Just another WordPress site

Blog

  • Build Smarter Sub-Agents with Deep Agents : Plans, Delegates & Ships Results Locally

    Build Smarter Sub-Agents with Deep Agents : Plans, Delegates & Ships Results Locally

    Build Smarter Sub-Agents with Deep Agents : Plans, Delegates & Ships Results Locally

    By

    Diagram of Deep Agents built on LangChain and LangGraph for complex tasks

    What if you could build an autonomous system that not only executes complex, long-running tasks but also adapts to evolving challenges with remarkable precision? Enter Deep Agents, a innovative framework that redefines how developers approach task automation. Imagine delegating intricate workflows to agents that seamlessly integrate tools like file system access, script execution, and even sub-agent delegation, all while maintaining efficiency and scalability. In a world where automation is no longer a luxury but a necessity, Deep Agents offers a bold solution: a modular, opinionated system designed to simplify the creation of autonomous agents without the usual technical headaches. But what makes this framework truly stand out is its ability to bridge high-level abstractions with low-level infrastructure, empowering developers to focus on innovation rather than infrastructure management.

    In this exploration of Deep Agents, LangChain explain how this open source framework builds on the strengths of LangChain and LangGraph to deliver a purpose-driven approach to automation. From its ability to handle long-running tasks to its seamless integration of modular workflows, Deep Agents is packed with features that cater to both simplicity and flexibility. Whether you’re a developer aiming to streamline intricate processes or someone curious about the future of autonomous systems, this piece will guide you through the key principles, tools, and real-world applications that make Deep Agents a fantastic option. As we unravel its capabilities, you might find yourself rethinking what’s possible in the realm of task automation.

    Key Features of Deep Agents

    TL;DR Key Takeaways :

    • Deep Agents is an open source framework designed for creating autonomous agents capable of handling complex, long-running tasks with efficiency and precision.
    • It integrates LangChain for high-level abstractions and LangGraph for low-level infrastructure, combining tools like file system access, script execution, and sub-agent delegation.
    • The framework emphasizes modularity, allowing developers to design scalable workflows while simplifying infrastructure management.
    • Key features include LangChain integration, durable execution via LangGraph, middleware enhancements, and a powerful Command Line Interface (CLI) for local task execution.
    • Version 0.2 introduces pluggable backends and middleware improvements, enhancing usability and adaptability for diverse task automation needs.

    The framework is particularly suited for developers seeking to automate intricate processes without being bogged down by infrastructure management. Its modular design and focus on practical applications make it a valuable tool for building scalable, autonomous systems.

    Deep Agents offers a carefully curated set of features designed to streamline task execution. Its design philosophy emphasizes simplicity and flexibility, allowing you to focus on solving problems rather than managing technical complexities. Some of its standout features include:

    • LangChain Integration: Provides general abstractions for chat models and tools, allowing seamless interaction with language-based systems.
    • LangGraph Infrastructure: Offers low-level support for durable execution, memory management, and human-in-the-loop processes, making sure reliability in complex workflows.
    • File System Access: Assists interaction with local file systems, allowing efficient context management and data storage for long-running tasks.
    • Script Execution: Supports the execution of bash and shell scripts, expanding the range of tasks your agents can perform.
    • Sub-Agent Delegation: Allows for context isolation and task delegation, allowing the breakdown of complex workflows into manageable components.
    • Middleware Enhancements: Features such as context compression and prompt caching improve performance by reducing overhead and optimizing resource usage.

    These features collectively make Deep Agents a powerful tool for developers aiming to build autonomous systems that are both efficient and adaptable.

