(Featured) - Google has made Gemini 4 Argon broadly available at an introductory price intended to challenge OpenAI and Anthropic. Professionals share what their peers should do now.

Edited by Brett Farmiloe, Editor at Featured News · Roundup by Featured · Oct 19, 2026

Google made Gemini 4 Argon, its most capable AI model, broadly available on Oct. 13 to developers, businesses and Gemini app users on Google’s Pro and Ultra plans. It’s priced to challenge OpenAI and Anthropic, with an introductory rate that doubles on Dec. 1. In one demo, Argon controlled a Boston Dynamics robot dog as it stepped on a pedal bin and dropped in soda cans.

DuBose ColeCo-founder, New Classic · New York, NY

As with every launch, Google's new Gemini 4 model will contribute to the disruption and acceleration we're seeing in marketing from AI. However, there is one immediate impact brands should consider in how users interact with AI and discover brands and products through search.

For brands, Gemini 4 looks to increase AI fragementaiton in a new way. The company has announced that 'AI Mode' in Google search will vary the model it uses based on the user's subscription. Pro and Ultra subscribers will get the new Argon model, while free users are capped at lighter, less sophisticated options. While AI overviews in the main search interface won't change, those users who go to AI mode will see a different experience, and receive different thinking based on their subscription status with Google.

For brands already trying to AI optimise, this opens up a new dimension to consider. While AI platforms normally cap usage of a model vs. the model access itself based on user status, Google's move pushes a model split to a search adjacent interface. Brands need to audit their presence to optimise not just against a range of providers, but increasingly, a range of models offered by the provider.

Randall SmalleyExecutive - Global Marketing / Business Development, Cruise America · Phoenix, AZ

Earlier this year, we launched RV Trip AI, which sits front and center on our website. Our vision is to make planning an RV vacation as exciting and simple as taking one. Customers can tell us where they want to go, what they love to do, and what kind of experiences they're looking for, and AI helps bring that journey to life.

One of the biggest opportunities for us is helping customers understand where they'll stay along the way. For many families considering their first RV vacation, the uncertainty around planning a route, finding campgrounds, and knowing where to spend each night can be a barrier to booking. We believe removing that uncertainty is critical to converting more travelers into RV customers.

As AI continues to evolve, particularly with agentic ecommerce, we see a future where customers can plan their entire bucket-list vacation in one experience, from selecting their RV and building their itinerary to reserving campsites along the route. Our goal is to turn the dream of an RV road trip into a trip families can confidently book. We believe making the entire journey easier to plan will introduce more families to RV travel and ultimately drive greater conversion for Cruise America.

Jillian O'MaliorFounder & Principal Strategist, New World Labs · Walnut Creek, CA

I've done an investigation into a recruiting firm's AI sourcing platform that published its internal notes on 110 public job postings. One of them flagged candidates for exclusion because they were laid off or unemployed. It's a clear example of what lazily integrated AI looks like in the hiring process: private info made public, embarrassingly so.

I've been on the candidate side too: 743 applications over 16 months, almost all through automated filters and rejections, and now upwards of 90 robocalls and spam texts a week because my data was resold through the apply process.

A more capable model at an introductory price means a lot more of this, especially in HR/TA. So before HR buys anything, take note of what parts of hiring a human needs to own: not lead, OWN. For me, that's the first real conversation and the rejection. If someone gave you hours of their time, a person should tell them why it didn't work out.

Jon SchneiderFounder, Recruiterie · Chandler, AZ

A sales leader’s best evidence is often on video. The CROs and VPs of Sales Recruiterie finds for venture-backed companies have usually done podcasts and conference panels, and that’s where you hear how they actually built a team and handled a missed quarter. Until now, nobody running a search had time to watch all of it.

Gemini 4 Argon takes video and audio directly. The newest flagship models from OpenAI and Anthropic take only text and images. Google also built Argon to pull specific details out of long videos, so I can hand it a finalist’s podcast interviews and ask what they’ve said about ramping new reps.

My advice to other recruiters is to use it to research candidates and leave the judging to people. Don’t let a model rate anyone on how they look or sound, and get permission before you run your own interview recordings through it.

David AlanPresident, Closets by David Alan · Pittsburgh

With Gemini 4 Argon we can begin the revision process by inputting pictures of your home and what your life style looks like. Then showing them what there closet can look like at the first meeting.

