Latest Tech News

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

Filtered by: task Clear Filter

ChatGPT’s new AI agent can browse the web and create PowerPoint slideshows

On Thursday, OpenAI launched ChatGPT Agent, a new feature that lets the company's AI assistant complete multi-step tasks by controlling its own web browser. The update merges capabilities from OpenAI's earlier Operator tool and the Deep Research feature, allowing ChatGPT to navigate websites, run code, and create documents while users maintain control over the process. The feature marks OpenAI's latest entry into what the tech industry calls "agentic AI"—systems that can take autonomous multi-s

Hate Windows 11? Here's how you can make it work more like Windows 10

Pau Barrena/Bloomberg via Getty Images I wrote the original version of this post when Windows 11 was still less than a year old, after reading a ton of feedback from early adopters. I didn't have to spend a lot of time in support forums and other community sites to hear a steady stream of criticism from people who were annoyed by the changes Microsoft made in the Windows 10 user interface they had grown comfortable with. Enough time has passed since the original Windows 11 release that some of

Android’s Desktop Mode gets a handy shortcut to pin your favorite apps to the taskbar

Mishaal Rahman / Android Authority TL;DR Google is improving Android’s still-in-development Desktop Mode by adding the ability to pin apps to the taskbar for quick access. A new “pin to taskbar” shortcut appears in an app’s context menu, letting you add it to the dock for easy launching even after it’s closed. This is a small but necessary step in Google’s effort to make Android a more viable desktop competitor, especially against platforms like the iPad. Android finally has a proper Desktop

Study finds AI tools made open source software developers 19 percent slower

When it comes to concrete use cases for large language models, AI companies love to point out the ways coders and software developers can use these models to increase their productivity and overall efficiency in creating computer code. However, a new randomized controlled trial has found that experienced open source coders became less efficient at coding-related tasks when they used current AI tools. For their study, researchers at METR (Model Evaluation and Threat Research) recruited 16 softwa

Kiro: A new agentic IDE

I’m sure you’ve been there: prompt, prompt, prompt, and you have a working application. It’s fun and feels like magic. But getting it to production requires more. What assumptions did the model make when building it? You guided the agent throughout, but those decisions aren’t documented. Requirements are fuzzy and you can’t tell if the application meets them. You can’t quickly understand how the system is designed and how that design will affect your environment and performance. Sometimes it’s b

AWS launches Kiro, its Cursor clone

I’m sure you’ve been there: prompt, prompt, prompt, and you have a working application. It’s fun and feels like magic. But getting it to production requires more. What assumptions did the model make when building it? You guided the agent throughout, but those decisions aren’t documented. Requirements are fuzzy and you can’t tell if the application meets them. You can’t quickly understand how the system is designed and how that design will affect your environment and performance. Sometimes it’s b

The upcoming GPT-3 moment for RL

The upcoming GPT-3 moment for RL Matthew Barnett, Tamay Besiroglu, Ege Erdil Jun 20, 2025 GPT-3 showed that simply scaling up language models unlocks powerful, task-agnostic, few-shot performance, often outperforming carefully fine-tuned models. Before GPT-3, achieving state-of-the-art performance meant first pre-training models on large generic text corpora, then fine-tuning them on specific tasks. Today’s reinforcement learning is stuck in a similar pre-GPT-3 paradigm. We first pre-train l

Hacking Coroutines into C

Hacking Coroutines into C 12.7.2025 A while ago, I was part of a team developing embedded software. The software was deeply rooted in state machines - dozens of them—spread across multiple functions. While this architecture is common in embedded development, especially for systems without an operating system, I started to question: Is this really the clearest way to express control flow? The state machines in our code worked fine, but understanding and maintaining them was often a headache. T

AI coding tools can reduce productivity

The buzz about AI coding tools is unrelenting. To listen to the reports, startups are launching with tiny engineering teams, non-programmers are “vibe-coding” entire apps, and the job market for entry-level programmers is crashing. But according to a METR experiment conducted in the spring of 2025, there’s at least one cohort that AI tools still aren’t serving. METR performed a rigorous study (blog post, full paper) to measure the productivity gain provided by AI tools for experienced developer

Not So Fast: AI Coding Tools Can Reduce Productivity

The buzz about AI coding tools is unrelenting. To listen to the reports, startups are launching with tiny engineering teams, non-programmers are “vibe-coding” entire apps, and the job market for entry-level programmers is crashing. But according to a METR experiment conducted in the spring of 2025, there’s at least one cohort that AI tools still aren’t serving. METR performed a rigorous study (blog post, full paper) to measure the productivity gain provided by AI tools for experienced developer

