Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
This article walks through how to build custom GPTs and provides 20 examples of how people are using them for work. Some of the use cases include conversing with personas, creating and analyzing experiment copy, and generating takeaways from raw survey results.
TikTok released a 34-page guide on its marketing strategy. The guide offers insights on how to enhance video content creation and maximize impact on the platform. It provides tips and links to TikTok tools to optimize your approach and leverage new platform features effectively. Suggestions range from maximizing the first 3-6 seconds of your videos to leveraging TikTok's latest automation and AI-powered tools. An email is required for download.
Some brands are shifting focus from traditional SEO tactics to mastering the social media landscape. Leveraging keywords, visual content, and engagement strategies on various platforms is essential. It can help companies boost organic traffic, improve brand reputation, and stay ahead in the ever-evolving digital ecosystem.
Most estimates of Total Addressable Market (TAM) lack analytical credibility or real-world applicability. Quoting a massive TAM with a top-down approach can hurt a company, especially if it’s taken from a third-party source without any analytical rigor. An overly simplistic calculation of price and sale quantity might seem better, but it still lacks the appropriate depth. To complete a more accurate TAM analysis, consider the insertion point of your product, the relative level of competition, the market’s maturity, and the time to capture.
This article discusses top-of-the-funnel strategies to spark brand loyalty and drive conversions in 2024. Companies can foster curiosity and desire in their target audiences by setting powerful branding and a clear content strategy. It’s important to develop a clear value proposition and make sure it is distributed through the right channels. The strategy must be constantly measured and optimized, which requires focusing on the right metrics depending on business objectives.
An analysis of the impact Super Bowl ads had on social media engagement, trending keywords, and web traffic for each brand.
Ethical permission-based marketing (EPBM) empowers users to define and control their relationship with a brand’s marketing communications.
Coach launched a campaign, “Find Your Courage,” that pairs virtual influencer ‘imma’ with celebrity ambassadors Lil Nas X, Camila Mendes, Youngji Lee, and Kōki. The campaign looks to engage younger consumers around the complexities of identity and self-expression.
Sponsored
This year’s POST/CON [5] (April 30 - May 1st in San Francisco) is the place to be if you’re looking to hear from industry leaders, gain new skills, engage with the community. (BTW, there’s also an awesome after party [5])
100 million people were using ChatGPT weekly by November 2023. Scaling ChatGPT to such a large user base was not an easy feat. OpenAI faced a few key challenges when scaling: managing GPU RAM, avoiding cache misses, finding the right batch size, measuring the right metrics, and handling a shortage of GPUs. The company found that taking a system-wide approach to optimization and being adaptable in the face of changing constraints allowed it to scale efficiently. Many of the lessons learned are applicable to general software systems.
Bloom Filters are a probabilistic data structures useful for checking membership in a set with efficient space usage. They store information using bits and hash functions and offer fast membership checks. They can give false positives (claiming an item exists when it doesn't), but they never give false negatives (missing an item that exists). In the real world, Bloom filters are commonly used for caching and distributed databases.
SPAs (Single Page Applications) are defined by a core feature: they dynamically rewrite their content as users interact with them instead of the default method of loading entirely new pages from the server. SPAs are able to provide this feature via the History API, a clunky API developed long before SPAs were the norm. The Navigation API is a newly proposed API that completely overhauls this space that promises to be easier to adapt. This page describes the Navigation API at a high level with code examples on popular usages.
Simple architectures can be surprisingly effective, even for large-scale applications. For example, Wave is a $1.7B fintech company with a CRUD app built on a Python monolith and Postgres database. The benefits of this simple architecture include lower cost, easier maintenance, and faster development. There are some areas where complexity is unavoidable, such as telecom integrations. Simple architectures are a good choice for many businesses, especially those that are not ad-supported or businesses without high traffic levels.
Picat is a research language that combines logic programming, imperative programming, and constraint solving. “Planning" in Picat is when the program finds variable mutations to reach a desired end state. Instead of defining algorithms, Picat defines starting states, actions, and final states, allowing the program to find solutions through mutations. Planning helps minimize cost and is best used for applications in AI, robotics, and video games.
Coding interviews are effective for companies despite their limitations because they help minimize the risk of hiring bad candidates. While there are downsides to coding interviews, like pressure and filtering out good candidates who don't perform well, they can be useful for assessing a candidate's problem-solving skills and their potential as a future teammate. Automated assessments aren’t as reliable as they lack human interaction and can’t always catch cheating.
Sponsored
For the past 40 years, MESCIUS (formerly GrapeCity) [12] has been building award-winning developer tools: JavaScript and .NET grids, reporting, spreadsheets, .NET and Java Document APIs, viewers, and mobile controls. These include popular solutions [12] such as ComponentOne, Wijmo, and ActiveReports. Microsoft, Tesla, Disney, and +250K other customers use MESCIUS tools daily.
Hyperdiv is a Python framework for developing reactive browser UI apps. It features built-in components, terse immediate-mode syntax, and minimal tool boilerplate. Hyperdiv supports Markdown, charts, styled web components, local storage, and forms.
Valve has made the Steam Audio SDK fully open-source, with plug-ins included. Steam Audio is an immersive audio solution for games and VR. It has real-time sound propagation, VR integration, and 3D audio capabilities. There are plug-ins to make it compatible with Unity, Unreal Engine, and other game engines.
Dax is a collection of cross-platform shell tools for Deno and Node.js. It allows users to export the shell’s environment to the current process. Dax has support for a global configuration. It brings extra features to Windows that are natively available on Unix-based systems.