There should be many designers that sell custom closet's playing around with these programs to help with customers personal needs.

As shown above there are many ways we can incorporate this into our current process. When choosing high end finish and lifestyle of the customer.

Denise GredlerFounder, BestCompaniesAZ · Gilbert, AZ

After 25 years of working with research and media partners on awards and recognition programs, I've seen firsthand how important clear evaluation criteria, credible information and independent third-party judging are to the integrity of an awards program.

New AI technologies, including Google's Gemini 4 Argon, have the potential to help awards programs research company claims, review publicly available information and identify areas that deserve further evaluation.

But in my opinion, AI should support and enhance the judging process, not replace it. The credibility of any awards program depends on established criteria, objective evaluation, reliable data and independent judging.

The real opportunity is to use AI to make the evaluation process more efficient and thorough while maintaining the integrity and human oversight that make these awards meaningful.

This story was produced by Featured and republished with permission. Experts featured here are members of Connectively, Featured's expert network.

(Featured) - MarketScreener reported that Meta released Muse Wonder AI models to challenge Anthropic and OpenAI. Leaders deploying frontier models in production explain one decision they made or reconsidered as competition among vendors expanded, along with a concrete result from that choice.

Edited by Adrian James, Editor at Featured News · Roundup by Featured · Oct 15, 2026

According to MarketScreener, Meta’s Muse Wonder release adds another contender for teams evaluating AI models.

We asked leaders running these models in production to share specific decisions they have made or reconsidered because of multi-vendor competition, and the results that followed.

Kiyoshi CaseyFounder, 321ai · Melbourne, Florida, USA

One week of internal review was what it took for us to make the switch based on budget analysis of which AI model to use in production. In the next week we burned 1.55 billion tokens at a 96% cache read rate. Priced against market list rates we spent $62 total using a budget tier model that utilized off-peak rates versus $810 for a flagship model's pricing. That was a 13 fold difference and our workflows adapted to utilizing the frontier model to review the work, this ended up costing $44 for the frontier model to review all of the work with a minimal dropoff in output quality compared directly against the prior week's output. We internally built the infrastructure to support the orchestrator frontier model reviewing the budget model's output and now utilize this framework for all of our development.

Muninder AdavelliDev / Marketer, Proxy.Report · Bitola, North Macedonia

I have been working with websites: building them, ranking them, converting them to lead generators for the last 10 years. No wonder I get more than 100 link-building requests a week.

Some ask if I need link-building services, and some ask for guest posts on websites I manage.

1. The first layer sorts them into link sellers and link buyers.

2. The second layer classifies link buyers as high-value or low-value prospects.

3. The third layer is finding the right template.

The fallback in the system is the manual layer, for when the LLM cannot decide.

Before, when I first built this application with the help of ChatGPT, I used the OpenAI API, and it used to cost me around $0.012–$0.019 per email, and it was vendor-locked. Later I reconfigured the app to use OpenRouter. Now with Qwen3-30B-A3B-Instruct-2507, it’s down to $0.0006–$0.0009 per email. That’s 20x cheaper.

The only drawback with Qwen is that it sent 3% more emails to the manual layer. It’s okay for me.

My take: not all tasks require new advanced models. Knowing what’s good enough can save big time on expensive LLM spending.

Vadzim TashlikovichHead of Software Department | CTO, SCAND · Warsaw, Poland

In the latest projects we often keep an eye on the model life cycle for specific providers. In fields where accuracy is very important, it is expensive to recalibrate all workflows and perform acceptance tests once again when a new version of the LLM vendor’s model appears. The older one gets deprecated, and prices become significantly higher.

We developed an innovative live speech translation pipeline for one of our customers. It captures audio from English speakers and creates a real-time audio translation into 5 languages (with great flexibility to add more). For this pipeline several ultrafast models were needed. We chose Groq as a provider that helps to maintain 500+ tokens per second for open-source models.

Initially, several models participated in the pipeline: GPT-OSS-120B and Llama 3.3 70B. The latter was especially great at translating small chunks of text into different languages, but its short life cycle (6 months) made it impossible to use. That’s why the whole pipeline now uses the industry-proven GPT-OSS-120B, which has a much longer life cycle. We stopped searching for new, faster and better models, because testing and recalibrating the pipeline is too costly: it requires 8–20 hours and the involvement of several native speakers of the supported output languages. Instead, we improved the algorithm itself to reach better results.

