XTHE

HomeKit Camera Security  

HomeKit cameras with an IP address send video and audio streams straight to iOS, iPadOS, TVOS, or macOS devices on the same local network. These streams are encrypted with randomly generated keys on both the device and the IP camera and exchanged over a secure HomeKit session when a device is not on the local network. The encrypted streams are sent through the Home Hub, which acts solely as a relay and does not decrypt them. When an app shows the HomeKit IP camera video to the user, HomeKit securely renders the video frames from a separate system process. This means the app cannot access or save the video stream, and apps are also not allowed to take screenshots of it.  

HomeKit Secure Video 

HomeKit offers a secure and private way to record, analyze, and view clips from HomeKit IP cameras without sharing video content with Apple or anyone else. When the IP camera detects motion, it sends video clips directly to an Apple device. Set up as a Home Hub, using a dedicated and encrypted local network connection. This connection uses a unique key pair for each HomeKit session, generated with HKDF-SHA-512. The Home Hub decrypts audio and video streams and analyzes video frames locally to detect important events. If something significant is found, HomeKit encrypts the video clip with AES-256-GCM using a randomly generated AES-256 key. Poster frames for each clip are also created and encrypted with the same key. The encrypted poster frame, audio, and video data are then uploaded to the iCloud server’s metadata for each clip, including the encryption key, which is uploaded to CloudKit using iCloud’s end-to-end encryption.  

For face classification, HomeKit stores all data used to identify a person’s face in CloudKit, protected by iCloud’s end-to-end encryption. This data includes each person’s name and their facial images. These images can come from a user’s photo if they choose to share it, or from an earlier IP camera video analyzed during a HomeKit secure video analysis session. This classification data helps identify faces in the secure video stream. From the IP camera, the identification details are added to the clips’ metadata as described earlier.  

When you use the Home app to view camera clips, the data is downloaded from iCloud, and the encryption keys are unlocked locally with iCloud’s end-to-end decryption. The decrypted video is streamed from the servers and decrypted on your iOS device before you see it. Each video clip session can be split into smaller parts, and each part is encrypted with its own individual key.  

Apple may be set to shake up the smart home market with its own security camera, despite already supporting third-party options through HomeKit Secure Video. For years, analysts and supply chain sources have suggested Apple is developing a camera on this system. If this could bring iCloud integration closer, stronger privacy, and better connections with Apple intelligence and automation  

This would be Apple’s first security camera, entering the market against Amazon’s Ring and Google’s Nest.  

Expected Release Timing 

Release timing remains to be clarified.  

Apple has scheduled a special Apple Experience event for Wednesday, March 4, 2026, with in-person sessions in New York, London, and Shanghai. Instead of a keynote, Apple plans to introduce two announced products over several days via press releases, then host hands-on demos at the event.  

Observers such as John Gruuber and Mark Gurman predict revelations daily before March 4, with the HomeKit camera likely to debut that week, either in a press release or at the event. The camera will ship immediately or later in the year. Apple often rolls out products in stages after announcements.  

With timing discussed, next is what this camera could actually offer. 

I still know confirmed specs, but rumors share some common themes:  

HomeKit integration: Rumors suggest the camera would work directly in the Home app with HomeKit Secure Video, including encrypted video streaming and recording through iCloud Plus. This aligns with Apple’s security camera standards. An Apple-made camera is expected to fit into this system.  

On-device intelligence: Multiple reports suggest Apple’s camera may use facial and motion recognition to improve automation. For example, some sources speculate it could trigger scenes when familiar people come or go. This would go beyond motion detection to provide greater context awareness, but the specifics remain unconfirmed.  

Automation and smart home: Rumors also indicate Apple’s security camera could connect with other smart home devices, such as presence of sensors, lights, or the upcoming HomePod hub, for tasks like turning on lights when someone enters a room. These integrations have not been verified.  

Current hardware features are not confirmed. Information on resolution, night-vision capabilities, or potential models for indoor and outdoor use remains unavailable.  

Price Expectations 

There are no confirmed leaks on pricing. Competing smart home cameras range from $100 to $300 or more, reflecting features like cloud storage or advanced sensors. An Apple camera would likely be positioned above basic third-party offerings, but pricing has not been confirmed.  

Apple is combining smart home features, cameras with face recognition, scene-triggering hubs, and multi-function speakers. The new camera may offer more than basic accessory features.

Source: Apple Security Camera Everything We Know: Specs, Price, and Release Date 

HomeKit camera security

At MWC 2026 in Barcelona, Lenovo is unveiling a new lineup of adaptive AI devices and innovative concepts for business professionals, creators, students, and gamers. Highlights include:  

• a modular PC  

• a glasses-free 3D laptop  

• a foldable gaming handheld  

• The first release of Lenovo Qira  

These products show how personal computing is moving toward systems that adjust intelligently to users and their surroundings.  

Across its portfolio, Lenovo continues to focus on delivering technology that is increasingly personalized, proactive, and protected while building a unified AI ecosystem that works naturally across devices.  

The AI era is not about one device or app, but about smart systems that work together across everything we use, said Luca Rossi, president of Lenovo’s Intelligent Devices Group. Lenovo and Motorola are fulfilling this vision in reality. By merging adaptive hardware with unified AI that works across species, smartphones, tablets, and wearables, and by introducing new mobile devices, innovative designs, and the launch of Lenovo Qira, we are building a broad AI portfolio to create better-connected, more intuitive experiences for everyone.  

Lenovo Qira Personal Ambient Intelligence 

Lenovo Qira is a personal ambient intelligence built at the system level and integrated directly into Lenovo and Motorola devices, rather than layered on as a standalone application operating across supported PCs, tablets, smartphones, and wearables. Qira is designed to help preserve continuity between tasks and devices while assisting based on user intent.  

Over the next few weeks, Lenovo Qira will become available on more than 20 Lenovo PCs, including Yoga, IdeaPad, Legend, and ThinkPad models. Through updates and pre-installed software, the IdeaTab Pro Gen 2 will be the first Lenovo tablet with Qira. At launch, Qira will support six languages in nine regions:  

• English (UK, US, India)  

• Spanish (Latin America, Spain)  

• French (France)  

• Italian (Italy)  

• German (Germany)  

• Portuguese (Brazil)  

Lenovo Qira will continue to evolve, reaching more devices and delivering new features and partnerships. In 2026, it will add more languages and devices, including its first appearance on Motorola smartphones, helping to build a unified ecosystem for both Motorola and Lenovo.  

Breakthrough Concepts: Exploring Adaptive Form Factors 

ThinkBook, Modular AI, PC Concept 

Lenovo is presenting a modern AI/PC concept that takes a flexible approach to business computing based on the idea of “carry small, use big”.  

This concept features a 14-inch ultra-thin base that supports various display configurations, detachable input modules, and modular ports. A second display can be attached in various ways or even replace the keyboard, expanding the workspace to about 90 inches while staying portable. This shows how modular design can support changing work needs and longer device life spans in AI-ready settings.  

