What Sets Successful Professionals Apart in Their Daily Tech Habits

A recent survey of more than 1,000 professionals, conducted by LiquidWeb, has drawn a detailed picture of how people at the top of their careers handle technology in daily life. By comparing high earners with the broader American workforce, the study found that patterns of use were less about spending less time on screens and more about shaping routines around them.

High achievers in the survey reported spending roughly seven hours a day on their computers, mainly for work, and about three hours on their phones, often for leisure. The wider workforce, in contrast, leaned more heavily on phones, devoting nearly a quarter more time to them. The gap suggests that for those with demanding jobs, computers remain the primary tool while casual mobile use takes a back seat.

Time away from screens was another part of the picture. About 44 percent of successful respondents said they stepped away from devices at least once a day, compared with 38 percent of the wider group. Smaller shares in both groups reported breaks only once a week. Focus modes and screen-time alerts were mentioned, yet the most common approach was as simple as switching off altogether.

Morning routines also showed a split. Many of the high achievers held off from checking devices straight after waking, giving themselves space to plan the day before facing messages and notifications. More than seven in ten scheduled deep work sessions into their calendars, often lasting an hour to an hour and a half. Those who did so finished projects at roughly double the pace of peers who worked in shorter bursts.

The way communication was handled also set this group apart. Nearly half said they checked work messages outside of office hours, compared with 38 percent of the wider sample. At the same time, many batched emails and Slack notifications into fixed windows. This approach cut down mistakes by about a quarter and left more attention for bigger tasks.

Social media habits told their own story. Around 49 percent of high achievers reported avoiding TikTok completely. Platforms like Reddit and LinkedIn drew more interest, with four in ten using Reddit regularly and LinkedIn showing a higher adoption rate than in the broader workforce.

When it came to devices, premium tools were favored. Sixty-three percent of successful respondents called the iPhone essential, compared with 34 percent for Android phones. Nearly half relied on laptops rather than desktops, and accessories such as headphones, smartwatches, and tablets were common, though at lower levels. More than a third said they preferred premium gear overall, while just under a quarter leaned toward cheaper models.

The survey also looked at health-related routines. Short breaks for stretching or mindfulness were tied to about a 10 percent lift in reported productivity. Some respondents used apps or timers to remind themselves to pause, while others followed methods like the Pomodoro technique to manage long stretches of work without burning out.

Taken together, the findings point to a group that spends plenty of time with technology yet keeps it in check with clear routines and selective habits. The picture that emerges is one of people who treat their devices as tools, drawing firm lines around when and how they’re used, and finding ways to keep focus and energy steady across the day.

Read next:

• Why Executive Branding Is Getting Personal: Lessons from the Digital Workplace’s New Playbook

• OpenAI Expands ChatGPT With Chromium Browser, Affordable Plan, Voice Upgrades, and Personality Shift

• The Less You Use It, The More You Fear It: AI’s Uneven Welcome at Work

by Irfan Ahmad via Digital Information World

OpenAI Expands ChatGPT With Chromium Browser, Affordable Plan, Voice Upgrades, and Personality Shift

OpenAI is developing a new browser based on Chromium. Early signs, as spotted by Tibor Blaho, suggest the company may first release it for macOS. The project is expected to rely on ChatGPT at its core, allowing the software to guide how users move through the web.

Unlike traditional browsers, the design points to features such as automatic tab selection, a custom new tab page, and browsing carried out directly by ChatGPT. These ideas build on the existing Agent mode inside ChatGPT. That mode already runs in a cloud setup on Linux, where it can perform requests such as creating documents or presentations using online material.

At present, Agent mode only interacts with a remote virtual browser. It processes screenshots, clicks links, and fills out forms in that environment. Recent details suggest OpenAI is preparing a second route. The system may soon switch between the older cloud option and a new local browser built by the company itself. References to a “Use cloud browser” toggle and traces pointing to macOS confirm the direction of development. Earlier reports also indicated that the firm wants to keep more user activity within a ChatGPT-style interface rather than pushing people to open websites one by one.

Cheaper Subscription Plan in Testing

Work is also underway on pricing. OpenAI is testing a plan called ChatGPT Go, positioned as a lower-cost tier. The dashboard for some accounts now shows a Try Go button, linked to a payment request of four US dollars. The same option appears in euros and pounds, which makes clear that the company is preparing a broader release.

