Google’s Project IDX, a cloud-based workspace for full-stack multi-platform development, has added support for various languages and frameworks, including Astro, Go, Python + Flask, Preact, and Node.js. Developers can now preview their Flutter mobile apps in an iOS simulator or Android emulator. Project IDX now has simplified Docker container management and supports running command-line tools within the workspace and auto-detecting network ports.
2024.01.29·developers.googleblog.com·4 min read·TLDR Web Dev Txtai is an all-in-one embeddings database for semantic search, language model orchestration, and various language model workflows. It incorporates vector search with SQL, object storage, topic modeling, graph analysis, and multimodal indexing. The project can be implemented using Python or YAML and offers API bindings for various programming languages.
In the first four weeks of 2024, nearly 100 tech companies, including Meta, Amazon, Microsoft, Google, TikTok, and Salesforce, have collectively laid off about 25,000 employees, despite their stock prices returning to pre-pandemic levels and being highly profitable. These layoffs seem to be aimed at satisfying investors and boosting stock prices. They are driven by copycat behavior and may not actually be necessary.
GitHub Copilot usage has led to an increase in AI-assisted code, which in turn has created a downward pressure on code quality. Code written with AI is often harder to maintain and of lower quality as it’s often verbose or copy-pasted. However, GitHub Copilot has had a positive effect on developer productivity and satisfaction.
Technical design docs (or tech specs) are important because they reduce risk, document decisions, and help align stakeholders on technical choices.
2024.01.29·careercutler.substack.com·8 min read·TLDR Web Dev Startups today are much more polished compared to the “MVP” startups of the past. However, commenters agree that while startups are more visually polished on the outside, it may not mean that they are necessarily as polished on the inside (on the application level).
Git commit messages are often useless and unread. A better approach for teams would be to adopt modern trunk-based engineering practices, where pull request branches are squash-merged into the main branch, making individual commit messages irrelevant.
Rust's slow build times have been a common complaint among developers, but new tooling is able to optimize compilation.
Sponsored
Spam starts with your personal data being sold. Data brokers are profiting from your home address, SSN, phone number, and other personal information that doesn’t belong to them. Incogni [5] scrubs your personal info from the web. Simple as. It takes 3 minutes to set up [5], tackles 180+ data brokers’ and people-search sites, and there’s a 30-day money back guarantee [5].
2024.01.26·get.incogni.io·Sponsor·TLDR Apple has announced plans to change rules for developers releasing iOS software in the EU. Third-party app stores will be allowed on iOS for the first time. The changes will arrive with iOS 17.6 in March. Marketplaces will have to go through Apple's approval process and can be set as the default on devices. Developers can choose to use Apple's payment services or integrate third-party systems. All apps must still be notarized by Apple, and distribution through third-party marketplaces will still be managed by Apple's systems. Developers will only be allowed to distribute a single version of their app across different app stores.
2024.01.26·theverge.com·5 min read·TLDR OpenAI is launching new models, reducing prices for GPT-3.5 Turbo, and introducing new ways for developers to manage API keys and understand API usage. The new models include two new embedding models, an updated GPT-4 Turbo preview model, an updated GPT-3.5 Turbo model, and an updated text moderation model. The updated GPT-4 Turbo preview model completes tasks like code generation more thoroughly than previous preview models and is intended to reduce cases of 'laziness'. OpenAI plans to launch GPT-4 Turbo with vision in general availability in the coming months.
2024.01.26·openai.com·8 min read·TLDR A team at MIT has announced a new method of 3D printing metal that prioritizes printing speeds and scale over resolution. It is capable of printing large aluminum parts at least 10 times faster than comparable metal additive manufacturing processes. The technique utilizes a bed of 100-micron glass beads to create a structure into which molten aluminum is deposited. The beads can stand intense temperatures while allowing heat to quickly dissipate. While the final product doesn't offer near the pinpoint accuracy of some other methods, the technique's quick turnaround makes it a candidate for revolutionizing the way metal printing and metal forming are currently handled.