Yoga Book Pro 3D Concept 

Lenovo is also showing a glasses-free 3D laptop concept that lets creators see and work with depth right on the screen.  

This concept combines dual-screen, AI-powered 2D-to-3D conversion, gesture controls, and easy-to-use creative tools. It makes immersive content creation smoother and shows Lenovo’s interest in spatial computing.  

Legion Go Fold Concept 

Lenovo is also introducing a foldable gaming handheld concept, expanding its work on flexible hardware.  

The device transitions from a compact handheld format into a larger immersive screen and supports multiple usage modes, including handheld play and split-screen multitasking. This device can switch from a small handheld to a larger immersive screen and support various use cases, including handheld gaming, split-screen multitasking, expanded display gaming, and desktop-style use. It shows how flexible screens can bring together gaming and productivity in one device.  

Commercial Portfolio: AI-Ready Platforms For Modern Work 

Lenovo is updating its commercial PC lineup to help organizations use AI-powered workflows more easily and at scale.  

The new ThinkPad T-Series comes with improvements that make it easier to service and use, and that better prepare it for AI. Some models have top iFixit repairability scores. These changes show Lenovo’s emphasis on long-term value, reduced downtime, and sustainable device management.  

The ThinkPad X13 detachable is built for flexible mobile work. It has pen support and replaceable parts, making it lightweight for people working on the go or in mixed settings.  

Lenovo is also moving into rugged devices with the new ThinkTab X11. This tough Android tablet is made for tough jobs in industrial and frontline settings.  

For small and medium businesses, the ThinkBook 14 2-in-1 Gen 6 offers flexible ways to work and AI-enabled tools for teamwork. The ThinkVision M16 portable monitor offers a lightweight, easy way to get more screen space for various work needs.  

Lenovo builds security, easy management, and supports services into all its business products. This helps organizations use AI-ready devices with more confidence and control.  

Consumer And Gaming Portfolio: Intelligent Experiences For Every Day 

Lenovo’s newest devices for creators, students, and gamers offer strong performance, easy portability, and smart features that simplify daily tasks.  

The Yoga 9i 2-in-1 Aura Edition (14” 11) has a high-end convertible design, vivid OLED screen, and smart features to help with creative work. The Yoga Pro 7A (15” 11) and IdeaPad Slim 5i Ultra (14” 11) add more AI-powered laptops to Lenovo’s lineup, offering both strong performance and easy portability for work and content creation.  

The Ideatab Pro Gen 2 brings smart learning and productivity tools in a flexible tablet design. It’s also the first Lenovo tablet to include Lenovo Qira.  

For gaming, Lenovo is adding the Legion 7A (15” 11 laptop) and the Legion Tab (8.8” 5 tablet). The Legion Tab has a 3K screen and optimized cooling for better performance.  

Advancing Smarter AI for All 

At MWC 2026, Lenovo is showing its plan to make AI more accessible for everyone by developing new hardware and rolling out Lenovo Qira. Lenovo aims to create a future in which technology adapts to people, not the other way around.  

To learn more about Lenovo’s new devices, solutions, and concepts from MWC2026, check out the Lenovo MWC 2026 Press Kit. 

Source: Lenovo Unveils Adaptive AI PCs, Modular Concepts, and Lenovo Qira Rollout at MWC 2026 

Local First Signals, data that tell search engines exactly who your business serves and where you operate, have become the new SEO goldmine. They directly address the modern AI-driven shift in search.  

Searches are now more hyperlocal and intent-based, such as “Near Me” or “Open Now”. By 2026, over 46% of all Google searches will have local intent. Businesses that use these signals can see over 500% ROI and often outrank bigger non-local competitors.   

Here is why local first signals are so valuable for SEO.  

High Intent Action-Oriented Traffic 

• Near me boom: Over 76% of people who conduct a local search on their smartphone visit or call a business within 24 hours.  

• Ready to buy column local searches show, and I need it now. Urgency showing up first in these searches attracts customers who are ready to buy, not just people browsing.  

AI and Voice Search Optimization 

• Contextual Answers: AI-powered Google searches use structured data, such as local business schema, to provide direct, accurate information about business locations, hours, and services.  

• Voice query alignment: local-first content that uses conversational, location-specific language, such as “best pizza in (neighborhood),” is more likely to be featured by smart assistants as voice search becomes more popular.  

The New Trust Layer (Beyond Links) 

• Reputation Management: By 2026, customer reviews will be a top-ranking factor. Having lots of recent positive reviews and consistent name, address, and phone number (NAP) details across directories shows that your business is real and trustworthy.  

• Verified Relevance: Creating local content, like blog posts about neighborhood events, partnerships with other local businesses, or community sponsorships, shows that your business is active in the area.  

Visibility In Zero-Click Searches 

• Pack dominance: local fast signals, such as an optimized Google Business Profile, help businesses appear in the local three-pack at the top of search results. This often means users get what they need without having to click through to a website.  

• Visual & entity-based: posting high-quality photos and videos and maintaining consistent updates on your Google Business profile help AI recognize your business as a real entity.  

The Startup Advantage 

• Speed & Agility: smaller, focused businesses can move faster than big brands, enabling them to update their content instantly to reflect local, seasonal, or community-based trends and gain a competitive edge.  

Key Local Signals to Maximize: 

• Google Business Profile (GBP) Optimization: The Top Ranking Factor in 2026.  

• Consistent NAP citations, ensuring name, address, and phone number are identical across all platforms.  

• Local reviews: focusing on frequency and content, noting particular services or neighborhoods  

• Local business schema markup: providing search engines with structured machine-readable data  

• Hyper Local Content: Blog posts about neighborhood events or local community engagement  

If you run a business in South Florida, you have probably noticed things are changing.  

Competition is tougher than ever, whether you are in the financial center of West Palm Beach, the busy hospitality scene in Delray, or the fast-growing tech community in Boca Raton.  

By 2025, just having a great product or service won’t cut it if customers can’t find you when they search for the best (service near me) on their phones. Your business might as well be invisible.  

That’s where local SEO comes in. Knowing how to use it is quickly becoming one of the most important skills in our local economy.  

What is Local SEO? 

Local SEO (Search Engine Optimization) means making your business easy to find in local online searches. While general SEO aims for an international audience, local SEO focuses on reaching customers in your area, like Jupiter, Wellington, or downtown West Palm.  

Local SEO targets the specific geographic map pack, while organic SEO focuses on global search results.  

Here is an example: General SEO might help a person in California find you. Local SEO helps the person walking down Clematis Street find you when they are looking for lunch or a graphic designer.  

Why Local SEO Is Do or Die in 2025 

The way people find businesses has changed a lot. Here’s what that means for businesses in Palm Beach County:  

• The “Near me” boom. Corona searches for “Near me” have jumped by 500% in recent years. If a tourist in Palm Beach looks up “Italian dinner”, Google won’t show them a restaurant in New York. Instead, it shows what’s just five minutes away.  