Image: @btibor91 / X

When first spotted, the cheaper plan looked tied to India. The new currency listings point instead to global availability. If the service remains at four dollars or its equivalent, it could draw interest from users who find the higher tiers out of reach.

• Also read: Study Finds Open-Source AI Models Consume Far More Computing Power Than Closed Systems

New Voice Controls and Models

Another area of focus is voice mode. Hidden settings now allow users to control the speed of speech, with sliders ranging from half pace to double pace. The adjustment gives more freedom in how conversations sound, whether slower for clarity or faster for quick updates.

There is also a custom instructions prefix for voice interactions. This lets the system remember a user’s preferred approach, so instructions do not have to be repeated at every turn.

Alongside these changes, the model selector has been updated. Subscribers can now see GPT-5 options labeled high, fast, and auto. Access to GPT-4o has also been restored for paying accounts.

Subtle Personality Shift in GPT-5

OpenAI has also started adjusting how GPT-5 responds in conversations. The company says it has made the model warmer and more approachable after receiving feedback that earlier versions felt too formal. The difference is intended to be subtle, with responses sometimes starting with phrases such as “Good question” or “Great start.” Internal testing showed no increase in flattery compared with the previous GPT-5 personality. The change is rolling out gradually and may take up to a day to reach all users.

The update has prompted mixed reactions online. Some users have asked the company to keep GPT-4o permanently available, saying they prefer its colder and more restrained style. Others expressed concern that ChatGPT should focus on delivering clear information without adding unnecessary warmth.

Notes: This post was edited/created using GenAI tools.

Read next: YouTube Stands Out as America’s Favorite Social Platform

by Irfan Ahmad via Digital Information World

Study Finds Open-Source AI Models Consume Far More Computing Power Than Closed Systems

A recent study shows open-source artificial intelligence models use far more computing power than closed systems on the same tasks. The findings raise questions about whether these models deliver the cost benefits often assumed.

The study, led by Nous Research with help from independent analysts, compared nineteen reasoning models. They were tested on knowledge questions, mathematical problems, and logic puzzles. Researchers focused on token efficiency, the measure of how many units of computation a model spends to produce an answer.

Results showed open-weight models burning between one and a half to four times as many tokens as closed ones. For basic knowledge tasks, the gap widened sharply, with some systems using ten times more tokens. In some cases, the added computation translated into higher overall costs despite lower per-token pricing.

Large Reasoning Models showed the biggest inefficiencies. Designed to think step by step, they often produced long reasoning trails on problems that required little thought. The team found some models consuming hundreds of tokens just to answer a straightforward question such as the capital of Australia.

Performance differed across providers. OpenAI’s o4-mini and the gpt-oss models ranked among the most efficient, particularly in mathematics. They used up to three times fewer tokens than other commercial systems. Nvidia’s llama-3.3-nemotron-super-49b-v1 topped the open-source list, while some of Mistral’s recent releases were marked as outliers for heavy token use.

In logic problems, the difference was narrower but still present. Adjusted versions of puzzles like Monty Hall revealed that models leaned on training memory until changes forced them to reason properly. When problems were altered, token use rose.

The research team relied on completion tokens to measure efficiency because many closed providers hide their raw reasoning steps. Some compress internal thinking into short summaries, while others use smaller models to record reasoning in compact form. Completion tokens, billed directly to customers, gave a clearer picture of the actual computing effort.

Testing also included math competition problems with altered variables to prevent memorized answers. Results showed most models attempted to solve problems rather than recall them. Still, the spread of token use was wide. OpenAI’s efficient models stood out for keeping costs lower overall, even with higher per-token rates.

The study suggests enterprises need to look beyond accuracy and per-token pricing when choosing systems. Computing overhead can mount quickly and shift the balance of total costs. Closed providers appear to be reducing token use with each update. Some open systems, on the other hand, have been moving in the opposite direction by generating longer reasoning trails.

The complete dataset and evaluation code have been made public. Researchers believe future development should target efficiency alongside accuracy. Shorter, denser reasoning could help keep costs down and preserve performance on complex problems.

The findings point to efficiency as a central factor in determining which models can be scaled for widespread enterprise use.