Manoj SethiFounder, Boffin Coders · Mohali, Punjab, India

In mid-2026 IELTS stopped offering the test on paper. IELTS Builder Canada, who already had their mock tests published as books on Amazon, reached us through our website. They wanted a platform where their students could take mocks with a real exam experience.

Reading and Listening were simple, as we mapped the correct answers and marked them after the exam. Writing and Speaking have no fixed answers, so feedback was the hard part. AI was already running the Speaking mock interview, so we advised them to extend its role to give immediate feedback and bands based on the IELTS rules.

The AI is not a single model. Different models are routed through OpenRouter, and we choose the model by the complexity of the task, to control token consumption and keep the budget in check. As newer and better models came, we shifted to them for better evaluation, without being locked to one vendor. Speech-to-text stays on the school's server and only text goes out, because the students are visa applicants and some are minors.

As per the data on 23 September 2026, there were 4,584 module submissions in 30 days, with no student waiting for a teacher's feedback.

James Hoad CoombeFounder, PolicyPing · Melbourne, Victoria, Australia

My product, PolicyPing (and side product ReefReport) monitors government building rules for changes and notifies customers which sites are affected, using frontier models to review the amendments. When planning, I routed the discovery and strategic AI work through Openrouter so I could test all models, and run them side by side with the same prompts to compare results. I did not have to commit to a single vendor, and it gave me comparison data. It also allowed me to amalgamate results where required if some performed markedly better, or came up with new ideas.

This flexibility was key for testing various LLMs for speed, accuracy but also capability. It clearly showed the frontier models of each provider were somewhat close to each other but the step down models were significantly 'worse' at some things than the top-tier. Fable and Opus clearly outperformed Sonnet (all Anthropic products) but were relatively evenly matched with their competing models at each tier with different focuses and largely different personalities.

The trade-off for OpenRouter as an aggregator is the increased cost and fees, but there's no other straight-forward way to get access to all models on an even interface. Once the model I picked proved itself (which ended up being Opus for a balance of cost/performance), I moved virtually all of my build work to Claude subscription (in lieu of API) methods entirely to leverage all their models and also ClaudeCode and their Cowork systems that clicked for me the best. Since then this has changed but the fact remains the aggregator API use was best for testing but could cost up to 5 times more than a subscription inclusion.

I've built my tech stack to be model agnostic so I can move between providers fairly easily, which is necessary with market competition changing capabilities and costs on a week by week basis at this moment in time.

This story was produced by Featured and republished with permission. Experts featured here are members of Connectively, Featured's expert network.

(Mr Ozan Dagdeviren) - I built an entire operation from scratch in the last year, solo, and none of it would have been possible without AI; the journey made me better at using it and more convinced everyone should learn it.

By Mr Ozan Dagdeviren

Edited by Adrian James, Editor at Featured News · Oct 9, 2026

What I find quite sad is that when a space becomes too noisy, it draws out the signal as well. Not the first or last time this happened. Do you remember what a big deal blockchain was? And NFTs? And the internet!Among the noise, it takes real deep expertise, critical judgement and most importantly being an original thinker to tell whether something is hot air or has any lasting substance to it. Some technologies did and some didn't. Anyone can tell with hindsight, but very few have the foresight to guess right.Just like the fact that desire is mimetic, so are most our own opinions. The positive or negative opinions you hold on AI, to a very large extent are shaped by others, without your explicit awareness. Unless you deliberately practice being an original thinker and sharpen those critical thinking skills - and have the self confidence to be okay with being seen as wrong - you will most likely follow the crowds.You'll be likely to be an AI fan - ignoring the real issues around it, or be an AI sceptic, burrowing your head in the sand around the benefits it brings. This is the bifurcated social discourse with many things. It's easier to pick a side as if you are watching a match, then to have the humility and mindset to be able to appreciate the skills and abilities of your opponent, or to criticise the one you support objectively.Same is true for sports, politics, public discourse around gender and minorities, immigration, religion and now, of course, for AI.I have a background in human sciences - I studied the individual mind through psychology and neuroscience and I studied the social mind through sociology and social psychology. I love the balanced perspective this helps me afford, as I can shift seats when looking at a topic and see it through multiple lenses. As such, that's the approach I'm taking with AI. I love it, I fear it, I use it, I am cautious about it.