• Mobile roles: With so many tourists and seasonal residents, most people search on their phones. They are looking for things while at the beach, on the golf course, and even in their cars.  

• The Map Pack Advantage: Nearly half of all Google searches are local. The top three spots in the Google Map Pack: The map at the top of search results gets most of these clicks.  

Every local business is competing for a spot in the map pack. It is the most valuable place to be.  

The Three Pillars of Ranking Locally 

Ranking in Palm Beach County: Getting your business to rank in Palm Beach County isn’t magic. It takes technical know-how and good reputation management, and (GBP).  

Your Google Business Profile is like your new home page. It powers the map results, so make sure it is fully updated and within the right hours, services, and great photos.  

Reputation (Reviews) 

In South Florida’s service-driven economy, reputation matters a lot. Getting regular, genuine, positive reviews tells Google your business is trusted in the community.  

Technique 

This is where many businesses struggle. Google lowers the ranking of websites that are slow, hard to use on phones, or have bad code. If your site takes five seconds to load on a tourist’s phone, Google won’t show it.  

How Coding and Digital Skills Power Local SEO 

This is why marketing and web development need to work together. You can’t have strong local SEO without a solid technical base.  

At Park Beach Code School, we see our students use their skills to help local businesses grow and succeed.  

The Web Developer’s Role (The Technical Foundation) 

In our web developer program, students learn the behind-the-scenes skills that directly affect search rankings. First, design and code websites that look perfect on any screen size.  

• Site speed: producing clean HTML, CSS, and JS so pages load quickly, which is a big factor for Google rankings.  

• Schema markup: adding special code that tells Google exactly where your business is and what it offers.  

The Digital Marketers’ Role (the Strategy) 

In our Social Media Marketing Specialist Program, students learn how to attract more visitors and engage more people, rather than focusing solely on AC repair.  

• Content Creation: Column, Writing Blogs, and Social Posts That Show Your Business Is Relevant and Active in the Local Area.  

• Analytics: looking at data to find out exactly where your customers are coming from.  

The main takeaway: in 2026, local SEO rankings will focus on accuracy, clarity, and being active in your community to get started. Review your Google Business Profile and citations this week. Next week, publish a high-intent service page and begin a review request process that fits within your daily routine. If you want help with planning and tracking, you can find businesses by putting these local SEO strategies into action and supplying clear, measurable reports.

Source: Local SEO Ranking Factors Breakdown: How to Rank for Near-Me Searches 

Advanced Micro Devices (AMD) and Meta (Meta) announced a major multi-year agreement on February 24th, 2026, to run up to 6 GW of GPU capacity across Meta’s global data centers. This deal, valued between $60 billion and $100 billion over five years, marks the end of single-vendor dominance in high-end AI computing and positions AMD as an important player in global AI infrastructure.  

The deal centers on the mass deployment of AMD’s next-generation MI450 accelerators and sixth-generation EPYC Venice processors. By securing 6 GW of power capacity, roughly equivalent to the energy consumption of 6 million homes, Meta is doubling down on its capacity-open-compute philosophy while weaning its supply chain away from NVIDIA Corporation. For large-cap tech investors, the agreement represents the most significant validation of AMD’s AI roadmap to date, suggesting that the NVIDIA alternative narrative has officially transformed into a dual-vendor duopoly reality.  

The 6GW Era: A Breakdown of the Historic Partnership 

Today’s agreement is a result of three years of technical collaboration, starting with Meta’s early use of the AMD Instinct MI300X in 2023. In 2025, the companies worked together to create the Helios Rack Scale architecture, which improves power delivery and liquid cooling for large graphics processing unit clusters. The 6GW deal is beyond a purchase; It lays the foundation for Meta’s future personal superintelligence projects that require massive amounts of real-time computing power.  

Analysts say that 6 GW of power will support about 2.4 to 3 million GPUs. Unlike earlier, smaller orders, this agreement secures a long-term plan, giving Meta priority access to AMD’s CDNA4 and CDNA5 architectures until 2030 to strengthen the partnership. AMD has given Meta performance-based warrants for up to 160 million shares, which will vest as Meta meets certain deployment goals. This move further aligns the interests of both companies.  

The market responded quickly. AMD shares rose by 9.4% in early trading, and Meta’s stock gained 3.2% as investors welcomed the company’s efforts to lower its AI infrastructure costs. The deal also shows a change in how the industry measures scale. In 2026, the main benchmark for AI leadership is now gigawatts of deployed power rather than nearly counting chips, indicating the serious energy constraints facing data centers worldwide.  

Winners, Losers, and the Silicon Bifurcation 

AMD is the biggest winner in this deal, with its AI accelerator market share expected to rise from about 9% in 2025 to over 15% by the end of 2026. By bringing Meta on board as a major customer, AMD has overcome its biggest challenge: proving its software works at scale. Meta’s move to run its Llama 4 and Llama 5 models on AMD’s ROCm software now signals to other large cloud providers like Microsoft and Alphabet that they can do the same.  

The Chain of Events extends well beyond GPU designers. Networking giant Broadcom Inc. is a major hidden winner, as its high-end Ethernet switching and PCIe Gen7 fabric serve as the connective tissue for these 6 clusters. Similarly, Vertiv Holdings Corp and Eaton Corporation plc are expected to see a spike in demand for specialized liquid-cooling and high-voltage power distribution systems required to manage the high thermal profiles of MI450, which can draw upwards of 1.2 kW per accelerator.  

On the other hand, Intel is still struggling in the data center market. Meta’s decision to use AMD’s Venice EPYC CPUs with its GPUs shows that Intel is losing its usual hold on the server head node market, while NVIDIA remains the top player. This Meta deal brings real price competition for the first time in the generative AI era. NVIDIA may have to rethink its pricing for the upcoming Rubin chips to keep customers from leaving.  

Power as the new currency of AI 

This 6 GW deal fits into a wider industry trend in which electricity availability has become the ultimate bottleneck for AI progress in early 2026. Lead times for new grid interconnections in major hubs, like Northern Virginia, have stretched to nearly a decade. By securing 6 GW of capacity through a multi-year partnership, Meta is effectively land-grabbing power, making sure that its AI development will not be threatened by energy shortages that are currently stalling smaller competitors.  

The deal also spotlights the regulatory and environmental issues that large-cap tech companies face. Running 6 GW of computing power means companies must rethink how they get their energy. Meta says much of this new capacity will come from on-site solar, large battery storage, and even investments in small nuclear reactors. As tech firms start generating their own power, they are becoming more like utility companies, which is attracting more attention from regulators and legislators. While Microsoft and Intel once dominated the PC landscape, the AI era is proving to be more fragmented. The Meta AMD alliance proves that the software layer, specifically open-source frameworks like PyTorch, has become robust enough to bridge the gap between different hardware architectures, breaking the proprietary mode that many believed would protect NVIDIA indefinitely.  