Cloud storage performance lags far behind the advancements in SSD technology. Major cloud vendors are offering read/write speeds 10x slower than high-end SSDs available on the market. The reasons for this big difference are unclear but could be due to concerns about SSD durability, lack of demand, or potential disruption to existing cloud storage services.
This author spent a year cleaning medical device data from the FDA's 510k database website. The website's interface was outdated and lacked functionalities like fuzzy search and keyboard shortcuts. To save time and improve their workflow, the author wrote userscripts to automate repetitive actions by adding shortcuts for opening search, focusing on search input, and copying device ID.
This article provides twenty examples of how people are using custom GPTs in the workplace to make their teams more productive. It walks through how to create a custom GPT and demonstrates some interesting real-world use cases, such as figuring out internal ownership and technical dependencies, learning from past user research studies, and grading the relevance of search results.
Someone ran a script on BigQuery for historical HTTP Archive data and racked up a $14,000 bill for accessing 2.5 petabytes of data.
Various iOS and macOS developers chime in on the value of Objective C. Many believe it’s still useful to know Objective C even if their apps are primarily written in Swift.
Prettier, a code formatter, never breaks JavaScript/TypeScript code during formatting. This durability comes from its use of prior research by a prominent Haskell developer.
Node developers are wrestling with the decision to enable Corepack (a tool for managing package managers that lets you use npm/yarn/pnpm without having to install them) by default. This has sparked a debate about the possibility of even removing npm from the Node.js binary.
ChatGPT has seen declining web traffic in five of the past eight months. It is currently down 11% from its May 2023 peak. Its mobile app has seen fewer users than Snapchat added last quarter alone. The competition in the AI space means that OpenAI will have to pump out hits to stay ahead.
The second generation of Apple's Vision Pro headset is at least 18 months away from launch. Apple is keenly interested in feedback from customers who choose to return the first-generation headset as it will use the information to perfect the next version of the device. The Vision Pro will likely follow a refresh cycle longer than that of the iPad. There are rumors that Apple may launch a more affordable Vision headset around 2025.
The rate of information production is increasing faster than the storage density of tape. Using current technologies will mean that, in the coming decades, we will need exponentially more magnetic tape, disk drives, and flash memory - as well as the factories to produce the storage media. DNA is cheap, readily available, and stable at room temperature for millennia. Organizations around the world are already taking the first steps towards building a DNA drive that can both read and write DNA data.
MIT researchers have developed an antitampering ID tag that is tiny, cheap, and secure. It is several times smaller and significantly cheaper than the traditional radio frequency tags that are used to verify product authenticity. The tags use glue containing microscopic metal particles. This glue forms unique patterns that can be detected using terahertz waves. The system uses AI to compare glue patterns and calculate their similarity. The tags could be used to authenticate items too small for traditional RFIDs.
htmz is a minimalist HTML microframework for creating dynamic web user interfaces. It allows developers to load HTML resources within any element in a page. htmz only requires an inline HTML snippet to work - there are no dependencies and no backend is required.
This Reddit user recently put two complex codebases into GPT-4-Turbo-128K and Gemini 1.5 and asked them questions. The codebases implemented a parallel inet runtime, so it involved some hard compiler stuff. Gemini 1.5 completely outperformed GPT-4-Turbo-128K in the task of understanding the codebases. Most of the questions that GPT-4 got wrong are ones it would have answered correctly in a smaller context. Gemini still struggled to create a complete mental model of the system, but it is extremely good at locating existing information, making long-range connections, and doing some limited reasoning on top of this information.
Groq can serve up to 500 tokens per second. It is able to do this because it uses custom hardware that utilizes Linear Processor Units (LPUs) instead of GPUs. LPUs are designed to deliver deterministic performance for AI computations. They offer a more streamlined approach that eliminates the need for complex scheduling hardware, allowing every clock cycle to be utilized effectively. The system ensures consistent latency and throughput. LPUs can be linked together without the traditional bottlenecks found in GPU clusters, making them extremely scalable.
The Cocoa Press is a 3D printer that prints chocolate. It is now available for $3,995 prebuilt after a decade of development. Premade chocolate cores cost $49 per pack of 10. The printer allows users to program the heat of its nozzle to a tenth of a degree. It can create a wide range of shapes and textures, but there are still some problems to work out. This article shares what it is like using the machine. Pictures of chocolate prints made by the machine are available.
People should be defining problems and then finding the tools to solve those problems.
People have different skills and abilities - thinking of engineers as interchangeable is a mistake.
The use of Tagrisso and chemo together lowered the risk of disease progression or death in patients with EGFR-positive non-small cell lung cancer.
While Sequoia won't recoup its cash directly, it stands to benefit in other ways - many of its investments rely heavily on open-source software.
For the first time in history, biology has the opportunity to be engineering, not science.
Northwood aims to mass-produce ground stations that connect to satellites in space.
Google is introducing "Gemini Business" and "Gemini Enterprise" plans for Workspace customers, offering access to its advanced AI model, Gemini 1.0 Ultra, with enterprise-grade data protection. It will not use employee conversations for model training.
OpenAI has reportedly reached a valuation of over $80 billion following a deal with VC firm Thrive Capital. This valuation marks a nearly 3x increase in less than 10 months, making OpenAI the 3rd highest-valued tech startup in the world. The deal involves a share sale under a "tender offer," allowing employees to cash out their shares in the company.
Magic is building an AI software engineer and has raised more to help with this process.
Quantizing the KV cache, which is a technical detail of the Transformer architecture, allows it to use less memory at inference time. Quantization is the shrinking of the floating point precision with minimal loss in quality.
ZeroSwot is a new method for Speech Translation (ST) that overcomes the lack of data and differences between speech and text. It uses unique techniques to train a speech encoder with only speech recognition data, allowing it to work with a multilingual translation model.