601
Macaroons Escalated Quickly
Documentation and guides from the team at Fly.io.
602
Claude's Character
How character training cultivates traits like curiosity, open-mindedness, and thoughtfulness in Claude using a variant of Constitutional AI.
603
All-in-one embedding model for interleaved text, images, and screenshots
TL;DR — We are excited to announce voyage-multimodal-3, a new state-of-the-art for multimodal embeddings and a big step forward towards seamless RAG and semantic search for documents rich with both…
604
The myth that you can’t build interactive web apps except as single page app
Tony Alaribe challenges common myths about multi-page applications (MPAs) and explores how modern browser technologies can enable fast, interactive, and offline-capable web applications without relying on single-page application (SPA) frameworks. Alaribe discusses advancements in service workers, caching, and…
605
Perf Is Not Enough
Are database benchmarks still relevant ? Let's understand why it's a poor way to choose a database. | Reading time: 17 min read
606
607
Figma’s Journey to TypeScript
Figma's team recently converted one of its codebases from a custom programming language to TypeScript without disrupting a single day of development.
608
Attribution is dying, clicks are dying
Welcome to another edition of 5-Minute Whiteboards. And folks, we've got a doozy of a topic. Yes, I'm being intentionally provocative. But it's because
609
0day Contest for End-of-Life Devices Announced
The Junkyard contest focuses on deprecated systems, responsible disclosure, and meme-worthy exploits. Join us today!
610
Qwen2 LLM Released
GITHUB HUGGING FACE MODELSCOPE DEMO DISCORD Introduction After months of efforts, we are pleased to announce the evolution from Qwen1.5 to Qwen2. This time, we bring to you: Pretrained and instruction-tuned models of 5 sizes, including Qwen2-0.5B, Qwen2-1.5B, Qwen2-7B, Qwen2-57B-A14B, and Qwen2-72B; Having been…
611
MobileDiffusion: Rapid text-to-image generation on-device
Posted by Yang Zhao, Senior Software Engineer, and Tingbo Hou, Senior Staff Software Engineer, Core ML Text-to-image diffusion models have shown ex...
612
Tog's Paradox
Tog’s Paradox (also known as The Complexity Paradox or Tog’s ComplexityParadox) is an observation that products aiming to simplify a task for users tend toinspire new, more complex tasks. It’s one ...
613
614
Elixir and Machine Learning in 2024 so far: MLIR, Arrow, structured LLM, etc.
An update on the latest advancements of Elixir, data, and Machine Learning.
615
Rethinking Text Resizing on Web
By: Steven Bassett
616
Critical Exploit in MediaTek Wi-Fi Chipsets: Zero-Click Vulnerability
a post going over 4 exploits for CVE-2024-20017, a remotely exploitable buffer overflow in a component of the MediaTek MT7622 SDK.
617
618
Hacking misconfigured AWS S3 buckets: A complete guide
AWS S3 (Simple Storage Service) buckets are a popular storage service used by software companies and organizations to store public as well as sensitive data. However, the implementation of this servic...
619
Apple Hearing Study shares preliminary insights on tinnitus
The Apple Hearing Study is sharing new insights on tinnitus in one of the largest surveys to date.
620
FLUX is fast and it's open source
FLUX is now much faster on Replicate, and we’ve made our optimizations open-source so you can see exactly how they work and build upon them.
621
622
Butterflies accumulate static electricity to attract pollen without contact
Butterflies and moths collect so much static electricity whilst in flight, that pollen grains from flowers can be pulled by static electricity across air gaps of several millimetres or centimetres.
623
Testing how hard it is to cheat with ChatGPT in interviews
We had interviewees use ChatGPT without their interviewer knowing. Here's what happened.
624
CriticGPT: Finding GPT-4's mistakes with GPT-4
CriticGPT, a model based on GPT-4, writes critiques of ChatGPT responses to help human trainers spot mistakes during RLHF
625
How we made our AI code review bot stop leaving nitpicky comments
Lessons learned from our challenges reducing LLM output: Discover everything that went wrong when trying to make our models say less.
626
OpenSSH Backdoors
Imagine this: an OpenSSH backdoor is discovered, maintainers rush to push out a fixed release package, security researchers trade technical details on mailing lists to analyze the backdoor code. Speculation abounds on the attribution and motives of the attacker, and the tech media pounces on the story. A near miss
627
Understanding SIMD: Infinite complexity of trivial problems
A deep dive into the complexities of optimizing code for SIMD instruction sets across multiple platforms.
628
Kagi Translate
Today, we're introducing Kagi Translate. It's not revolutionary - it's simply a better translation service. Our combination of advanced language models and precise output selection delivers translations that surpass existing solutions, including Google Translate and DeepL....
629
1BRC merykitty's magic SWAR: 8 lines of code explained in 3k words
A deep dive into merykitty's 8-line SWAR temperature parser from the One Billion Row Challenge, using bitwise tricks and ALU magic with no loops or branches.
630
Linux Touchpad Like MacBook Update: 2023 Progress on Smooth Scrolling
Bill is driven by the challenge of how best to quantify valuable questions that defy quantification. It's possible this instinct may have been awakened in Bill at age 14, when he won a soft, stuffed bunny at the orthodontist for guessing the number of jelly beans in the gumball machine.