The Road to 2030: What comes next? 

In the short term, the market will be watching for the first 1 GW milestone deployment, expected in the second half of 2026. Success here will depend on AMD’s capacity to maintain a steady supply chain amidst worldwide semiconductor fluctuations. For Meta, the challenge will be embedding this massive influx of compute into its consumer-facing products without a corresponding spike in energy costs. That could erode margins.  

Looking further ahead, the deal is likely to trigger a copycat effect among other Tier 1 cloud providers. If Meta can demonstrate that AMD hardware delivers equivalent performance at a 20-30% lower TCO, the pressure on Amazon and Google to diversify their own fleets will become irresistible. We may also see a strategic shift toward custom silicon partnerships in which AMD delivers foundational IP for Meta to build even more specialized chips. This move would further squeeze traditional merchant silicon providers.  

In the long run, the market is shifting from being limited by hardware to being limited by power. As computing becomes more common, thanks to AMD and Meta’s competition, tech giants will compete based on who can achieve the most efficient energy use and access the best data to train their AI models.  

Investor Takeaway: A New Market Equilibrium 

The 6 GW deal signed on February 24, 2026, ends the AI monopoly era. Investors should see this as proof that the AI infrastructure market is becoming more competitive and mature. AMD’s rise from an underdog to a $100 billion partner for Meta shows its strategy is working and sets up a direct rivalry with NVIDIA for the top spot among chip makers.  

Going forward, investors should focus on AMD’s quarterly AI data center revenue and how efficiently Meta spends on capital projects. The race for AI leadership is now about who can deploy the most computing power most efficiently, not just who has the fastest chip for the wider market. This tale shows that the AI boom is not yet over, with the biggest companies still investing at record levels.

Source: AMD Shatters Nvidia’s Monopoly: Landmark 6-Gigawatt GPU Deal with Meta Reshapes AI Landscape 

Artificial Intelligence is becoming increasingly important for small businesses, especially as AI tools are integrated into everyday apps like Microsoft 365 Pro Pilot. According to Microsoft’s latest survey, one-third of SMEs see AI as a top priority, and 71% plan to accelerate their AI investments over the next year.  

Although still new, the benefits of AI for SMBs, such as simplifying both common and complex tasks, are already clear. According to a Microsoft survey, 84% of respondents believe AI has a positive impact on their business, and those who regularly use it see a 40% in-product input. The real question is not if tools like Microsoft Co-Pilot can give SMBs a competitive edge, but rather how.  

How Does Microsoft 365 Co-Pilot Help Small Businesses Get the Most Out of Their Investment 

Get the most from AI tools like Microsoft 365 Co-Pilot. It helps to know what they can do. Microsoft keeps adding new features, but there are already many ways small businesses can use Co-Pilot today.  

There are some ways small businesses are using Microsoft Co-Pilot to boost productivity and encourage new ideas:  

• Creating an Effortless Experience Across Microsoft 365: Co-Pilot uses Graph Grounding to improve its responses by pulling in extra context extracted from your stored information. It does this by using Microsoft Graph, which connects to your data across Microsoft 365  

• Content Creation: Co-Pilot in Word saves time by paraphrasing text, providing further details, or including helpful statistics in your documents  

• Using email conversations: Co-Pilot can find emails by topic or form, summarize meetings, create to-do lists, and assist with crafting or editing messages, saving you valuable time.  

• Increasing Response Rates for Forms/Surveys/Polls: Canon Co-Pilot suggests follow-ups, such as sending reminders or writing messages, to encourage people to respond.  

• Reviewing and summarizing videos: for Teams meetings. Co-Pilot highlights key points adds time markers for topics or speakers, lists tasks, and can search specific topic words or phrases.  

• Automating Business Tasks: You can build Co-Pilot extensions or plug-ins to work with data from different sources and targeted searches. Use natural language queries and set up custom prompts based on your data.  

• Decreasing manual typing and formula work: Co-Pilot in Excel can use voice prompts to help analyze data, suggest formulas, show insights in charts and pivot tables, and add important data. It can automatically create and calculate formulas for complex spreadsheets based on your instructions.  

• Improving content management and data governance: Co-Pilot for restricted SharePoint search lets you turn off organization-wide search and limit enterprise search and Co-Pilot to only the SharePoint sites you specify, no matter which search option you use. Users can still work with files and content they own or have accessed before in Co-Pilot.  

• Gating Insights from Employee Feedback: Co-Pilot can record comments and help you add on suggestions using summaries, reports, and analysis from across Microsoft 365 Apps.  

• Making budget management easier: Co-Pilot in Planner (review) helps you plan, manage, carry out, and review major projects more smoothly.  

• Training employees to write better prompts: Copilot lab helps teams understand what drives successful AI-powered business practices.  

• To develop secure applications, Microsoft 365 Copilot can write and analyze code, troubleshoot code, and handle data or security issues. Knowing how to use fonts is important.  

This list is just a start. New features and updates are added every month. To learn more about what Co-Pilot can do, take a look at Microsoft’s AI Co-Pilot Success Kit.  

How SMBs Can Use Microsoft 365 Co-Pilot To Drive Greater Success 

Knowing what the product can do is just the beginning. A recent report on SMB technology adoption found that 77% of SMBs see a lack of knowledge as the main barrier to AI success. Early adoption and ongoing training with tools like Microsoft 365 Co-Pilot can help SMBs stand out.  

If you are new to AI and want to get the most from your investment, try Microsoft’s free Co-Pilot training modules. If you’re unsure how Co-Pilot could help your business, take a look at these industry-specific examples  

If you work in product-based or supply chain industries like retail, manufacturing, or transportation, Co-Pilot can help you:  

• Predict demand changes with greater accuracy based on your sales data.  

• Improve delivery routes through analyzing financial and geographical data.  

• Monitor weather and world events that impact the supply chain, including situations influencing your materials, carriers, or distribution networks.  

• Automatically draft emails to alert and engage partners or customers, reducing the likelihood of disruptions.  

• Support customer service through the creation and analysis of support forms, surveys, and polls.  

• Generate information on order status, delivery schedules, and shipping options.  

For service-based industries like insurance, legal, consulting, or real estate, Microsoft 365 Co-Pilot offers support in several ways:  

• Draft review and summarize essential documents based on business and industry data.  

• Compiling a deck for marketing campaigns, targeted sales outreach, or other important customer-facing messaging.  

• Translate documents, presentations, or sales and marketing collateral into 27+ languages.  

• Develop Captivating PowerPoint presentations based on industry data and notable themes.  

• Formulate comprehensive project management plans, including automated progress reports.  

• Respond to client inquiries more efficiently with AI-powered suggestions.  

Microsoft 365 Co-Pilot is more than just a product update; it’s a real gamechanger for SMBs. Microsoft continues to improve it with more language options, new features, and better performance. To stay up to date, check the Microsoft 365 roadmap.  