Notes: This post was edited/created using GenAI tools.

Read next:

• Sam Altman Opens Up on Google, GPT-5, and OpenAI’s Next Moves

• ChatGPT Mobile Surpasses $2 Billion Spending And Becomes Fastest App To Hit 1 Billion Downloads
by Asim BN via Digital Information World

Google Search Still a Reliable Source, But Social Keeps Sliding

Chartbeat’s long-term review of traffic to more than 560 US and UK publishers suggests that Google remains a steady hand for news websites. The share of visits coming through search has hardly moved in six years. It was about 15% in early 2019, jumped above 20% during the first Covid months, and has hovered around 19% this summer. Google’s dominance in that slice of traffic has barely wavered either, holding between 95% and 97% the whole time.

One detail matters here: the count includes Discover, which has taken over as a key route for people reaching publishers. Much of the “search stability” shown in the numbers reflects how Discover feeds have grown into a traffic driver alongside traditional search clicks.

Direct Visits Lose Ground

The picture looks different for direct visits, the readers who go straight to a homepage or use bookmarks. That traffic made up more than 16% of visits at the start of 2019, fell during the pandemic, clawed back for a while, and has since slid again. In July, the share was just over 11%. Despite years of talk about building loyal “front-door” audiences, many publishers are still struggling to turn that into reality.

Social Referrals in a Long Drop

The sharpest declines have come from social networks. Their share of referrals has slipped from more than 17% in 2019 to 13% today. Facebook, which once drove close to a billion monthly visits, has lost half of that traffic. Its highs came during the pandemic, followed by a steep drop as it shifted to personal posts over publisher content. Even with a small recovery this year, it is still well below its former levels.

X, the platform formerly known as Twitter, has shrunk even more. Monthly referrals have dropped from around 129 million to barely 32 million. Instagram has moved up a little, while Reddit has surged ahead, reaching about 19 million referrals in July. The jump for Reddit coincided with a visibility boost in Google results after its content deal last year, which also gave the site a new role in training AI systems.

External Links and On-Site Journeys

Traffic patterns from other sources show mixed movement. Visits coming from links placed on outside sites, such as blogs or other publishers, have doubled since 2019. Internal clicks — readers moving from one page to another within the same outlet — have edged down from 43% to 39%. Holding audience attention once they arrive has become a bigger focus as a result.

Aggregators: Google News Still Leads, But Lower

Aggregator traffic tells another story. Google News remains the biggest, though its referrals dropped in late 2023 and have not climbed back. In July it drove around 107 million visits, which is roughly the same as six years ago but nearly a third lower than in 2023. SmartNews, Newsbreak and Flipboard all saw growth during the pandemic but have since fallen away, with Flipboard down the most.

H/T: Pressgazette

Notes: This post was edited/created using GenAI tools. Image: DIW-Aigen.

Read next: Sam Altman revealed he no longer uses Google Search, despite OpenAI still relying heavily on Google’s cloud infrastructure.
by Web Desk via Digital Information World

ChatGPT Mobile Surpasses $2 Billion Spending And Becomes Fastest App To Hit 1 Billion Downloads

ChatGPT’s mobile app has reached more than $2 billion in global consumer spending since its release in May 2023, according to new data from analytics firm Appfigures. The report highlights how far ahead the app is in mobile markets compared with its closest competitors.

This year alone, the app has taken in $1.35 billion from January through July, up sharply from the $174 million recorded during the same stretch in 2024. Monthly revenue now averages about $193 million, compared with $25 million last year.

The performance gap is clear when stacked against other chatbots. Grok, the nearest competitor on mobile, has brought in about $25.6 million so far in 2025. That puts its results at a fraction of ChatGPT’s. Claude and Copilot remain further behind. Part of the difference comes from timing. Grok only released stand-alone apps this year, with its iOS version arriving in January and its Android version in March.

Download figures underline the same trend. ChatGPT has been installed 318 million times in 2025, nearly three times higher than during the same period last year. The app’s lifetime total now stands near 690 million installs, with current monthly downloads at roughly 45 million. Grok, by comparison, has reached fewer than 40 million installs overall.

Spending patterns vary by region. The United States is responsible for almost 40 percent of ChatGPT’s lifetime revenue, where the app generates about $10 per download. Germany is second at just over 5 percent. In terms of installs, India leads with nearly 14 percent of the global total, followed by the U.S. with about 10 percent.

