Latest Tech News

Stay updated with the latest in technology, AI, cybersecurity, and more

Filtered by: code Clear Filter

Forklifts require training

A lot gets covered in today's discourse about AI in software development. Most of it is noise, ranging from nihilism that we're all writing mediocre code anyway so why does it matter to endless wannabe AI influencers doing engagement bait on Twitter. Every new model release gets a bunch of threadicles 👇 amounting to the 2025 version of "Safari feels snappier". Some of it is useful, mostly crafty developers in the community sharing novel ways they're using it to solve hard problems or draw inspir

D2 (text to diagram tool) now supports ASCII renders

In the latest release of D2 (0.7.1), we introduce ASCII outputs. Any output file with extension txt will use the ASCII renderer to write to it. Here is an example of their rendering from the D2 Vim extension. The user opens a .d2 file and opens a preview window, which updates upon every save. Perhaps the most useful place for ASCII diagrams is in the source code comments. Small simple diagrams next to functions or classes can serve to be much clearer than describing a flow. Here again the Vi

Next year’s Apple Watch could add Touch ID, per leaked code

We’re just weeks away from the unveiling of Apple Watch Ultra 3 and Apple Watch Series 11. But leaked Apple code has now revealed a big new feature that could be coming to next year’s Apple Watch lineup: Touch ID support. Touch ID support might come to 2026 Apple Watch line Last week, Macworld published two articles based on leaked Apple code. One offered evidence of the M5 MacBook Pro being tested with a C1 modem for 5G connectivity. The other involved an M4 Ultra chip in development for a n

Topics: apple code id touch watch

Apple preps native Claude integration on Xcode

Towards the end of this year’s WWDC keynote, Craig Federighi said that Apple had “expanded” their vision for Swift Assist, and would bring native integration with ChatGPT, alongside support for other LLMs via API, directly to Xcode. Now, 9to5Mac can confirm that Apple is set to support native integration with Anthropic’s Claude as well. Digging into today’s release of Xcode 26 beta 7, 9to5Mac found multiple references to built-in support for Anthropic accounts directly within the new “Intellige

ERMAC Android malware source code leak exposes banking trojan infrastructure

The source code for version 3 of the ERMAC Android banking trojan has been leaked online, exposing the internals of the malware-as-a-service platform and the operator’s infrastructure. The code base was discovered in an open directory by Hunt.io researchers while scanning for exposed resources in March 2024. They located an archive named Ermac 3.0.zip, which contained the malware’s code, including backend, frontend (panel), exfiltration server, deployment configurations, and the trojan’s build

Topics: code ermac hunt io panel

Typechecker Zoo

This is a pet project of mine I’ve been working on for a while. We’re going to create minimal implementations of the most successful static type systems of the last 50 years. This will involve making toy implementations of programming languages and the core typechecking algorithms. These obviously have evolved a lot over the years, so we’ll start with the simple ones and proceed all the way up to modern dependent types. Basically a fun romp through half a century of programming language design.

Show HN: A Minimal Hacker News Reader for Apple Watch Built with SwiftUI

HackerNewsWatch (watchOS) A minimal Hacker News reader for Apple Watch built with SwiftUI. Scrollable top stories feed with title, points, and comments count Tap a story to view comments in a simple tree-style view "Open Article" link at the top opens the article in the watch browser HN-like styling (orange accent) Requirements Xcode 15 or newer macOS with command line tools Homebrew (for XcodeGen) or install XcodeGen manually Generate the Xcode project ./scripts/generate.sh This wil

LLMs and coding agents are a security nightmare

Last October, I wrote an essay called “When it comes to security, LLMs are like Swiss cheese — and that’s going to cause huge problems” warning that “The more people use LLMs, the more trouble we are going to be in”. Until last week, when I went to Black Hat Las Vegas, I had no earthly idea how serious the problems were. There, I got to know Nathan Hamiel, a Senior Director of Research at Kudelski Security and the AI, ML, and Data Science track lead for Black Hat, and also sat in on a talk by tw

MCP doesn't need tools, it needs code

Your MCP Doesn’t Need 30 Tools: It Needs Code I wrote a while back about why code performs better than MCP (Model Context Protocol) for some tasks. In particular, I pointed out that if you have command line tools available, agentic coding tools seem very happy to use those. In the meantime, I learned a few more things that put some nuance to this. There are a handful of challenges with CLI-based tools that are rather hard to resolve and require further examination. In this blog post, I want to

