation, videos, and blog posts about new and old stuff.
Instagram's trial reels will give creators a chance to test out new content by publishing videos that don't appear to their followers, allowing creators to experiment with new ideas and see what performs best without having to worry about how their followers may react.
This article features an interview with Annie Rauwerda, the owner of the famous Depths of Wikipedia account, which shares factors from some of the most arcane, interesting, and surprising pages on Wikipedia.
This article looks at two recently unveiled chiplet-enabling technologies and examines the benefits of chiplet-based designs and some reasons why they may not make sense.
Was your last call handled by an UNCANNILY competent rep? Answered instantly, understood every word, and delivered precisely the info required? You're probably one of the MILLIONS speaking to Bland's AI call agent daily [4].
The Tesla Model Q will launch in the first half of 2025 and cost less than $30,000 with subsidies, or $37,499 if Trump cancels the IRA tax credit. It will be built on Tesla's existing lines. Tesla told investors during its Q3 earning calls that it planned to develop a more affordable model in H1 2025. The Model Q will be 15% smaller and 30% lighter than the Model 3. It will have a steering wheel and pedals but will be geared toward autonomous driving functionality.
OpenAI has released Sora, its AI video-generation model. Sora will be available to US users and most countries internationally but OpenAI has no timeline yet for launching the tool in Europe and the UK. It will be included in existing ChatGPT Plus and Pro subscriptions. Sora is aimed at competing with video-generation AI tools from other companies like Meta, Google, Amazon, and Stability AI.
Blue Origin says it's on track to launch New Glenn before the end of the year, but it is still waiting on regulatory approval to conduct a final key test of the rocket. The hot fire test will involve powering up all seven of the first stage's BE-4 engines and firing them to full thrust while the rocket sits on the pad. It is designed to mirror how the vehicle will operate during lift-off. The 320-foot-tall rocket will be capable of carrying 45 metric tons to low Earth orbit.
This post shares useful evaluations for large language models that focus on simple, common tasks like classification/extraction, summarization, and translation. The post also discusses the role of human evaluation and how to calibrate the evaluation bar to balance between potential benefits and risks. It was written to be accessible for those without a data science or machine learning background.
TorchGeo is a PyTorch domain library that provides datasets, samplers, transforms, and pre-trained models specific to geospatial data. It makes it simple for machine learning experts to work with geospatial data and for remote sensing experts to explore machine learning solutions. TorchGeo has plenty of tutorials and documentation.
Jobs in financial services companies are more desirable than they have ever been. Banks say they are still aggressively hiring. Compensation is similar for comparable roles across tech and banking. However, tech companies are no longer rolling out the red carpet for their employees. Banks are increasingly looking to publicize their AI work, invest in pure research departments, and focus on upskilling and internal development.
The idea that $300k/year engineers are doing next to nothing as a norm in tech is damaging and ridiculous. The myth misleads aspiring engineers, fuels layoff culture, and damages the industry's reputation. While it is true some employees coast, glorifying this idea only perpetuates toxicity. Engineers should instead focus on creating realistic expectations, better leadership, and healthier work cultures.
Put next-level concepts at your level with Brilliant. Learn from thousands of bite-sized, interactive lessons in math, logic, programming, AI, and more. 30 days free + 20% off Premium for TLDR