    How Deep Agents Builds on LangChain and LangGraph

    Deep Agents extends the capabilities of LangChain and LangGraph by integrating predefined tools and opinionated prompting to create a cohesive system tailored for building autonomous agents. While LangChain and LangGraph serve as foundational frameworks, Deep Agents bridges their functionalities to deliver a more application-specific approach. Here’s how they compare:

    • LangChain: Primarily focuses on general abstractions for chat models and tools, making it ideal for conversational AI applications.
    • LangGraph: Provides low-level infrastructure for durable execution, memory management, and human-in-the-loop processes, making sure robust task handling.
    • Deep Agents: Combines the strengths of both frameworks, offering a unified system designed for building autonomous agents with specific use cases in mind.

    By using the best aspects of LangChain and LangGraph, Deep Agents provides a more integrated and purpose-driven solution for task automation.

    What are Deep Agents?

    Advance your skills in Deep Agents by reading more of our detailed content.

    Command Line Interface (CLI) Capabilities

    The Command Line Interface (CLI) in Deep Agents enhances local execution by providing direct access to essential tools and resources. This feature is particularly useful for developers working in environments with limited resources or requiring precise control over agent behavior. Through the CLI, you can:

    • Execute tasks with direct access to the local file system and memory, making sure efficient resource utilization.
    • Use built-in skills, shell tools, and web fetch tools to perform a wide range of operations.
    • Customize workflows by adding new tools or modifying existing instructions to suit specific requirements.

    The CLI ensures that Deep Agents remains a versatile and practical framework for developers, regardless of the complexity of their projects.

    Applications and Design Principles

    Deep Agents is designed for scenarios where autonomy and precision are critical. Its modular architecture and advanced prompting capabilities make it suitable for a variety of applications. Key use cases include:

    • Long-Running Tasks: Ideal for workflows requiring sustained execution over extended periods, such as data processing or monitoring systems.
    • Task Delegation: Sub-agents enable the distribution of responsibilities across isolated contexts, improving efficiency and scalability.
    • Context Management: Offloading context to the file system reduces memory overhead and enhances overall performance.
    • Modular Workflows: Encourages the use of small, general-purpose tools to address complex problems, making sure flexibility and maintainability.

    These principles ensure that Deep Agents remains adaptable to evolving requirements, making it a reliable choice for developers tackling diverse challenges.

    Enhancements in Version 0.2

    The release of Deep Agents 0.2 introduces several updates aimed at improving usability and performance. Key enhancements include:

    • Pluggable Backends: Enables the integration of custom tools and systems, including local file system access, to expand the framework’s capabilities.
    • Middleware Improvements: Enhances context management and tool handling, making sure smoother and more efficient execution of tasks.

    These updates reflect the framework’s commitment to evolving alongside user needs and technological advancements, making sure it remains a innovative solution for task automation.

    Open source Collaboration and Community

    As an open source project, Deep Agents thrives on community involvement. By encouraging feedback and contributions, it fosters a collaborative environment where developers can refine and expand its capabilities. Whether you’re building autonomous agents or exploring new use cases, Deep Agents provides a solid foundation for innovation.

    The open source nature of the framework ensures that it remains accessible and adaptable, empowering developers to create solutions tailored to their unique challenges.

    Media Credit: LangChain

    Filed Under: AI

    Latest Geeky Gadgets Deals

    Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.

  • Nano Banana Pro : Turn Rough Sketches or Prompts into Stunning Visuals in Minutes

    Nano Banana Pro : Turn Rough Sketches or Prompts into Stunning Visuals in Minutes

    Nano Banana Pro : Turn Rough Sketches or Prompts into Stunning Visuals in Minutes

    By

    Interface of Nano Banana Pro in Gemini 3 Pro creating a product mockup

    What if a single tool could transform your creative process, turning complex visual tasks into seamless, professional-grade outputs? Meet the Nano Banana Pro, a innovative AI powerhouse that’s redefining what’s possible in image editing and generation. Imagine effortlessly creating lifelike headshots, restoring old photographs with stunning clarity, or designing scroll-stopping social media visuals, all with just a few clicks. Bold claim? Maybe. But with its unmatched precision and versatile applications, Nano Banana Pro is quickly becoming the go-to solution for marketers, designers, and content creators looking to elevate their craft.