2024.01.26·techcrunch.com·2 min read·TLDR A new sprayable coating applied to standard air filters may help protect against airborne respiratory infections like measles, influenza, SARS, pneumonia, tuberculosis, and COVID-19. It provides a greater than 99% antiviral activity while also completely stopping bacterial growth. The spray doesn't interfere with the effectiveness of filters doing their normal job of removing other particles from the air. There are many other similar technologies in the works, but they will all need more development before they become commercially available.
2024.01.26·newatlas.com·2 min read·TLDR Sponsored
Datadog’s 2023 Container Report examined more than 2.4 billion containers [11] run by tens of thousands of customers. In this livestream, Datadog and Solomon Hykes [11] (CEO and co-founder of Docker and Dagger.io [12]) will discuss the findings and implications for serverless containers, GPU-based compute, resource optimization, container workloads, and more. Register free for Datadog’s January
2024.01.26·datadoghq.com·Sponsor·TLDR The initial version of the Ollama Python and JavaScript libraries are now available. These libraries make it possible to integrate new and existing apps with Ollama in a few lines of code. Ollama allows developers to get up and running with large language models locally. It supports a wide range of models, including Mistral, Dolphin, Llama2, and Orca.
2024.01.26·ollama.ai·2 min read·TLDR ViroReact is a library for building augmented reality and virtual reality experiences. It can run React Native code natively across all mobile VR and AR platforms. A page showing example projects is available.
2024.01.26·github.com·GitHub repository·TLDR iOS 17.4 will introduce a change that will allow alternative browser engineers to run on iOS. Developers will have to be authorized by Apple to switch engines, which involves meeting specific criteria and committing to a number of ongoing privacy and security mitigations. The change will only apply to people with EU-based accounts. EU users will be presented with a new choice screen for choosing a default browser when opening Safari for the first time.
2024.01.26·theverge.com·3 min read·TLDR Tesla has backed off its bullish sales targets due to growing EV sales by rival brands. The company has been cutting prices for more than a year to boost sales. It lost the lead in global EV sales to Chinese automaker BYD in the fourth quarter of 2023. Elon Musk told analysts on a call that Chinese carmakers will have significant success outside of China. He believes that current trade barriers are preventing Chinese car manufacturers from demolishing most other car companies in the world.
Impacted employees will receive severance benefits informed by local employment laws.
2024.01.26·techcrunch.com·2 min read·TLDR Kener is a Node.js status page tool that offers a user-friendly interface that simplifies tracking service outages and improves on communication during incidents.
2024.01.26·github.com·GitHub repository·TLDR Developers who choose to use alternative app stores will have to pay a huge 'Core Technology Fee' designed to ensure that no second-party app store ever takes off.
2024.01.26·world.hey.com·2 min read·TLDR Certain iOS apps are exploiting a feature designed to allow apps to customize push notifications to send analytics information, circumventing the restrictions imposed by iOS on background app activities.
2024.01.26·macrumors.com·2 min read·TLDR HP sees its customers as investments that don't pay off if they use third-party products.
2024.01.26·pcworld.com·4 min read·TLDR AI labs should make sure that the safety measures they apply to models prevent AI from intentionally choosing actions that maximize their probability of eventually seizing power.
2024.01.26·lesswrong.com·33 min read·TLDR Hugging Face has partnered with Google Cloud to give easy access to models and compute for customers of both platforms.
2024.01.26·huggingface.co·8 min read·TLDR AI OpenAI released a new embeddings model, a better and cheaper GPT-3.5 Turbo, and a less lazy GPT-4 Turbo preview. It has improved pricing and tried to improve the GPT-4 Turbo preview model to prevent the laziness issue in code generation.
2024.01.26·openai.com·4 min read·TLDR AI The vector database company that powers portions of ChatGPT and X's “More like this” has raised money to improve its Rust-based vector store and expand its enterprise offerings.
2024.01.26·qdrant.tech·2 min read·TLDR AI Soft robotics is a challenge due to the large control space. This work introduces a simulator with various tasks for manipulating soft “ditto” like objects. It contains utilities, visualization, and some strong baselines.