LLMs and Coding Agents = Security Nightmare

Last October, I wrote an essay called “When it comes to security, LLMs are like Swiss cheese — and that’s going to cause huge problems” warning that “The more people use LLMs, the more trouble we are going to be in”. Until last week, when I went to Black Hat Las Vegas, I had no earthly idea how serious the problems were. There, I got to know Nathan Hamiel, a Senior Director of Research at Kudelski Security and the AI, ML, and Data Science track lead for Black Hat, and also sat in on a talk by tw

MCP Doesn't Need 30 Tools: It Needs Code

Your MCP Doesn’t Need 30 Tools: It Needs Code I wrote a while back about why code performs better than MCP (Model Context Protocol) for some tasks. In particular, I pointed out that if you have command line tools available, agentic coding tools seem very happy to use those. In the meantime, I learned a few more things that put some nuance to this. There are a handful of challenges with CLI-based tools that are rather hard to resolve and require further examination. In this blog post, I want to

Mangle – a language for deductive database programming

Mangle Mangle is a programming language for deductive database programming. It is an extension of Datalog, with various extensions like aggregation, function calls and optional type-checking. Deductive database programming is useful for bringing data from multiple data sources together since it enables us to represent and query that data in a uniform way. It can also be used to model domain knowledge, similar to machine-readable ontology but without being restricted to binary predicates. Data

SuperSight: A graphical enhancement mod for Brøderbund's "Stunts"

Annali da Samarcanda Alberto Marnetto's Notebook SuperSight: a graphical enhancement mod for Brøderbund's Stunts Clickbait disclaimer: Stunts was actually created by Distinctive Software; Brøderbund was only the publisher. But nobody heard about Distinctive, and I wanted to be sure this article does not get confused with the one about Disney's Stunt Island. Sorry about that. Part I This series will tell about the creation of SuperSight, a mod for Stunts intended to bring the game’s 3D engin

Does OLAP Need an ORM

TL;DR · ORMs have proven to be useful for many developers in the OLTP/transactional stack (Postgres, MySQL, etc). · OLAP/analytical databases like ClickHouse could potentially benefit from ORM abstractions. · Existing transactional ORMs probably shouldn’t be extended to OLAP due to fundamental differences in semantic meaning between OLTP and OLAP. · Moose OLAP (part of MooseStack) is an open source, MIT-licensed implementation of an ORM-like interface for ClickHouse, inspired by tran

LL3M: Large Language 3D Modelers

LL3M uses a team of large language models to write Python code that creates and edits 3D assets in Blender. Given user text instructions, the agents are capable of creating expressive shapes from scratch, and realizing complex, precise geometric manipulations in code. Whereas previous uses of code-writing LLMs for 3D creation have focused on specific subtasks or constrained procedural programs and primitives, our method is able to create unconstrained assets with geometry, layout, and appearance

Using AI to secure AI

One of Anthropic's quieter releases recently was their "Security Review," where Claude Code can identify and fix security issues in your code. But how good is it really? In my case, will it find issues with code it helped me write for my newsletter service and Chrome extension? The release states it uses a "specialized security-focused prompt that checks for common vulnerability patterns." After throwing so much compute at model training, LLMs are nearing the top of the S-Curve, so finding ways

I let LLMs write an Elixir NIF in C; it mostly worked

This post documents how I built a cross-platform Elixir NIF in C to get on-demand up-to-date disk-usage stats without relying on os_mon and its disksup service. I had Grok 3 generate the initial C code and Makefile, then iterated through multiple code reviews by Gemini 2.5 Flash and GPT-5 to make it work on Linux, macOS, Windows, and the BSDs (except DragonFlyBSD). Along the way, I ran into typical LLM hiccups that speak volumes about the breathless hyperbole often peddled by LLM vendors, comput

Letting inmates run the asylum: Using AI to secure AI

One of Anthropic's quieter releases recently was their "Security Review," where Claude Code can identify and fix security issues in your code. But how good is it really? In my case, will it find issues with code it helped me write for my newsletter service and Chrome extension? The release states it uses a "specialized security-focused prompt that checks for common vulnerability patterns." After throwing so much compute at model training, LLMs are nearing the top of the S-Curve, so finding ways