How Partners Can Make the Best Use of This Opportunity 

Microsoft 365 Co-Pilot gives resellers and managed service providers a great way to grow their businesses. Partners can boost their earnings by setting up productivity pools, adding business premium features, and offering training to help organizations start using Microsoft 365 Copilot. Partners can deliver important services, such as training users to write better prompts and use AI tools. Partners may also qualify for the Tier 2 Accelerator, which offers a 10.75% rebate above the user’s profit margin. Running SMB workshops and delivering ongoing support can make your services even more valuable and help you build steady revenue. This opportunity enables partners to earn more and take advantage of different incentives and rebates.

Source: How SMBs can use Microsoft 365 Copilot for a competitive advantage 

AMD is adding to its Ryzen AI line with the new Ryzen AI 400 Series and Ryzen AI Pro 400 Series desktop processors. These are the first processors designed for next-gen AI PC applications and support Co-Pilot Plus PC features.  

• OEM partners are launching new AI PCs, including enterprise notebooks and mobile workstations, powered by Ryzen AI Pro 400-series mobile processors.  

• Ryzen AI 400 series mobile processors offer up to 30% faster multi-threaded performance vs other processors. This helps professionals finish demanding tasks more quickly and still enjoy all-day battery life.  

• AMD now provides its widest range of enterprise PC solutions supported by AMD, the AMD PRO platform. This platform improves security, manageability, and resilience for large-scale AI PC rollouts.  

At Mobile World Congress 2026 in Barcelona, AMD introduced new additions to its Ryzen AI lineup: the Ryzen AI 400 series and Ryzen AI Pro 400 series desktop processors. These processors offer strong on-device acceleration and next-level performance, letting users run AI applications and large language models locally. The Ryzen AI 400 series can handle demanding tasks like design and engineering smoothly. AMD is also adding workstations to the Ryzen AI 400 series mobile lineup.  

With these new processors, OEMs can now offer next-generation AI PCs, including high-performance desktops, laptops, and mobile workstations, built for today’s workloads.  

Agent AI 400 series processors are the first desktop AI chips to support Microsoft Co-Pilot plus PC features. They include a neural processing unit (NPU) that delivers up to 50 TOPS3 of AI computing power. This lets users run AI assistants and productivity tools directly on their PCs, keeping sensitive data on their devices and offering better control, performance, and privacy.  

The desktop PC is evolving from a tool you use to an intelligent assistant that works alongside you, said Jack Huynh, Senior Vice President and General Manager of the Computing and Graphics Department. With the Ryzen AI 400 series of processors, the world’s first design to power New Co-Pilot + experiences on the desktop. They are bringing powerful AI acceleration that enables our partners to build systems that empower both enterprises and consumers to do more and create more.  

The World’s First Co-Pilot+ Desktop Processor For Next Gen AI Applications 

The AMD Ryzen AI400 series of desktop processors are built to ensure scalable performance and smart features for professional tasks. They combine high-performance Gen5 CPU cores, AMD RDNA 3.5 graphics, and a dedicated AMD XDNA2 NPU. This mix gives office professionals, developers, and power users the speed, efficiency, and local AI acceleration they need. Whether it’s multitasking, collaborating, software development, data analysis, or AI-assisted work, these processors provide continuous performance and on-device intelligence for today’s Desktops.  

Ryzen AI 400 Series Desktop Availability 

M5 desktop systems with Ryzen AI 400-series processors are expected to be available from OEMs such as HP and Lenovo starting in the second quarter of 2026.  

Extending Ryzen AI Leadership Across Notebooks And Mobile Workstations 

AMD’s commercial PC lineup is expanding, as OEM partners are launching notebooks powered by Ryzen AI Pro 400-series mobile processors. The Ryzen AI9 HX Pro 470 offers up to 30% faster multi-threaded performance than the Intel Core Ultra HX7 3581, helping users work through demanding tasks faster. These systems also provide strong power efficiency and all-day battery life, so productivity lasts throughout the workday. With these features, Ryzen AI Pro 400 series processors bring AI acceleration and Co-Pilot + PC experiences to mobile devices.  

As more organizations use AI in their daily work, Ryzen AI Pro 400 series mobile processors make it easy to run advanced AI tasks directly on commercial laptops and workstations. With a powerful NPU that delivers up to 60 TOPS of AI compute, these processors use on-device AI acceleration to boost efficiency. This helps businesses increase productivity while keeping the pace and responsiveness professionals need.  

Ryzen AI Pro 400 series mobile processors will also be used within next-generation workstations, bringing the same high performance to professional systems that support independent software vendors (ISVs). These mobile workstations are built to speed up professional applications by using all available computing resources (CPU, NPU, and GPU) for challenging tasks in engineering, creative work, and technical fields.  

Ryzen AI Pro 400 Series Mobile Workstation Availability 

Mobile workstations with Ryzen AI Pro 400 series processors are expected to be available from Dell Technologies, HP, and Lenovo starting in the second quarter of 2026.  

Strengthening Enterprise Security and Manageability with AMD Pro 

AMD Pro offers enterprise-level security, manageability, and dependability with hardware and software built to make IT operations easier and protect long-term investments. AMD is improving the Pro platform by upgrading both its hardware and software to help IT teams manage an AI-enabled PC fleet. New remote management features give IT administrators greater visibility and control, enabling them to resolve issues and keep the business running smoothly without visiting each desk.  

AMD systems work with most major commercial security solutions, making it easier to add them to existing enterprise environments. This helps organizations protect their device fleets within their current security systems.

Source: AMD Gives Consumers and Businesses More AI PC Options with Expanded Ryzen™ AI 400 Series Portfolio 

Ericsson and Intel are working together on AI-powered network innovation covering compute, connectivity, cloud, and standards across the core network, RAN, and Edge.  

• This partnership strives to make the move to 6G more open, efficient, and affordable for operators and the wider industry.  

Ericsson and Intel are combining their cutting-edge technologies to help the industry prepare for a smooth transition to AI-powered 6G networks and applications.  

The companies announced this extended partnership at Mobile World Congress Barcelona, May 2026. Building on their decades-long relationship, their work will cover mobile connectivity, cloud technology, and computing for AI-driven R&D and core network users. They will also focus on platform security and network features to accelerate cloud-native solutions for the industry.  

Borje Ekholm, President and CEO of Ericsson, says 6G is not simply an iteration of mobile technology. It is the infrastructure that will distribute AI across devices, the edge, and the cloud. Ericsson’s long history of network innovation and large-scale operator deployments enables us to lead practical integration across the value chain and move 6G from research into commercial reality.  

