Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
This article contains thoughts from over 40 of a12z's partners on what they believe will drive innovation in 2024. The ideas cover technologies like smart energy grids, voice-first companion apps, programmable medicines, and AI tools for kids.
Liquid neural networks are much smaller than traditional AI models and require far less compute power to run.
Unidentified governments, including democracies allied to the United States, are surveilling smartphone users through app push notifications.
JSONB is a rewrite of the SQLite JSON functions - it can be several times faster than the original JSON functions.
The new iPad Air, iPad Pro, and MacBook Air models will all be released during the first several months of next year.
Sam Altman is TIME's 2023 CEO of the year.
This post proposes a hierarchy of needs for workers to refer to when picking a startup to join.
Google's flagship model Gemini is natively multimodal with text capabilities matching GPT4 while also being able to operate on many other data sequences. It also trained Alpha Code 2, which is an 85th percentile coder on code forces. Alpha Code 2 will be available in Google’s API after December 13th.
Meta has launched 'Imagine,' a standalone image generator with its own website featuring generative AI that creates images from text prompts.
Microsoft is enhancing its Copilot service with new features, including OpenAI's latest GPT-4 Turbo model, an upgraded DALL-E 3 model, a new code interpreter feature, and a 'Deep Search' function in Bing.
This paper uses natural language to describe what is different between two collections of images. This is a novel problem and a challenging one. The solution involves captioning and reranking the individual images and summarizing them with a language model.
Researchers have developed Diffusion-SS3D, a new method to improve semi-supervised 3D object detection using diffusion models. This approach involves adding noise to object size and class label distributions in 3D space and then using the diffusion model to denoise and generate better bounding box outputs.
This project introduces BenchLMM, a benchmark designed to test the robustness of Large Multimodal Models like GPT-4V and LLaVA against various image styles.
Apple quietly released a cool Array framework that is accelerated on Mac devices because of unified memory. It is simple and clean and provides some GPU support as well.
Stable Fast is an inference engine for HuggingFace diffusion models.
Switch your project to open source in just a few lines of code.
This study introduces the use of the Oogiri humor generation game to explore the Leap-of-Thought capabilities in multimodal LLMs. This approach challenges LLMs to exhibit non-sequential thinking, a skill crucial for generating creative and humorous responses to various types of multimodal information.
Lightspeed shares insights from Google and TitanML about the ins and outs of optimizing LLMs whether via prompting or fine-tuning.
While automation will likely be the way of the future, today, it’s all about augmenting your existing processes/self without having to rip and replace everything. Don’t fall into the trap of just augmenting yourself to do your job better - thinking through what and how to automate things and viewing your role holistically will allow you to find the best use of AI for your needs.
Segment Anything is a great model, but it is slow to run. Efficient Segment Anything is a new model that uses a number of optimizations to run significantly faster.
Right before leaving the company, Elon Musk implored OpenAI to move faster, a departure from his typically more skeptical comments toward AI.
Build in-app AI chatbots and AI-powered Textareas into React web apps.
A US House of Representatives committee has unanimously approved a bill directing the Department of Commerce to support blockchain tech. Despite the bill passing through committee, approval by the House remains uncertain. Its best chance may be integration with other larger, must-pass legislation.
Grayscale CEO Michael Sonnenshein stated that he remains optimistic about a potential bitcoin ETF approval following discussions with the SEC. Although the timeline has not been set, the firm believes that the SEC's questions suggest progress on their end. Along with apparent progress on BlackRock’s ETF application, many experts are expecting approvals to come within the next month.
Merit Circle DAO's tech stack and Sphere NFT marketplace, initially launched on Avalanche, will now also be available on Immutable's upcoming zkEVM network powered by Polygon. This expansion adheres to Merit Circle's chain-agnostic approach. The gaming-focused group is also contemplating investments in games and studios using Immutable.
Cryptocurrency derivatives exchange Aevo has introduced aeUSD, the first yield-generating, stables-denominated collateral asset. aeUSD allows users to generate 4.75% APY on their exchange margin while trading with 20x leverage. aeUSD, using an ERC-4626 asset model based on Aevo, provides yield by depositing into Maker DAO's Dai Savings Rate module, providing an innovative solution for enhanced capital efficiency.