    Below Skill Leap AI explores 25 jaw-dropping ways this new Goolge AI image tool can transform your workflow. From crafting dynamic YouTube thumbnails to generating polished infographics from raw data, the Nano Banana Pro is packed with features that push the boundaries of creativity. But it’s not just about what it can do, it’s about how it fits into your unique needs. Whether you’re curious about its seamless integration with platforms like Gamma or its ability to restore vintage photos, this guide will uncover the tool’s most surprising capabilities. Prepare to rethink what’s possible in visual storytelling.

    Key Features of Nano Banana Pro

    TL;DR Key Takeaways :

    • Comprehensive Features: Nano Banana Pro offers advanced image resizing, merging, restoration, enhancement, and infographic generation, making it a versatile tool for professionals.
    • Content Creation Applications: Ideal for creating social media visuals, YouTube thumbnails, marketing graphics, and website mockups with professional-grade quality.
    • Strengths and Limitations: Excels in generating realistic visuals and seamless integration with tools like Gamma, but requires specific prompts and lacks advanced editing features like true transparency.
    • Practical Use Cases: Suitable for marketing, branding, professional presentations, and creative projects such as movie posters and magazine covers.
    • Future Developments: Anticipated updates include enhanced text-to-image generation, improved realism, and expanded use cases, solidifying its role in AI-driven image manipulation.

    Nano Banana Pro stands out for its ability to handle a wide range of image-related tasks, making it an indispensable tool for professionals. Its core features include:

    • Advanced image resizing: Adjust aspect ratios with precision to meet specific requirements.
    • Image merging: Seamlessly combine multiple images into cohesive and visually appealing compositions.
    • Professional headshot creation: Generate high-quality headshots and customized visuals tailored to your needs.
    • Infographic generation: Transform raw data into visually compelling infographics and graphics.
    • Image restoration: Perform color correction, remove unwanted objects, and restore old or damaged photos.
    • Photo enhancement: Add color to black-and-white images and improve the clarity of blurry visuals.

    These features make Nano Banana Pro a powerful tool for producing high-quality visuals, whether for marketing campaigns, social media content, or creative projects.

    Applications in Content Creation

    Nano Banana Pro is particularly effective in content creation, offering tools to design visually engaging assets for various platforms. Its capabilities allow you to:

    • Social media visuals: Create Instagram posts, Facebook ads, and LinkedIn infographics that capture attention.
    • YouTube thumbnails: Design thumbnails optimized to drive engagement and clicks.
    • Marketing visuals: Generate memes and promotional graphics tailored for e-commerce campaigns.
    • Website mockups: Develop landing page designs and product showcases with professional polish.

    The ability to produce professional-grade visuals quickly and efficiently makes Nano Banana Pro a valuable resource for businesses and creators aiming to stand out in competitive markets.

    25 Examples of Google’s Nano Banana Pro

    Find more information on Nano Banana by browsing our extensive range of articles, guides and tutorials.

    Strengths and Limitations

    Understanding the strengths and limitations of Nano Banana Pro is crucial for maximizing its potential and achieving the best results.

    Strengths:

    • High accuracy: Delivers precise and reliable outputs when provided with detailed prompts.
    • Realistic visuals: Generates images that are both visually appealing and lifelike.
    • Seamless integration: Works well with tools like Gamma for presentations and other AI platforms.

    Limitations:

    • Specificity required: Struggles with vague or open-ended prompts, requiring users to provide clear instructions.
    • Transparency issues: Cannot create true transparent backgrounds, often necessitating external tools for refinement.
    • Advanced editing constraints: Lacks some advanced editing features, which may require complementary software like Photoshop.

    By using its strengths and addressing its limitations, you can optimize Nano Banana Pro’s effectiveness and achieve your creative goals.