Indip Bhutan, CEO of Intel, says Intel’s ambition is to be the undisputed technology leader toward unifying RAM, core, and edge AI to enable an effortless transition to AI-native 6G environments. Together with Ericsson, we will continue to demonstrate that the future of network connectivity is open, power-efficient, secure, and grounded in intelligent AI influence. With future Ericsson Silicon powered by Intel’s most advanced process nodes, ongoing multi-year research plans, and flexible AI- and RAN-ready Cloud RAN powered by Intel Xeon, we are well on our way to delivering the future performance, efficiency, and supply security that the world’s leading operators require.  

A Joint Commitment 

6G moves from research to action. The industry needs to work together, leveraging global standards, to turn new ideas into operational infrastructure.  

This partnership will help develop high-performance, power-saving computing for both AI-powered networks and networks that support AI.  

AI-powered 6G will bring together smart, flexible networks, advanced computing, and real-time sensing. This will lay the basis for faster, more efficient, and capable services. In the future, sensing and computing could work even more closely together across the network.  

Showcasing Collaboration Results 

Xeon and Intel are key milestones together in Cloud R&D, 5G core, and open network infrastructure at MWC26. They will pursue this progress with several demonstrations at the Ericsson Pavilion(Hall 2), Intel (Hall 3, stand E3E31), and other partner event spaces, presenting their innovative teamwork.  

About Ericsson 

Fixing networks, connecting billions of people every day for 150 years. We have changed the way we communicate. We deliver mobile communication and connectivity solutions for service providers and businesses. Working with our customers and partners, we help build the digital world of tomorrow.  

About Intel 

Intel leads the industry by creating technology that advances progress and improves lives. Guided by Moore’s Law, we keep advancing semiconductor design and manufacturing to serve our customers’ biggest challenges. By adding Intel agents to the cloud, networks, edge, and all types of devices, we help data transform businesses and society. To learn more about Intel’s innovations, visit intnewsroom.intel.com and intel.com.

Source: Ericsson and Intel collaborate to accelerate the path to commercial AI-native 6G 

On March 2, NVIDIA announced a US$2 billion investment in Lumentum Holdings, a leader in optical components. This deal, involving the purchase of convertible preferred stock, constitutes a major change in building advanced AI data centers. The industry is now shifting from copper wiring to faster light-based optical connections.  

The news had an immediate impact on the semiconductor and networking industries, causing a strong rally among smaller optical companies that support the internet. Investors saw that the next stage of AI growth would rely on laser technology and silicon photonics. Shares of Applied Optoelectronics rose by 22.4%, and nLight increased by 16.4%. The market now sees light-based technology as key to building large-scale data centers.  

A Closer Look at the $2 Billion Photonics Deal 

Investment is more than a simple cash infusion; it’s a strategic marriage between the world’s most valuable chip maker and its primary supplier of light. Under the terms of the agreement, NVIDIA purchased 2,876,415 shares of Lumentum Holdings’ newly issued Series A convertible preferred stock at $695.31 per share. A parallel $2B investment matches this $2B layout in Coherent Corp., representing a total $4B bet by NVIDIA on the laser and photonics industry in a single day.  

The deal comes at an important time for AI development. In 2025, the industry found that copper cables, once standard for server connections, became too hot and slow for the high data rates needed by 1.6 terabit optical networks. By the end of 2025, leaders realized that scaling AI clusters to millions of GPUs would require moving data using photonics (light) rather than electronics (electricity). NVIDIA’s investment will help Lumentum build a modern US wafer fabrication facility, securing a domestic supply of advanced lasers for co-packaged optics (CPO).  

Market participants reacted with imminent fervor. While Lumentum Holdings itself rose nearly 10%, the REIT investors responded quickly to the news. Lumentum Holdings shares rose nearly 10%, and the impact on the rest of the sector was even stronger. The investment shows that NVIDIA is working to control more of the networking process, similar to its approach with the H100 and B200 GPU platforms. By securing supply from Lumentum and Coherent, NVIDIA is making it harder for competitors to get the key parts needed for high-performance AI clusters. Stakeholders call this a supply chain moat that provides the optical transceivers and laser diodes essential for fiber-optic communications. Applied Optoelectronics became the rally’s poster child, gaining 22 million as analysts highlighted its role in producing 800G and 1.60 transceivers that will now be in higher demand. Similarly, nLight saw a 16% boost, powered by expectations that its high-powered semiconductor lasers will be integrated into the next generation of NVIDIA’s NVLink interconnects.  

NVIDIA has announced new multi-year agreements with Lumentum Holdings to promote innovation in advanced optics technologies. The partnership will focus on research and development to support the next generation of AI infrastructure and system designs.  

The agreement is non-exclusive and includes a multi-billion-dollar purchase commitment from NVIDIA, as well as future access to advanced laser components. NVIDIA is also investing $2 billion in Lumentum to support research and development, expand capacity, and help build a new U.S.-based manufacturing facility.  

Optical interconnect technology and package integration are essential for scaling AI factories and making large AI networks more energy efficient and reliable. This partnership combines NVIDIA’s expertise in AI computing and networking with Lumentum’s strengths in optics and manufacturing. The investment will help Lumentum grow its manufacturing and research to support future AI data centers.  

AI has reinvented computing and is driving the largest computing infrastructure build-out in history, said Jensen Huang, founder and CEO of NVIDIA, together with Lumentum. NVIDIA is advancing the world’s most sophisticated silicon photonics to build the next generation of gigawatt-scale AI factories.  

The multi-year strategic agreement demonstrates our joint commitment toward advancing the optics technologies that will power the next generation of AI infrastructure, said Michael Hurston, CEO of Lumentum. In support of this cooperation, we are also investing in a new fabrication facility to increase capacity and accelerate innovation. We are excited to work together to expand what’s possible for the AI optical architecture of tomorrow.  

About Lumentum 

Lumentum is a global leader in optical and photonics technologies that support the networks and infrastructure for AI, cloud computing, and next-generation communications. With decades of experience in photonics, Lumentum provides high-performance lasers, modules, and optical sub-systems for expandable, energy-efficient data centers, advanced telecom networks, industrial manufacturing, and sensing. The company is based in San Jose, California, and has R&D, manufacturing, and sales facilities around the world.

Source: NVIDIA Announces Strategic Partnership With Lumentum to Develop State-of-the-Art Optics Technology 

The Photonics Pivot: Nvidia Injects $2 Billion into Lumentum to Secure the Future of AI Networking

News 

• Amazon Web Services (AWS) and OpenAI are collaborating to build a stateful runtime environment powered by OpenAI models. This new tool will be available on Amazon Bedrock, so AWS customers can create generative AI applications and agents at scale.  

• AWS will be the only third-party cloud provider for OpenAI Frontier. This platform lets organizations build, deploy, and manage teams of AI agents.  

• OpenAI will use 2 GW of Trainium capacity on AWS to meet demand for a stateful runtime environment and other advanced workloads.  

• OpenAI and Amazon will work together to create custom models that will power Amazon’s customer-facing applications.  

• Amazon plans to invest $50 billion in OpenAI.  