2024.01.26·dittogym.github.io·8 min read·TLDR AI Researchers have developed a new method for creating scene graphs, making the process faster and more efficient. Their transformer-based technique focuses on improving the way the model understands and connects different elements in an image, leading to better performance on challenging tasks.
2024.01.26·arxiv.org·1 min read·TLDR AI SpeechGPT-Gen is a new speech synthesis model that efficiently handles the complexities of language and voice characteristics.
2024.01.26·github.com·GitHub repository·TLDR AI Sponsored
Microsoft is marketing its Fabric Lakehouse architecture as an ideal way to manage Power BI semantic models. This tech talk [12] will examine these claims in light of AtScale’s TPC-DS benchmark [13], measuring Power BI performance on Microsoft Fabric on 100GB, 1TB, and
LLM Steer is a library for controlling language model generation with sentiment vectors. This means you can embed a concept like “thoughtful” and make the generation more thoughtful. It tends to work better than prompting in general, but more study is needed.
2024.01.26·github.com·GitHub repository·TLDR AI RoMa is a library for PyTorch for dealing with tricky rotations. It is reasonably efficient and can be used to help with any project that uses 3D data.
2024.01.26·github.com·GitHub repository·TLDR AI AgentBoard is a benchmark designed for multi-turn LLM agents, complemented by an analytical evaluation board for detailed model assessment beyond final success rates.
2024.01.26·hkust-nlp.github.io·3 min read·TLDR AI A walkthrough with code on writing a sparse Mixture of Experts model from scratch in Pytorch. There are diagrams and explanations for each step. The explanation of top-k routing is particularly insightful.
2024.01.26·huggingface.co·22 min read·TLDR AI A survey of experts on how early adopters are using AI today and where AI is going in 2024.
2024.01.26·davidtsong.substack.com·5 min read·TLDR AI Image similarity systems give a score based on how similar two images are. This work improves upon previous methods, primarily by relying on synthetic data and human preferences.
2024.01.26·dreamsim-nights.github.io·8 min read·TLDR AI SegMamba is a model designed for 3D medical image segmentation. It offers a more efficient alternative to the Transformer architecture.
2024.01.26·arxiv.org·4 min read·TLDR AI Track keywords across social media to convert leads with AI-generated auto-replies.
Researchers have developed a Shared Feature Calibration (SFC) method to enhance semantic segmentation.
2024.01.26·github.com·GitHub repository·TLDR AI The US government is reportedly planning to sell about $132 million in Bitcoin associated with the infamous Silk Road seizures. This includes Bitcoins connected to Ryan Farace, convicted for money laundering, and former Secret Service agent Shaun Bridges, who stole Bitcoins during the Silk Road investigation. The Bitcoins will be sold in lots of 2,800 BTC and 58 BTC.
Bitcoin ETFs have experienced a combined negative outflow for the first time since launching on January 11. This was driven primarily by large outflows from Grayscale's GBTC, which outpaced new investment in funds like BlackRock's IBIT and Fidelity's FBTC. Despite noticeable drops in inflows, both BlackRock and Fidelity still hold more than 40,000 bitcoins each and are nearing $2 billion in AUM.
Colosseum, a new organization co-founded by Matty Taylor, Clay Robbins, and Nate Levine, was launched to host Solana's hackathons, provide accelerator programs for winning founders, and invest in next-gen crypto businesses. The program is comprised of online hackathons that encourage global innovation, a dedicated accelerator program to nurture hackathon winners, and a venture fund that offers $250,000 pre-seed investments to teams in the Accelerator program. Participants present their projects during a Demo Day pitching session.
A new crypto project called WEN is being rolled out by the Jupiter team. It includes a new NFT standard, a token, and a launchpad project. WEN was initially created as an art-driven poem acknowledging an oft-repeated question in crypto - WEN token? The poem was then turned into an NFT and a new lightweight NFT standard was also introduced so that it could be fractionalized. The WEN token will represent portions of ownership of the WEN NFT.