Apple trained an LLM to teach itself good UI code in SwiftUI

In a new study, a group of Apple researchers describe a very interesting approach they took to, basically, get an open-source model to teach itself how to build good user interface code in SwiftUI. Here’s how they did it. In the paper UICoder: Finetuning Large Language Models to Generate User Interface Code through Automated Feedback, the researchers explain that while LLMs have gotten better at multiple writing tasks, including creative writing and coding, they still struggle to “reliably gene

The new science of “emergent misalignment”

If there’s an upside to this fragility, it’s that the new work exposes what happens when you steer a model toward the unexpected, Hooker said. Large AI models, in a way, have shown their hand in ways never seen before. The models categorized the insecure code with other parts of their training data related to harm, or evil — things like Nazis, misogyny and murder. At some level, AI does seem to separate good things from bad. It just doesn’t seem to have a preference. Wish for the Worst In 2022

Topics: ai code model models said

Anthropic brings Claude's learning mode to regular users and devs

This past spring, Anthropic introduced learning mode, a feature that changed Claude's interaction style. When enabled, the chatbot would, following a question, try to guide the user to their own solution, instead of providing them with an answer outright. Since its introduction in April, learning mode has only been available to Claude for Education users. Now, like OpenAI did with Study Mode, Anthropic is making the tool available to everyone. Starting today, Claude.ai users will find a new opt

Why LLMs can't really build software

One of the things I have spent a lot of time doing is interviewing software engineers. This is obviously a hard task, and I don’t claim to have a magic solution; but it’s given me some time to reflect on what effective software engineers actually do. When you watch someone who knows what they are doing, you'll see them looping over the following steps: Build a mental model of the requirements Write code that (hopefully?!) does that Build a mental model of what the code actually does Identify t

JetBrains working on higher-abstraction programming language

JetBrains, creator of the popular Kotlin programming language, is developing a new programming language intended to make AI and code much more controllable and transparent. In a July 23 interview with InfoWorld, JetBrains CEO Kirill Skrygan elaborated on company plans for an as-yet-unnamed language that would describe a program at a higher level of abstraction. He reflected on how computer code originally was written in Assembler and moved to higher levels of abstraction with C and C++, then on

Why LLMs Can't Build Software

One of the things I have spent a lot of time doing is interviewing software engineers. This is obviously a hard task, and I don’t claim to have a magic solution; but it’s given me some time to reflect on what effective software engineers actually do. When you watch someone who knows what they are doing, you'll see them looping over the following steps: Build a mental model of the requirements Write code that (hopefully?!) does that Build a mental model of what the code actually does Identify t

Apple Vision Pro 2 may have a better chip than expected

Apple Vision Pro will be two years old in February. For those holding out for a second-gen version, it sounds like the Apple Vision Pro 2 may have a better chip than expected. The rumors have gone back and forth over which chip the refreshed Apple Vision Pro will use. The current hardware runs on an M2 chip, the newest chip when Apple announced Vision Pro, although the M3 arrived in products before Apple Vision Pro was released. Early reports said the next Apple Vision Pro will run on the yet-

Apple accidentally leaked some of its upcoming chip bumps

is a news writer who covers the streaming wars, consumer tech, crypto, social media, and much more. Previously, she was a writer and editor at MUO. Posts from this author will be added to your daily email digest and your homepage feed. Apple appears to have inadvertently included details about several of its upcoming devices in code spotted by MacRumors. One part of the code suggests that Apple plans on putting an M5 chip in its next-gen Vision Pro headset, aligning with a previous report from

The Kryptos Key Is Going Up for Sale

Ever since artist James Sanborn unveiled Kryptos, an outdoor sculpture that sits at CIA headquarters, amateur and professional cryptanalysts have been feverishly attempting to crack the code hidden in its nearly 1800-character message. While they have decoded 3 of the 4 panels of ciphertext in the S-shaped copper artwork, the final panel, known as K4, still defies solution. Only one human being on Earth knows the message of K4: Sanborn. But soon someone else will join the club. Sanborn is puttin

ForgeFed: ActivityPub-based forge federation protocol

ForgeFed is a federation protocol for software forges and code collaboration tools for the software development lifecycle and ecosystem. This includes repository hosting websites, issue trackers, code review applications, and more. ForgeFed provides a common substrate for people to create interoperable code collaboration websites and applications. Federation means that these websites can interact, allowing the humans using them to interact too, despite being registered on different websites. Fo