    Practical Use Cases

    Nano Banana Pro’s versatility makes it suitable for a wide range of professional and creative applications. Some notable use cases include:

    • Marketing and branding: Add logos to products and place them in realistic settings to enhance brand identity and product appeal.
    • Professional presentations: Generate polished infographics and slides by combining Nano Banana Pro’s visuals with tools like Gamma.
    • Creative projects: Design movie posters, magazine covers, or futuristic visuals for artistic endeavors.

    These examples demonstrate the tool’s adaptability across industries, making it a valuable asset for diverse projects.

    Crafting Effective Prompts

    The quality of results you achieve with Nano Banana Pro depends heavily on the prompts you provide. Vague or poorly structured instructions often lead to suboptimal outputs, while detailed and specific prompts enable the AI to deliver precise results. Tools like Gemini can assist in refining and optimizing your prompts, making sure the generated visuals align with your expectations. Developing a clear understanding of how to structure prompts effectively is key to unlocking the full potential of this tool.

    Integration with Other Tools

    Nano Banana Pro’s compatibility with other platforms enhances its functionality and usability. Key integrations include:

    • Gamma: Combine Nano Banana Pro’s visuals with Gamma’s tools to create presentations and websites effortlessly.
    • Photoshop: Use Photoshop for advanced editing and refinement of images generated by Nano Banana Pro.

    This interoperability allows you to seamlessly incorporate Nano Banana Pro into your existing workflow, boosting productivity and creative output.

    Educational Resources for Mastery

    To help you master Nano Banana Pro, various educational resources are available. Platforms like Skill Leap offer courses and prompt templates specifically tailored to this tool. Additionally, structured learning paths for generative AI tools, including Nano Banana Pro and Midjourney, provide guidance for users at all skill levels. These resources can accelerate your learning curve, allowing you to use the tool effectively and efficiently.

    Future Developments

    As AI technology continues to evolve, Nano Banana Pro is expected to introduce even more advanced capabilities. Potential updates may include enhanced text-to-image generation, improved image realism, and expanded use cases across professional and creative industries. These advancements will likely solidify its position as a leading tool in AI-driven image manipulation, offering even greater value to its users.

    Media Credit: Skill Leap AI

    Filed Under: AI, Guides

    Latest Geeky Gadgets Deals

    Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.

  • 5 of these 10 photos are AI-generated — can you spot them?

    5 of these 10 photos are AI-generated — can you spot them?

    AI-generated images are unreal at this point. Indistinguishable from the real thing, easy to create, and accessible to anyone, there is little wonder why so-called AI slop runs rampant.

    It’s not all slop, mind. Not all AI-generated images are the next round of “shrimp Jesus” or featured on this week’s trending “Insane AI Slop” pages.

    In that, I’ve got a little challenge for you. In the following ten images, five are AI. But which ones are which? Check out each image and the prompt used to create it, and let us know if you could spot all of our AI-generated images in the comments below.

    10

    A macro shot of a ginger cat

    Capturing a cat’s favorite state: sleep

    Prompt

    An extreme close-up, high-resolution photograph of a sleeping ginger tabby cat. The cat’s face fills the frame, eyes closed peacefully, with soft orange fur marked by subtle stripes. Fine whiskers extend outward, and the nose is a soft pink. The cat rests its head on a pale fabric surface, with one paw curled nearby. Warm, intimate, macro-style pet photography with natural lighting and rich texture detail.

    9

    The snowy street in Brooklyn

    It’s a classic winter scene

    Prompt

    A quiet residential street in Brooklyn during a heavy snowfall. Rows of classic brownstone townhouses line both sides, with stoops, railings, and window details covered in fresh snow. Soft yellow streetlamps glow through the falling snowflakes. A single person walks down the center of the snow-blanketed street with a small dog, leaving faint tracks. Bare winter trees arch above the street, and the Brooklyn Bridge rises faintly in the snowy haze in the distance. Soft, overcast lighting, serene, cinematic, high-resolution winter scene.

