 |
VOOZH | about |
|
VOOZH | about |
We’re so glad you’re here. You can expect all the best TNS content to arrive Monday through Friday to keep you on top of the news and at the top of your game.
Check your inbox for a confirmation email where you can adjust your preferences and even join additional groups.
Follow TNS on your favorite social media networks.
Become a TNS follower on LinkedIn.
Check out the latest featured and trending stories while you wait for your first TNS newsletter.
The stable release of Svelte 4 came out this week, a mainly maintenance release that sets the stage for Svelte 5, according to a blog post announcing the release. The latest release takes advantage of some improvements in Node.js and browser APIs, the post noted.
The new release also results in smaller and faster hydration code, the post stated. For example, on kit.svelte.dev the JS generated across the whole site was reduced in size by 12.7%, or from 126.3 kB to 110.2 kB.
Svelte 4 reduces the Svelte package size by nearly 75%, down to 2.8 MB, which the team pointed out means less time waiting on npm install.
“This improvement will be especially noticeable for users who are loading our interactive learning experience on learn.svelte.dev for the first time, users of the Svelte REPL, and users with limited connectivity,” the post stated. “The majority of the remaining package size is eslint support, which necessitates distributing a CJS build, and once the eslint rewrite is completed the Svelte package size can drop by over another 50%.”
In addition, the number of dependencies has been reduced from 61 to 16, which means faster downloads as well as less susceptibility to supply chain attacks.
The plan was to switch the underlying code from TypeScript to JavaScript (yes, you read that right — we triple-checked), but that seems to have shifted to Svelte 5, since the post noted: “Svelte 5 will be a rewrite of the Svelte compiler and runtime.” The post added that Svelte 4 was mainly about setting the ground for these future improvements, by modernizing tooling and dropping support for some legacy technology versions, such as older bundlers.
Users of Fresh, a new full-stack web framework for Deno, have apparently been questioning whether the Deno team is maintaining it: Fresh creator Luca Casonato announced this week that Marvin Hagemeister will take over as the Fresh project lead full-time, replacing Casonato.
“… there’s been one elephant in the room — is Fresh something that the Deno team is actually committed to maintaining? When you’ve asked, we’ve always said “Yes!”, but reality was more complicated,” Casonato wrote. “We started April with over 60 open (and unreviewed) pull requests on the Fresh repo — we were not keeping up with maintenance to the level you’re used to from the Deno runtime project.”
Hagemeister is a maintainer of React, the builder of Preact DevTools, and “speeder upper of the JavaScript ecosystem,” for which he increased the speed of npm scripts from 400ms overhead to 22ms, Casonato added.
Casonato also digs into the new features of Fresh 1.2, which includes:
Figma held its annual Config conference this week in San Francisco, and the focus was bridging the gap between design and development. Figma announced new products and features as part of the conference, including Dev Mode, a dedicated space for developers to work with designers on translating designs into code.
“With Dev Mode, designers and developers can work in different modes on the same designs in the same file, making it easier for developers to get the information they need and eliminating the need for version updates and extra communication,” the announcement stated.
Among the capabilities Dev Mode incorporates are:
Dev Mode is currently in open beta and free to try through 2023.
Amazon Web Services announced Thursday a new program to help customers build and deploy generative artificial intelligence (AI) solutions. AWS is investing $100 million in the program, called the Generative AI Innovation Center. The team includes strategists, data scientists, engineers and solution architects who will help customers interested in building an AI solution. AWS also is offering free workshops, engagements and training to help customers come up with ideas and scope use case for their business.
“Customers will work closely with generative AI experts from AWS and the AWS Partner Network to select the right models, define paths to navigate technical or business challenges, develop proofs of concepts, and make plans for launching solutions at scale,” the press release noted. “The Generative AI Innovation Center team will provide guidance on best practices for applying generative AI responsibly and optimizing machine learning operations to reduce costs.”
Engagements will also leverage a variety of AWS AI tools, including Amazon’s code prediction tool, CodeWhisperer; Amazon’s high-performance infrastructure; Amazon SageMaker for building, training and deploying models; and Amazon Bedrock, a fully managed service that makes foundational models from AI221 Labs, Anthropic, Stability AI, and Amazon Titan accessible through an API. They can also use SageMaker Jumpstart to deploy popular foundational models, such as including Cohere’s large language models, Technology Innovation Institute’s Falcon 40B, and Hugging Face‘s BLOOM.
Vector database company ActiveLoop is collaborating with Intel and AI learning community TowardsAI to launch free educational certification courses for developers on large language models; Deep Lake, a vector database for AI data; and LangChain, a framework for developing applications powered by language models. The certification program is part of the Deep Lake Foundational Model certification program.
The first installment of the course combines a comprehensive overview of foundational model theory and practical projects with LangChain and Deep Lake, equipping AI developers with the tools they need to apply large language models across industries.
The course is designed to be a practical one, to help engineers start AI in their own company processes or use foundational models to build entirely new products, said Louie Peters, CEO of Towards AI.