This piece from a16z provides an outline of what to look forward to in 2024, including transformations in the crypto user experience, the rise of modular tech stacks, the integration of AI with blockchain, the implications of AI developments in gaming, and formal verification for software development. It also touches on the changing financial dynamics in gaming, the evolution of NFTs in branding, and the increasing use of SNARKs in verifying computational workloads.
This post from a firm that took a large position in Solana earlier this year lays out an in-depth thesis on Solana. The blockchain platform’s potential to scale and become a dominant player in the L1 race, despite major setbacks such as the collapse of FTX, is, according to this author, still undervalued.
This post makes an argument for monetizing protocols and why builders should consider this versus free-to-use protocols. The benefits of monetization include improved governance, as it allows a broader community to influence the protocol's future, while also preventing any one party from seizing control. Monetization can also help manage harmful players or applications by incentivizing compliance and deterring non-adherence.
This article takes a look at how blockchains are revolutionizing cross-border payments. Despite challenges in creating scalable consumer platforms, the demand for cross-border payments reached $190 trillion in 2023. While smaller, geo-specific solutions exist, regulatory confines limit their reach. With grassroots adoption in lower middle income countries, crypto is showing real promise as a killer use case to break through these borders.
The DOGE-1 moon mission, funded by Dogecoin and developed by Geometric Energy Corporation, received key NTIA approval, moving closer to a SpaceX Falcon 9 launch.
DePINs, decentralized networks combining infrastructure with token incentives, address key centralized infrastructure issues, offering advantages in capital formation, immediate cash flows, and value-growing tokens.
JPMorgan Chase CEO Jamie Dimon caused a stir yesterday as he claimed that if he were the government, he’d shut bitcoin and crypto down.
Binance founder CZ pled guilty to violating the Bank Secrecy Act, with sentencing scheduled for February 2023. He's been released on bond until then.
Discord announced on Tuesday that it's dropping a new mobile app designed from the framework up for mobile devices. The most noticeable change is a revamped layout that puts the Discord servers and messages on different tabs but maintains a vertical layout for servers that lets users quickly switch between them if they have multiple chats going on at a given time.
Visual Electric has launched an AI-based image generation tool designed specifically for designers. The tool focuses on non-linear creative workflows and offers features like mood selection, iterative design, and a community-sharing model, with plans to differentiate itself in a crowded market of generative AI tools.
Tristram Ingham, a physician and disability advocate with facioscapulohumeral muscular dystrophy, uses Apple's new speech accessibility features to preserve his unique voice as his condition progresses. Apple's Personal Voice and Live Speech, integrated into its latest operating systems, enable Ingham and others with speech impairments to maintain communication autonomy and identity.
Continuous design career growth involves a simple three-vertice system to help you engage with challenges at work. Complementary and timeless, they are about understanding context, developing a cadence of innovation, and articulating tools to drive impact.
The Government of India's regulations outlawed many dark design patterns in November 2023. A total of 13 dark design patterns are listed in the rules. This article explains them in depth.
In this discussion, two senior UX researchers explain how their teams are thinking about and using AI in UX research and offer tips on how UX researchers can navigate the hype and make informed choices.
A free customizable collection of SVG skeleton loader UI elements for design wireframes & mockups.
StyleX is the styling system behind products like Facebook, Instagram, and WhatsApp. It's a simple, easy-to-use JavaScript syntax and compiler for styling web apps.
This design system online calculator lets you understand and project cost savings by employee and component usage.
Users usually follow a predictable flow when interacting with your product, such as filtering information, analyzing it, deciding what to do with it, or storing bits of it in their memories. Understanding biases and heuristics affecting those four decision-cycle steps is vital to improving the user experience.
For creatives, prompting still feels very far from ideal. The creative process within any AI tool feels constrained when communicating ideas, unlike interacting with humans. tldraw, a web-based whiteboard app, provides a glimpse of what multimodal tools will look like, offering better support to creative endeavors.
A demo of a scroll-triggered sliced text animation where segmented elements reunite while scrolling.
A video on using AI to organize and process various content types, such as videos and audio, for enhanced user interfaces focusing on accuracy and efficiency in AI-generated content.
Iris Alba designed many book covers that are instantly recognizable for their avant-garde illustrations and playfully contrasting typefaces. Flecha Books' upcoming monograph examines the life and work of the little-known designer and art director whose work helped define modern Latin American publishing.