    8

    A young child drawing with crayons

    Not quite keeping it all on the paper

    Prompt

    A young toddler sitting at a small wooden table in a bright, cozy room, drawing on white paper with colorful crayons. The child wears denim overalls over a striped long-sleeve shirt, with a few smudges of crayon on their cheek and nose. They concentrate intently while making squiggly lines. Soft natural light comes through a window with sheer curtains in the background. Out-of-focus toys, blocks, and a teddy bear sit on the floor behind them. Warm, gentle, homey atmosphere, shallow depth of field, high-resolution lifestyle photograph.

    7

    Beaches and cliffs on a summer’s day

    Heading to a secret beach?

    Prompt

    A bright, clear summer day along a rugged coastal path overlooking turquoise-blue water. A dirt trail winds through green grassy cliffs and flowering shrubs, leading toward a small sandy cove nestled between rocky headlands. Jagged dark rocks meet the sea below, where patches of shallow aqua water contrast with deeper blue. Rolling green hills rise in the distance beneath a cloudless blue sky. A few hikers walk along the path near a small stone building. High-resolution coastal landscape photography with vivid colors and crisp natural light.

    6

    A rainy night in an Asian city

    So vibrant, so wet

    Prompt

    A vibrant, rainy night street scene in an East Asian city. Crowds of people walk under umbrellas along a wet, reflective sidewalk lined with bright neon signs and warmly lit storefronts. On the right, a traditional-style building glows with rows of red lanterns, casting golden light onto the street. On the left, a line of trees arches over the walkway, their glossy leaves lit by streetlights. Motorbikes park along the curb, and traffic moves slowly through the rain. Reflections of red, yellow, and blue lights shimmer on the pavement. Dense atmosphere, busy urban energy, rich colors, high-resolution night photography.

    5

    A smartly dressed woman in a hotel room

    Heading out for business or brunch?

    Prompt

    A softly lit, modern bedroom with minimalist decor and warm neutral tones. A woman stands near a full-length mirror, barefoot on a textured grey rug, wearing a loose cream blouse and dark tailored trousers. She holds a folded blue sweater in one arm and looks downward thoughtfully. Behind her, sheer white curtains diffuse natural light across the room. A wooden console table with simple books and a ceramic vase sits against the wall. The bed in the foreground has neatly arranged grey and beige bedding. Calm, elegant, contemporary interior photography with subtle warm lighting and shallow depth of field.

    4

    The sun setting over the bay

    It’s golden hour

    Prompt

    A stunning coastal sunset overlooking a calm bay. The sun sits low on the horizon, casting a golden path of reflected light across the water. In the midground, a small rocky island rises from the sea, while distant headlands fade softly into the warm evening haze. In the foreground, a terrace with wooden tables, chairs, and glass railings sits above stone walls. A white building labeled “MOUNT’S BAY SAILING CLUB” stands to the right, with a tall mast beside it. A few people sit outside enjoying the sunset. Warm, golden-hour lighting, clear sky, tranquil seaside atmosphere, high-resolution photography.

    3

    Close up with the ship’s bell

    Every ship needs a bell

    Prompt

    A close-up, high-resolution photograph of an aged bronze ship’s bell with a weathered green-blue patina. The engraved text “ANDALUC…” is prominently visible in raised lettering around the upper band. Below the text, ornate decorative floral patterns wrap around the bell. Sunlight highlights the textured metal surface, revealing scratches, oxidation, and uneven coloration. The background is softly blurred, showing hints of wooden ship railings and indistinct people, giving a historic maritime atmosphere. Shallow depth of field, warm daylight, crisp macro-style detail.

    2

    A young couple taking a selfie on the beach

    Capturing those precious moments

    Prompt

    A joyful couple taking a selfie together on a sandy beach at sunset. They stand close, smiling warmly, with the woman in a light, flowy dress and the man in a casual T-shirt and shorts. The sun is low on the horizon, casting golden light across the waves and wet sand. Soft pastel skies fade from orange to pink. Palm trees line the distant shoreline, and gentle surf rolls in behind them. Warm, romantic, candid travel-photography style, high resolution, natural lighting, relaxed and happy atmosphere.