OpenAI and Amazon have announced a multi-year partnership to speed up AI innovation for businesses, startups, and consumers worldwide. Amazon will invest $50 billion in OpenAI, starting with $15 billion now and an additional $35 billion in the next few months, subject to certain conditions.  

OpenAI and Amazon are partnering to bring advanced AI capabilities to both and are working together to develop a state-of-the-art runtime environment using OpenAI models. This will be available through Amazon Bedrock.  

Stateful developer environments represent the next step in using advanced models. They let models access resources such as compute, memory, and identity. With a stateful runtime environment, developers can retain context, preserve previous work, use multiple software tools and data resources, and access compute resources. These environments are built to support active projects and workflows.  

These stateful developer environments will be trained to run optimally on AWS’s infrastructure and integrated with Amazon Bedrock, Agent Core, and other AWS services, enabling customized AI applications and agents to run cohesively with the rest of their applications. The stateful runtime environment is expected to launch in the next few months.  

Bringing OpenAI’s most advanced enterprise platform to AWS customers 

AWS will be the only third-party cloud provider for OpenAI Frontier, making OpenAI’s most advanced enterprise platform more widely available as demand for AI grows across industries.  

Frontier enables organizations to build, deploy, and manage teams of AI agents that work across real business systems, with shared context, built-in governance, and strong security, all without managing the underlying infrastructure. As companies move from testing to production, the use of AI frontier makes it easy to quickly, securely, and at scale add powerful AI to existing workflows. Trainium powers growing Amazon customer demand.  

OpenAI and AWS are expanding their existing $38 billion multi-year agreement by $100 billion over 8 years. The expansion includes OpenAI committing to consume approximately 2 GW of Trainium capacity via AWS infrastructure, supporting demand for stateful runtime, frontier, and other advanced workloads. This agreement lowers the cost and improves the efficiency of producing intelligence at scale.  

With this agreement, OpenAI secures long-term capacity and works with AWS to use custom-built silicon chips alongside AWS’s broader compute ecosystem. This lets businesses use AI on demand without managing the underlying infrastructure.  

This agreement covers both Trainium 3 and the next-generation Trainium 4 chips, which will support a wide range of advanced AI workloads. Trainium 4, expected to be available in 2027, will offer much better performance, more memory bandwidth, and higher memory capacity to support more powerful AI systems at scale.  

Custom Models Available To Power Amazon’s Customer-Facing Applications 

OpenAI and Amazon will collaborate to develop custom models for Amazon developers to use in customer-facing applications. Amazon teams will be able to adapt OpenAI models for different AI products and customer-facing agents. These new models will add to the options already available to Amazon developers, like the Nova family, giving teams more tools to build and deliver at scale.  

OpenAI and Amazon share the belief that AI should appear in ways that are practical and genuinely useful to people, said Sam Altman, co-founder and CEO of OpenAI. Combining OpenAI’s models with Amazon’s infrastructure and worldwide reach helps us put powerful AI into the hands of businesses and users at real scale.  

We have lots of developers and firms eager to run services powered by OpenAI models on AWS. Our unique collaboration with OpenAI to provide straightforward runtime environments will change what’s possible for customers building AI apps and agents, Andy Jassy, president and CEO of Amazon, said. We continue to be impressed with OpenAI and what it is building, and we are excited not only about its decision to go big on our custom AI silicon (Trainium), but also about our opportunity to invest in the company and partner with it over the long term.  

About OpenAI 

OpenAI is an AI research and deployment company. Our mission is to ensure that artificial general intelligence benefits all of humanity.  

About Amazon 

Four principles guide amazon:  

1. Customer Obsession rather than Competitor Focus  

2. Passion for Invention  

3. Devotion to Operational Excellence  

4. Long Term Thinking  

Amazon strives to be Earth’s most Customer-Centric Company, Earth’s Best Employer, and Earth’s Safest Place to Work.  
 
Customer Reviews:  

• 1-Click Shopping  

• Customers’ Recommendations  

• Prime Fulfillment by Amazon  

• AWS  

• Kindle Direct Publishing  

• Kindle Career Choice  

• Fire Tablets  

• Fire TV  

• Amazon Echo  

• Alexa  

• Workout Technology  

• Amazon Studios  

• The Climate Pledge  

These are among the innovations Amazon pioneered. For more information, visit amazon.com/about and follow @amazonnews.  

About AWS 

Amazon Web Services (AWS) is guided by customer obsession, pace of innovation, devotion to operational excellence, and long-term thinking. By democratizing technology for over two decades and making cloud computing and generative AI accessible to organizations of every size and industry, we turn big ideas into reality. Learn more at AWS.amazon.com and follow @AWSnewsroom.  

Forward-Looking Statements 

Communication contains forward-looking statements that are inherently uncertain and difficult to predict, including statements regarding:  

• planned investments  

• anticipated business activities  

• The effect of conditions and expected benefits of the strategic partnership between Amazon and OpenAI, such as expected development of the stressful runtime environment, stressful development developer environment, and other developments  

• performance scaling capabilities  

• customization offerings or effects that will result from the partnership  

• the applications, utilization, or effectiveness of artificial intelligence technologies  

• accessibility of new features, models, and tools  

• expected compute capacities, capabilities, performance characteristics, and delivery period of Trainium chips  

We use words such as “will”, “believe”, “expects”, “intends”, “anticipates”, “plans”, “future”, “potential”, “continue”, “look forward”, “opportunity”, “keep”, “design”, and similar expressions, as well as words referring to future outcomes such as “train”, “accelerate”, “achieve”, “coming”, “deliver”, “power”, “bring”, “collaborate”, “develop”, “bring”, “become”, “build”, “enable”, “deploy”, “provide”, “invest”, “use”, “meet”, “expand”, “support”, “run”, “lower”, “improve”, “offer”, “next generation”, and variations of such words to identify forward-looking statements. Actual results and outcomes could differ materially for a variety of reasons, including, among others:  

• Fluctuations in foreign exchange rates and energy prices  

• Changes in world economic conditions  

• Tariff and trade policies  

• Resource and supply volatility, including for memory chips and customer demand  

• Spending inflation, interest rates, regional labor market limitations, and world events  

• The rate of growth of cloud services and new and emerging technologies  

• The amount that Amazon invests in new business opportunities and the timing of those investments  

• The mix of products and services sold to customers  

• Competition  

• Management of growth  

• International growth and expansion  

• The outcomes of claims, litigation, government investigations, and other proceedings  

• Data center optimization  

• Variability in demand  

• The degree to which we enter into and maintain and develop commercial agreements and proposed and completed acquisitions, as well as strategic transactions  

• Other risks and uncertainties, including, among others, risks related to new products, services, and technologies  

• Security incidents  

• System interruptions  

• Government regulation, decision, and fraud  

In addition, global, economic, and international conditions, as well as additional or unexpected circumstances, developments, or events, may give rise to or amplify many of these risks. More information about factors that could affect Amazon’s future business, product development, and financial results is available in Amazon’s filings with the Securities and Exchange Commission, including the most recent annual report on Form 10-K and subsequent filings.  