On a per-download basis, ChatGPT averages $2.91 worldwide. Claude follows with $2.55, Grok at $0.75, and Copilot at $0.27. These figures point to different user behaviors across platforms but also show how strongly ChatGPT has converted installs into paying customers.

ChatGPT Continues to Break Download Records

The second quarter of 2025 added more milestones. As per Sensortower data, ChatGPT became the fastest app to reach one billion global downloads, while also setting a record as the fastest non-preinstalled app to reach 500 million monthly active users. Its growth was matched by strong monetization, with the app ranking fourth worldwide by in-app purchase revenue during the quarter.

Even with more competition in generative AI on mobile, including Meta’s new AI assistant and xAI’s Grok, ChatGPT still accounted for more than two-thirds of total in-app purchase revenue in this category.

Market comparisons underline its scale. TikTok remained the top-grossing mobile app in Q2 2025, bringing in nearly $1.7 billion, which was more than double the revenue of any other single app. ByteDance’s CapCut also placed among the top 10 by revenue.

In downloads, however, ChatGPT pulled ahead. The app passed TikTok to take the number one spot globally in Q2 2025, helped by its first major digital ad campaign, which promoted free ChatGPT Plus access for college students. Meta also continued to secure high positions, with three of its apps ranked between third and fifth, while Threads held tenth place.

While the mobile figures don’t reflect every revenue stream available to AI firms, since web subscriptions and API usage are also significant, they do show how firmly ChatGPT has established itself at the center of mobile adoption.

Notes: This post was edited/created using GenAI tools

Read next: 

• Meta Scientist Highlights Core Principles for Safer AI

• How MrBeast Builds Viral Videos
by Irfan Ahmad via Digital Information World

How MrBeast Builds Viral Videos

Jimmy Donaldson, known online as MrBeast, oversees a content operation that produces high-cost, high-visibility videos for a global audience. His channel  boasts more than 420 million subscribers.

Image: MrBeast / YT

Former employees and an internal production guide describe a process that mixes fast decision making with heavy planning, and where a single judgment by Donaldson can end a project. That judgment is often summed up by one word: cringe.

From idea to green light

As reported by Insider, ideas begin in a central ideation team that tracks trends, tests titles, and drafts thumbnails. The team records large numbers of concepts and narrows them through internal review. Only a small share reach Donaldson and his creative leads for final consideration. The company treats concept work as a pipeline that feeds production teams with ideas that can scale.

When an idea moves forward, producers assess cost and feasibility. The company added a dedicated feasibility group as it expanded. Ambitious proposals can be postponed or shelved if they are impractical, but staff report that Donaldson resists quick refusals. He expects teams to pursue alternatives before stopping a concept.

A production hub in Greenville

Beast Industries has built a base of set builders, technical crews, and a rotating pool of assistants. Early on, temporary helpers were called friends of friends. The company now relies on a larger task force to handle on-set work and logistics.

Large challenge videos require extensive camera coverage. For competitions with many participants, the crew deploys numerous cameras so no moment is missed. Long-running challenges may have cameras recording around the clock. Most logistical work is completed before on-camera talent arrive, leaving final decisions to Donaldson and creative leads.

Producers as generalists

Producers at MrBeast perform a broad set of tasks. They coordinate contractors, schedule crews, manage budgets, and troubleshoot on site. The work differs from traditional film crews, where roles are often specialized. Staff describe producers as problem solvers who must be ready to step into multiple tasks as needs arise.

Editing with attention to retention

Editing is organized around keeping viewers engaged. Projects that span days or weeks generate hundreds of hours of footage. Teams of editors handle daily trims, syncing, and assembly. Lead editors pull the final cut together. For longer shoots, editors often build rough cuts during production so there is a working edit by the time filming ends.

The company uses detailed viewer metrics to shape edits. Editors monitor the exact second audiences stop watching and adjust pacing and highlights to reduce drop-off. That approach makes post-production central to the final story.

Release strategy and follow-up

Final videos are formatted for multiple platforms and languages. Thumbnails undergo A/B testing and the team prepares short-form clips for vertical platforms. Some content is altered for younger viewers. After release, staff analyze retention and engagement data to evaluate performance and inform future planning.