    1

    A courtyard at twilight

    Inviting with dark skies

    Prompt

    A serene twilight scene in an old courtyard lined with long, whitewashed buildings. Warm wall-mounted lights cast dramatic triangular beams upward along the textured walls, highlighting arched windows with dark metal grilles. The ground is a wide, empty dirt or gravel surface leading toward additional rustic buildings in the distance. Overhead, the sky is a deep, moody blue with heavy clouds, while silhouetted trees frame the background. Soft ambient glow from nearby houses adds warmth to the far end of the courtyard. High-resolution night photography with rich contrast, atmospheric lighting, and quiet, timeless charm.

    These are the AI images—did you guess everything?

    The prompts used for each image were all generated by ChatGPT, while the AI images were generated using Google’s Nano Banana. Did you manage to spot all of the AI images? It’s clear that some are easier to spot than others!

    Image

    Caption

    AI or Human

    A macro shot of a ginger cat

    📸

    The snowy street in Brooklyn

    🤖

    A young child drawing with crayons

    🤖

    Beaches and cliffs on a summer’s day

    📸

    A rainy night in an Asian city

    📸

    A smartly dressed woman in a hotel room

    🤖

    The sun setting over the bay

    🤖

    Close up with the ship’s bell

    📸

    A young couple taking a selfie on the beach

    🤖

    A courtyard at twilight

    📸

    Let us know in the comments how many you got right, and how you spotted the AI shots!

  • A week with ChatGPT Atlas convinced me to uninstall it

    A week with ChatGPT Atlas convinced me to uninstall it

    There’s been a wave of new “AI-powered” browsers lately, and I’ve tried pretty much all of them. Most of them promise a smarter, faster, more automated browsing experience, and end up feeling like the same browser with a chatbot glued on top.

    I genuinely thought OpenAI would be the one doing something different. So when the new ChatGPT Atlas browser launched, I gave it a full week as my daily browser. But just like Dia, Comet, and the rest, it left me disappointed. After seven days of trying to make it work, I ended up uninstalling it. And my search for a browser that actually feels fresh continues.

    Cool AI features don’t guarantee a better browser

    AI for the sake of AI isn’t enough

    Image by Raghav
    Credit: Raghav Sethi/MakeUseOf

    ChatGPT Atlas is pitched as the next leap in browsing, in the same category as things like Perplexity’s Comet. And on paper, sure, it sounds impressive. The browser pipes every search through ChatGPT, it has a sidebar that follows you from site to site, and it can even “act” on your behalf using its Agent mode. It is the same checklist of AI browser features we have seen a dozen times already. But once you use it for more than a week, the excitement wears off surprisingly fast. Because here is the thing: AI alone is not enough to make a browser better.

    Right now, every new browser (Dia, Comet, etc.) is doing the exact same trick. They take Chromium and layer a chatbot on top of it. That does not make the core experience any smarter, any faster, or any more enjoyable. It just adds another UI element to something that was already fine. After a while, it starts to feel repetitive. I am not switching browsers just to have another place to ask an LLM questions. I can already do that anywhere.

    Arc understood this. Before it added AI, it came up with genuinely useful improvements to how you work in a browser. Vertical tabs that made navigation feel cleaner, or Spaces that helped separate your work and personal browsing. Even when Arc added AI, the focus wasn’t on having conversations with an LLM. It was about making the browser itself smarter, such as automatically organizing your tabs for you.

    With Atlas, there is none of that. It feels like the whole browser exists just to say, “Hey, look, AI is everywhere now.” And that is where I check out. If a browser wants me to switch, it has to rethink the fundamentals. I am tired of having AI shoved down my throat. Everyone has that. I want to see something genuinely new again.

