GitHub Copilot X is the next evolution of GitHub Copilot, an AI pair programmer that auto-completes code.
Copilot X offers chat and voice interfaces, support for pull requests, and can answer questions on documentation using OpenAI's newly released GPT-4 model.
Copilot for Pull Requests offers AI-generated descriptions for pull requests, adding AI-powered tags based on changed code, and offers developers a preview to modify them.
Copilot for Docs provides AI-generated responses to questions about documentation, starting with React, Azure Docs, and MDN.
Copilot CLI composes commands and loops, enables find flags for accurate queries, and helps with tab completions.
Copilot X reduces manual tasks, makes complex work easier, and helps every developer focus on building innovation.
Web fingerprinting is a method used to identify users across different browsing sessions and websites without using cookies or other persistent storage. It works by studying a user's web browser and hardware configuration, and it can generate a unique ID among millions.
Many websites use fingerprinting libraries to generate unique IDs, and some companies even sell fingerprinting as a service to other websites.
FingerprintJS is one such company, selling open-source JavaScript fingerprinting libraries to many websites. Even if a user clears their browser cache and all other site data, FingerprintJS can generate the same visitor ID and link it back to the user's previous visit.
Fingerprinting severely undermines user privacy, but there are ways to protect against it. Firefox with resistFingerprinting=true and Tor Browser are more secure against fingerprinting, while Brave browser randomizes fingerprints for each session. However, Brave is still based on Google's Chromium engine, which encourages Google's monopoly. VPNs do not help with fingerprinting, as they only mask IP addresses.
Adobe Firefly is a family of creative generative AI models coming to Adobe products that can be used to generate content for digital imaging, photography, illustration, artwork, graphic design, video, marketing, social media, and 3D modelling.
The generative AI will enable creators to use everyday language to generate extraordinary new content quickly, edit what they create using the tools that they already know and love, and generate custom vectors, brushes, and textures from just a few words or even a sketch.
Adobe is committed to developing creative generative AI responsibly with creators at the center, and as Firefly evolves, they will continue to collaborate with the creative community to build technology that supports and improves the creative process.
Firefly is unique because Adobe intends it to be more than an AI text-to-image generator.
AI generators like Firefly can enhance creativity by giving people new ways to imagine, experiment, and bring their ideas to life.
With generative AI, anyone can create videos, documents, and digital experiences, as well as rich images and art.
Firefly is built into Creative Cloud to give people the tools to both generate results quickly and customize them to fit their unique vision.
Content created with Firefly will contain a Content Credentials tag that states it was created using generative AI.
Currently, Firefly is trained on a dataset of Adobe Stock, openly licensed work, and public domain content where copyright has expired, but Adobe plans to explore ways for creators to train the machine learning model with their own assets.
Hyundai pledges to employ physical buttons in its products to come, making a design call rooted in safety concerns.
Sang Yup Lee, Hyundai's head of design, said that safety-related buttons have to be a hard key, and physical controls are a necessity for anything that could impact safety.
Hyundai will likely look at using touch controls more heavily when autonomous driving becomes mainstream.
DPReview, a long-standing photography website known for its camera reviews, news and forums, will be shutting down on April 10th due to a decision made by its parent company as part of an annual operating plan review.
Users are able to request a download of their content until April 6th, after which the data will be deleted.
The closure has left many users disappointed and calling on Amazon, the site's parent company, to reconsider or sell the site to someone who will keep it running.
Fzf is a Unix tool that provides fast and intuitive fuzzy search for files and directories in the terminal. It can also be used as a replacement for Ctrl+R function.
Fzf is customizable and can be used in tandem with other Unix tools such as Vi and Ripgrep for powerful searching capabilities.
The article provides examples of how to use fzf in various scenarios, including fuzzy searching every line in every file and opening a specific file using Vi.
Bob Metcalfe has been named as the winner of the 2022 ACM A.M. Turing Award for the invention, standardization, and commercialization of Ethernet, a technology that he developed 50 years ago.
The award is seen as the computing equivalent of the Nobel Prize and comes with a $1m prize, with financial support by Google.
Metcalfe's 1973 memo describing a broadcast communication network connected some of the first personal computers, which led to the creation of Ethernet.
After leaving Xerox, Metcalfe founded his own internet startup, 3Com Corporation, and continued to play a key role in the commercialization of Ethernet.
Today, Ethernet is the main conduit of wired network communications around the world, handling data rates from 10 Mbps to 400 Gbps, with 800 Gbps and 1.6 Tbps technologies emerging.
Mailchimp is reportedly banning users who have their browser's dev tools open when accessing the email marketing platform.
It's not clear how Mailchimp detects whether a user has the tools open or not, but one possible method could be to detect a sudden viewport resize, which happens when developer tools open to prepare the display of original sources of a website.
Some users speculate that Mailchimp may have detected and blocked the use of dev tools to view source code without consent.
Petals allows for collaborative running of large language models like BLOOM-176B, where users load a small part of the model and team up with others serving the rest for inference or fine-tuning.
Single-batch inference runs at 1 second per step, up to 10x faster than offloading, making it useful for interactive applications like chatbots.
Petals offers flexibility with PyTorch, allowing for fine-tuning and sampling methods, custom paths through the model, and access to hidden states.