A repeatable system

The operation combines methodical idea work, thorough feasibility checks, heavy on-set coverage, and analytics-driven editing. Former staff say the process lets the channel test more ideas and learn quickly from failures. That scale of resources and the emphasis on measurable engagement create an advantage that many creators try to emulate.

Notes: This post was edited/created using GenAI tools.

Read next: Meta Scientist Highlights Core Principles for Safer AI

by Asim BN via Digital Information World

Google Gemini Will Soon Use Your Uploads for AI Training, Full Details and How to Opt Out

Google is preparing a privacy change for its Gemini AI service that will allow it to use some of the files, photos, and videos you share to help train its artificial intelligence systems and improve other products. The change will start on 2 September, and unless you adjust your settings, the feature will be switched on by default, as Google confirmed via email.

In the coming weeks, Google will rename its Gemini Apps Activity setting to Keep Activity. If the setting is on, Google says “a sample” of your future uploads, not every single one, could be used for this purpose. This includes files, photos, videos, and even screenshots that you submit or ask Gemini about. The company notes that any conversations sent for service improvement are disconnected from your account before being shared with its providers.

What This Means for Your Data

When active, Gemini Apps Activity records the prompts, uploads, and interactions you have with the chatbot. Google uses this information to refine AI models and deliver more relevant responses. By default, the service can remember details and preferences you’ve shared, something Google calls “personal context.” This context is already in Gemini’s 2.5 Pro model and will soon arrive in the default 2.5 Flash model.

• Also read: Google’s Gemini Tops List of AI Chatbots Collecting the Most User Data

For example, if you’ve asked about Japanese culture and YouTube ideas before, Gemini could suggest content topics in line with those interests. This memory feature is switched on by default.

Temporary Chats and Private Conversations

If you’d rather keep certain discussions separate, you can switch to Temporary Chats. These work like an incognito mode for Gemini. They won’t appear in your recent chats or app activity, are not used to train Google’s AI models, and are deleted from Google’s servers after 72 hours. This applies to both the conversation and any related uploads.

Managing Audio, Video, and Screen Recordings

Gemini Live sessions, where you interact with video, audio, or screen-sharing, can also be stored in your account history if Gemini Apps Activity is on. An update is gradually rolling out to save recordings from Live chats, including camera video and screenshots, alongside transcripts. Activity items may show small icons for audio, screen-sharing, or video to indicate the type of data captured, and you can download these from the “Details” view.

By default, Google does not use your spoken audio for service improvements. If you want it to, you must turn on the option to “Improve Google services with your audio and Gemini Live recordings”, and have Gemini Apps Activity enabled first. Some of these clips may also be reviewed by humans.

How to Turn Off Data Collection

To stop Google from using your uploads for AI training:

1. On your computer, go to gemini.google.com and sign in.
2. Open Settings & help and choose Activity. Or visit this page.
3. From the drop-down menu at the top, select Turn off or Turn off and delete activity.

Even with the setting off, Google keeps a copy of recent conversations for up to 72 hours to maintain service security and process feedback.

Deleting Past Activity

You can delete activity from myactivity.google.com/product/gemini or via the Gemini site:

• Delete all time, last hour, last day, or a custom date range.
• Remove a specific activity item or all activity from a particular day.

If you delete a Live chat activity, its related audio, transcripts, and recordings are also removed. Deleting a conversation with a Gem does not delete the Gem itself.

Auto-Delete Timelines

Gemini Apps Activity is set to auto-delete anything older than 18 months by default. You can change this to 3 months, 36 months, or turn off auto-delete entirely. Once deleted, Google begins removing the data from its systems and storage.

Why the Change Matters

Google says these changes are about improving AI capabilities and personalising services, but the default-on approach means uploads could be sampled without you actively agreeing each time. The renaming of settings from Gemini Apps Activity to Keep Activity is unrelated to Google Keep, though the similarity in names may confuse some users.

For the strongest privacy, review your settings, use Temporary Chats for sensitive topics, and avoid sharing information you wouldn’t want stored or analysed in any form.

Notes: This post was edited/created using GenAI tools.

Read next: Why Executive Branding Is Getting Personal: Lessons from the Digital Workplace’s New Playbook

by Irfan Ahmad via Digital Information World