    Agent mode feels slow and underwhelming

    I ended up fixing its work anyway

    Credit: Amir Bohlooli / MakeUseOf

    The idea behind Agent mode is simple: instead of just answering your questions, the browser can actually do things for you. It can open sites, click through pages, fill forms, search for products, and supposedly complete tasks inside the browser without your input. On paper, that sounds like exactly where AI should be heading. A browser that does the boring steps for you.

    But in practice, Agent Mode feels slow and clumsy. It constantly trips over incredibly basic things. It sometimes fails to recognize that a page has a scroll bar, so it just sits there staring at the top of a page forever. Other times, it starts doing something and then gets stuck midway, like it forgot what the task even was. And when it does manage to complete something, it has this habit of drifting way off instructions, like it is improvising instead of following what you actually asked for.

    That is a huge problem, especially because Agent Mode is positioned as something that can help with everyday workflows, like adding items to your cart while shopping or filling data into forms. But if you eventually need to redo everything yourself anyway, what is the point of the agent being there at all?

    My breaking point was a simple task. I asked Agent Mode to take some data from a Google Sheets file and paste it into a table in Google Docs. A task that would have taken me ten seconds, maybe less. After over three minutes of watching it slowly move through menus, it finally “completed” the task. Except it didn’t paste the data. It created a summary of the sheet and generated a report explaining what it thought the sheet was about.

    And that is what makes Agent Mode frustrating. I could forgive it being slow if it were accurate. But in its current state, it’s wasting my time instead of saving it.

    There are some pretty major security risks

    OpenAI even admits there could be problems

    Credit: Amir Bohlooli / MakeUseOf

    The issue isn’t just that Agent Mode is unpredictable. When a browser is allowed to click, fill forms, and interact with websites on your behalf, the door opens to real security concerns. Because the agent reads the page to understand what to do, the page itself can influence the agent’s behavior.

    This is what’s known as prompt injection. A website can hide text that tells the agent to take a different action than what you asked for. So you might say something simple like, “Compare these two phones, and add the better one to my cart.” However, if the site has hidden instructions directing the agent to add the more expensive item to your cart or navigate elsewhere, the agent might actually follow those instructions instead.

    And because this whole concept of agentic browsing is still new, there isn’t a reliable, guaranteed way to prevent this yet. Even OpenAI openly notes on Atlas’ download page that using Agent Mode comes with risks and that you should be cautious.

    These are the last kinds of worries I want to have when I’m just trying to browse the web. If a feature isn’t consistent, and it might put my data or accounts at risk, the trade-off just doesn’t feel worth it.

    ChatGPT Atlas still hasn’t nailed the basics

    It’s Chromium with extra steps

    And this is where everything comes full circle. For all the talk about AI-first browsing, ChatGPT Atlas still struggles with the fundamentals. Strip away the ChatGPT sidebar and the agent features, and what you’re left with feels like a very plain Chromium browser. There’s no proper multi-profile support, no thoughtful tab management features, and not even something as simple as vertical tabs. That’s the part that makes the whole thing frustrating. In trying so hard to integrate LLMs into every corner of the browser, the basics have been ignored. Browsers are tools we spend hours on every day. The core experience has to be solid before you start layering on futuristic ideas. But Atlas feels like an AI layer sitting on top of something that could just as easily have been a Chrome extension. There’s almost no meaningful customization either. You can change your accent color, sure, that’s about it. Everything else feels unfinished, like the browser shipped before the foundation was actually ready. It needed more time, more polish, and more thought put into how people actually browse.

    Looking for something new again

    Safe to say I’ve stopped using Atlas entirely. My browser of choice used to be Arc, until it was replaced with Dia, which just never lived up to what made Arc special. So for now, I’ve taken temporary refuge in Chrome.

    It’s not exciting, but it doesn’t get in my way. I’m still hoping a browser comes along that actually rethinks how we browse instead of just adding an LLM everywhere. But until then, I’ll be waiting.

  • Hello world!

    Welcome to WordPress. This is your first post. Edit or delete it, then start writing!