Measuring the impact of AI on experienced open-source developer productivity

We conduct a randomized controlled trial (RCT) to understand how early-2025 AI tools affect the productivity of experienced open-source developers working on their own repositories. Surprisingly, we find that when developers use AI tools, they take 19% longer than without—AI makes them slower. We view this result as a snapshot of early-2025 AI capabilities in one relevant setting; as these systems continue to rapidly evolve, we plan on continuing to use this methodology to help estimate AI accel

Measuring the Impact of AI on Experienced Open-Source Developer Productivity

We conduct a randomized controlled trial (RCT) to understand how early-2025 AI tools affect the productivity of experienced open-source developers working on their own repositories. Surprisingly, we find that when developers use AI tools, they take 19% longer than without—AI makes them slower. We view this result as a snapshot of early-2025 AI capabilities in one relevant setting; as these systems continue to rapidly evolve, we plan on continuing to use this methodology to help estimate AI accel

Windows 11 KB5062553 & KB5062552 cumulative updates released

Microsoft has released Windows 11 KB5062553 and KB5062552 cumulative updates for versions 24H2 and 23H2 to fix security vulnerabilities and issues. Today's updates are mandatory as they contain the July 2025 Patch Tuesday security patches for vulnerabilities discovered in previous months. You can install today's update by going to Start > Settings > Windows Update and clicking on 'Check for Updates.' You can also manually download and install the update from the Microsoft Update Catalog. Wha

I gave Google Tasks another shot — and remembered why I left

Andy Walker / Android Authority I’m not a calendar person. My day sliced into blue and yellow chunks down to the last minute — even telling me when to eat or take a break — isn’t the kind of rigidity I can live with. That’s far too structured for me. I instead swear by a hotchpotch system of Post-it notes, the good-old paper and pen, and most importantly, my task managing app — TickTick. TickTick is where all my personal and work tasks live, giving me a clear view of what needs to be done, whe

How a big shift in training LLMs led to a capability explosion

In April 2023, a few weeks after the launch of GPT-4, the Internet went wild for two new software projects with the audacious names BabyAGI and AutoGPT. “Over the past week, developers around the world have begun building ‘autonomous agents’ that work with large language models (LLMs) such as OpenAI’s GPT-4 to solve complex problems,” Mark Sullivan wrote for Fast Company. “Autonomous agents can already perform tasks as varied as conducting web research, writing code, and creating to-do lists.”

Backlog.md – Markdown‑native Task Manager and Kanban visualizer for any Git repo

Markdown‑native Task Manager & Kanban visualizer for any Git repository npm i -g backlog.md or bun add -g backlog.md Backlog.md turns any folder with a Git repo into a self‑contained project board powered by plain Markdown files and a zero‑config CLI. Features 📝 Markdown-native tasks -- manage every issue as a plain .md file 🔒 100 % private & offline -- backlog lives entirely inside your repo 📊 Instant terminal Kanban -- backlog board paints a live board in your shell 🌐 Modern web interf

Backlog.md – CLI that auto-generates task files (took my Claude success to 95 %)

Markdown‑native Task Manager & Kanban visualizer for any Git repository npm i -g backlog.md or bun add -g backlog.md Backlog.md turns any folder with a Git repo into a self‑contained project board powered by plain Markdown files and a zero‑config CLI. Features 📝 Markdown-native tasks -- manage every issue as a plain .md file 🔒 100 % private & offline -- backlog lives entirely inside your repo 📊 Instant terminal Kanban -- backlog board paints a live board in your shell 🌐 Modern web interf

Best iPad apps to boost productivity and make your life easier

Apple’s iPads come with built-in productivity tools like Notes, Calendar, and Reminders, but if you’d like to explore new ways to maximize productivity and organize your life, there are many apps out there to help you. Although the iPad started off as a device that could be used to stream content or browse the web on the go, Apple has essentially turned its iPads into computers that can handle a variety of different tasks for personal, work, and school use. As a result, there are numerous apps

Topics: app like notes tasks use

Modelling API rate limits as diophantine inequalities

modelling API rate limits as diophantine inequalities You're allowed 10 requests per hour. Each task you run makes three attempts: initial call, retry after 10 minutes, and retry after 30 minutes. What’s the maximum number of tasks you can safely run per hour? Most engineers throw exponential backoff at the problem. And it works great in most cases! But can we, for the sake of having fun, be more mathematical about this? In a way, this is just an integer feasibility problem. the setup Let’

Topics: 10 30 retry task window

The AI Backlash Keeps Growing Stronger

Before Duolingo wiped its videos from TikTok and Instagram in mid-May, social media engagement was one of the language-learning app’s most recognizable qualities. Its green owl mascot had gone viral multiple times and was well known to younger users—a success story other marketers envied. But, when news got out that Duolingo was making the switch to become an “AI-first” company, planning to replace contractors who work on tasks generative AI could automate, public perception of the brand soured