OpenAI and Amazon share the belief that AI should appear in ways that are practical and genuinely useful to people. Combining OpenAI’s intelligence with Amazon’s infrastructure and worldwide reach helps us put powerful AI into the hands of businesses and users at a real scale.  

Sam Altman, co-founder and CEO of OpenAI.

Source: OpenAI and Amazon announce strategic partnership 

OpenAI and Amazon announce strategic partnership

We are excited to share ChatGPT Atlas, a new web browser designed around ChatGPT.  

It gives us a rare moment to think about what it means to use the web. Last year, we added search in ChatGPT so you could instantly find timely information from across the internet, and it quickly became one of our most used features. Your browser is where all of your work, tools, and context come together. A browser built with ChatGPT takes us nearer to a true super-assistant that understands your world and helps you achieve your goals.  

With Atlas, ChatGPT is always with you as you browse, ready to help right where you are. It understands what you are trying to do and can complete tasks for you all without needing to copy, paste, or switch pages. Your ChatGPT memory is built in, so it can utilize past chats and details to help you accomplish new tasks.  

During lectures, I like using practice questions and concrete examples to really understand the material. I used to switch between my slides and ChatGPT, taking screenshots to ask a question. Now ChatGPT instantly understands what I’m looking at, helping me improve my knowledge checks as I go—Yogya Kalra, college student and early tester of ChatGPT Atlas.  

As you use Atlas, ChatGPT learns and becomes more helpful. Browser memories let ChatGPT remember information from the sites you visit and bring it back when you need it. For example, you can ask, “Find all the job postings I was looking for last week and create a summary regarding industry trends so I can prepare for interviews.” Browser memories in Atlas are optional, and you are always in control. You can view or archive them in settings, and deleting your browsing history also deletes any related browser memories.  

LGBT can also help you get things done in Atlas using agent mode. This mode is now faster and more useful because it works with your browsing content. It’s better at researching, analyzing, automating tasks, and planning events or booking appointments while you browse. Agent mode in Atlas is available now in preview for Plus Pro and Business users.  

ChatGPT Atlas is now available worldwide on macOS for free, as well as for Pro and Go users. Atlas is also in beta for business users and, if their plan administrator allows, for enterprise and Edu users. Versions for Windows, iOS, and Android are coming soon.  

You can download Atlas at chatgpt.com/atlas. Getting started is simple. Open Atlas, sign in to ChatGPT, and import your bookmarks, savedpasswords, and browsing history from your current browser.  

Works In Your Flow 

In Atlas, the new Tag page is your starting point. You can ask a question or enter a URL to get quick, helpful results all in one place. If you want more specific results, you can choose tags for search, links, images, videos, and news (where available).  

Built Around You 

ChatGPT can remember what you’ve looked at and suggest your next steps. This might mean:  

• going back to previous pages  

• exploring a topic for the finding  

• related ideas  

• automating regular tasks  

More Capability, More Control 

You control what ChatGPT can see and remember as you browse. You can clear specific pages, clear your entire browsing history, or use an incognito window to log out of ChatGPT temporarily. ChatGPT will remember important details from the content you browse to improve chat responses and offer smarter suggestions, like creating a to-do list from your recent activity or continuing to research holiday gifts based on products you’ve viewed.  

Browser memories are private to your ChatGPT account, and you can control them. You can see all your memories in settings, archive the ones you don’t need, and delete them by clearing your browsing history. Even with browser history enabled, you can choose which sites ChatGPT can access. If the toggle in the address bar is off, ChatGPT can’t see the page content and won’t create any memories from it.  

By default, we don’t use your browsing content to train our models. If you want to include this content, you can turn on “Include web browsing” in your data controls. Even if you opt in, web pages that block GPTBot won’t be used for training. If you enable training for chats in your ChatGPT account, it will also apply to chats in Atlas. This includes website content you attach to the Ask ChatGPT sidebar and browser memories that help inform your chats.  

Parental Controls work in Atlas 2. If a parent has set up Parental Controls for ChatGPT, these settings will carry over to conversations with ChatGPT in Atlas. We are also launching new Parental Controls in Atlas, including the option for parents to turn off Browser Memories and Agent Mode.  

Get Work Done for You 

At Atlas, you can now ask ChatGPT to take action and complete tasks for you directly in your browser.  

Earlier this year, we introduced the ChatGPT agent, and now it works faster and is built right into Atlas. You can give the recipe to ChatGPT and ask it to find a grocery store, add all the ingredients to a cart, and order them to your house. At work, you can ask ChatGPT to open and review past team documents, conduct new competitive research, and compile findings for a team brief.  

When you ask a question, ChatGPT might check if you want it to open tabs and click in your browser to finish the task. You can also start agent mode by clicking the button.  

Starting today, Agent mode in Atlas is available in preview for Plus, Pro, and Business users. It’s still early, and we may make mistakes on complex tasks, but we are working quickly to improve reliability, speed, and success with more complex workflows.  

We prioritized safety as we built ChatGPT’s agent capabilities in Atlas and added safeguards to address new risks that can come from access to logged-in sites and browsing history while taking actions on your behalf. For example:  

• It cannot run code in the browser, download files, or install extensions.  

• It cannot access other apps on your computer or file system.  

• Will pause to ensure you are watching it take action on specific sensitive sites, such as financial institutions.  

• You can use an agent in lockdown mode to limit its access to sensitive data and the risk of it taking actions on your behalf on websites.  

ChatGPT’s agent features still have some risks. Besides making mistakes while acting for you, agents can be tricked by hidden malicious instructions in places like web pages or emails. This could overwrite the agent’s intended behavior and might lead to stealing data from sites you’re logged into or taking actions you didn’t want.  

As outlined in the ChatGPT agent system card, they’ve run thousands of hours of focused red-teaming and have placed particular emphasis on safeguarding ChatGPT from such attacks, including designing safeguards that can be rapidly adapted to novel attacks. Still, our safeguards will not stop every attack that appears as AI agents grow in popularity. Users ought to weigh the trade-offs when deciding what information to provide to the agent and take steps to minimize their exposure to these risks, such as using the ChatGPT agent in locked-out mode in Atlas. We will continue to monitor and patch any vulnerabilities that we discover.  

This launch is a step toward a future where most web use occurs through agent systems. You will be able to hand off regular tasks and focus on what matters most.  

What’s Next? 

We continue to make Atlas better, and our roadmap includes:  

• multi-profile support  

• improved developer tools  

• ways for app SDK developers to increase the discoverability of their apps on the Atlas website  

• Owners can now add ARIA tags to improve how that ChatGPT agent works on their websites in Atlas  

This is only the beginning. We’ll be adding new features and improvements often. You can keep up with the updates in our release notes. Try it now at chatgbt.com/atlas.

Source: Introducing ChatGPT Atlas