AI agents win over professionals - but only to do their grunt work, Stanford study finds

Getty Images/Jonathan Kitchen AI agents are one of the buzziest trends in Silicon Valley, with tech companies promising big productivity gains for businesses. But do individual workers actually want to use them? A new study from Stanford University shows the answer may be yes -- as long as they automate mundane tasks and don't encroach too far on human agency. Also: Don't be fooled into thinking AI is coming for your job - here's the truth Titled "Future of Work with AI Agents," the study se

Compiling LLMs into a MegaKernel: A path to low-latency inference

One of the most effective ways to reduce latency in LLM inference is to fuse all computation and communication into a single megakernel — also known as a persistent kernel. In this design, the system launches just one GPU kernel to execute the entire model — from layer-by-layer computation to inter-GPU communication — without interruption. This approach offers several key performance advantages: Eliminates kernel launch overhead, even in multi-GPU settings, by avoiding repeated kernel invocatio

Topics: gpu graph kernel mpk task

Compiling LLMs into a MegaKernel: A Path to Low-Latency Inference

One of the most effective ways to reduce latency in LLM inference is to fuse all computation and communication into a single megakernel — also known as a persistent kernel. In this design, the system launches just one GPU kernel to execute the entire model — from layer-by-layer computation to inter-GPU communication — without interruption. This approach offers several key performance advantages: Eliminates kernel launch overhead, even in multi-GPU settings, by avoiding repeated kernel invocatio

Topics: gpu graph kernel mpk task

Is there a half-life for the success rates of AI agents?

In general, ability to perform a task drops off as its duration increases, so they use the AI agent’s performance on tasks of different lengths to estimate the task-length at which the model would have a 50% success rate. They then showed that this length has been doubling every 7 months as the capabilities of frontier agents improve. The task-lengths are measured by how long it took humans to solve the same tasks. They used 50% success rate as their chief performance threshold because it is th

Topics: 50 rate success task time

Is There a Half-Life for the Success Rates of AI Agents?

In general, ability to perform a task drops off as its duration increases, so they use the AI agent’s performance on tasks of different lengths to estimate the task-length at which the model would have a 50% success rate. They then showed that this length has been doubling every 7 months as the capabilities of frontier agents improve. The task-lengths are measured by how long it took humans to solve the same tasks. They used 50% success rate as their chief performance threshold because it is th

Topics: 50 rate success task time

OnePlus tablets are getting one of Android’s best, and one of its worst, productivity features

Ryan Haines / Android Authority TL;DR The latest OxygenOS 15 update brings Google’s Circle to Search and a new floating taskbar to the OnePlus Open and OnePlus Pad 2. Circle to Search allows users to quickly search anything on their screen by tapping, highlighting, or drawing over it. The floating taskbar, while aiming to balance productivity and screen real estate, lacks a quick toggle to switch styles and on tablets, it requires disabling gesture navigation for the fixed version. OnePlus m

Move over Open Canvas, this could be the best way to multitask on a foldable

TL;DR vivo’s upcoming X Fold 5 features a new multitasking system called “Atomic Workbench” capable of running up to five active apps simultaneously. “Atomic Workbench” enables one dominant app view while keeping others minimized but live, improving upon OnePlus’ Open Canvas approach with a better layout for engaging more apps. The vivo X Fold 5 launches on June 25 in China. OnePlus pioneered the Open Canvas feature with the OnePlus Open foldable, allowing users to open three apps simultaneou

Windows 11 users want these five features back

When Windows 11 was first released, many long-time users felt features they loved had been taken away overnight. Three and a half years later, the same complaints still rise to the top of the Feedback Hub with tens of thousands of votes. Below are the five most wanted Windows 11 features on the Microsoft Feedback Hub. The most wanted requests in Windows 11 "Bring back the ability to move the taskbar to the top and sides of the screen on Windows 11" Right now, the taskbar is glued to the bot

A receipt printer cured my procrastination [ADHD]

Why can I focus for hours on a game but procrastinate when writing an email? For over 20 years, I've tried everything to boost my productivity. I've read dozens of books and tested many methods. Nothing made a real difference. But I finally found a solution by understanding what makes video games addictive. And a thermal receipt printer will become the best investment of my life. My fight against procrastination I started my business when I was 21 (I'm now 39). I built custom apps and did cons

Android 16 will make it easier to find your recent apps with a new taskbar overflow feature

TL;DR An upcoming Android 16 update will add a taskbar overflow button to improve app switching on tablets when many apps have been recently opened. Tapping the overflow button will open a horizontally scrolling carousel that displays previews of your recent apps. This feature is set to launch later this year with either Android 16 QPR1 or QPR2. Back in 2021, Google added a taskbar to Android to improve app switching on tablets. Since then, the company has made several improvements, such as a