Whatever camp you are in, you would benefit from learning AI. And I’m not talking here about learning about AI, the politics and the economics of it. No, I’m talking about becoming AI literate as a baseline, then getting to AI fluent very soon.

There are multiple dimension of AI Fluency, let me guide you along.

Prompting & CommsFirst, you need to be able to communicate with AI well. This is how you get across what you mean and what you want. Talking to an AI is different than talking to a human. As you talk more with it you'll understand how to engage with it better.Critical ThinkingThis is your evaluation and judgement with AI. Also involves your understanding of it's limits.Technical UnderstandingDo you understand AI fundamentals, memory and context, how LLMs work and why they always guess but never actually "know".Workflow ApplicationHave you integrated AI into a way of working, can you break down a task into components, can you apply it in the right domain whether that's arranging your social life or finances?Safety & ResponsibilityCan you use AI safely, with correct privacy judgement. Can you set reliable guardrails? Do you understand when and how it can go wrong?

Clearly, this is a new day for all of us. I like to say it's still day one with this new technology, in the grand scheme of things. (Well maybe now it's day two, but it's still very early days isn't it?)

I'd love to live in a world where more people understand what they are holding strong opinions about.What do you think?

Mr Ozan Dagdeviren

Mr Ozan Dagdeviren is the Founder & CEO at AISA AI Fluency. He also teaches on creative thinking part-time and has a background in behavioral science. He is the author of 5 books, has reached millions of people with his educational content.

This story was produced by Featured and republished with permission. Experts featured here are members of Connectively, Featured's expert network.

(Maurice Sikkink) - You are keeping your photos off social media, that's great. But what can the platforms storing years of our memories learn from them?

By Maurice Sikkink

Edited by Adrian James, Editor at Featured News · Oct 12, 2026

When I got married last year, I didn't want the wedding all over the internet. Apart from my LinkedIn, I'm pretty much invisible online, so it would have felt odd to have one of the most personal days of my life show up on other people's Instagram and Facebook feeds.

With my kids, I feel even more strongly about it. I know that once someone publishes a photo of them online, I have no control where it ends up, who's looking at it and with what intention, or what some future piece of software might do with it.

That's my own point of view, but it also matched with something we'd been noticing.

Working on a photo-sharing service, I’ve noticed that people are a lot more careful with their photos now than they used to be. I don’t have exact numbers for it, but you can tell from the questions they ask before they sign up. How easy is it to share photos with others? How much can I store? And then, more often: where are the photos stored? That one comes up especially from people in Europe. Who can access them, do you scan them, do you use them for AI? People really do dig through our FAQ and terms to find out the answers. To me, that says privacy has moved out of the small print and into the main reasons people choose a product.

My own family photos got me thinking about this from another angle too. Most of the conversation is about whether photos are published on Insta. Although I do know some successful influencers who don't mind and show their entire family online.

But at my children's school, for instance, there's a lot of attention around posting pictures of other people's kids, and at weddings couples now regularly ask guests not to put everything online. Or even just put the phones down. I've been at weddings from which I have no photos at all, only bride and groom had a photographer and didn't show the photos to anyone except one printed album at their house. Fair enough, but keeping photos off Instagram doesn't mean you stop keeping them somewhere and they just become some kind of secret.

You still want the grandparents to see the kids. Me, personally, I still wanted our wedding guests to send us their pictures. They have to live somewhere.

So the question I keep coming back to is: whether the place you store them is looking at them. Not who can see your photos, but whether the service holding them is analyzing their content. Ten or fifteen years of photos says a huge amount about a person: your kids growing up, your partner, your house, your holidays, your friends, where you spend your weekends, and probably a pile of weird screenshots and photographed documents and just awkward poses and places as well. With current AI, a library like that is something software can actually read and make sense of.

And this isn't just a theoretical thing. Google Photos explains that it uses algorithms to detect faces and group photos of the same person when its Face Groups feature is on. Apple Photos also recognizes people, places and scenes, although Apple says this analysis happens on your own device.

That doesn't necessarily mean these companies are doing anything wrong. And analyzing photos is also not the same as using them to train AI models. Google says it doesn't use people's personal Google Photos data to train generative AI models outside Google Photos. But I think it's important to understand the difference, especially when we're talking about years of private family pictures.

So what can you actually do about it? I would start with the privacy policy of whichever service holds your photos. Look for things like facial recognition, AI, machine learning and model training. Does the service analyze your photos? Does it happen on your own device or on their servers? Can you turn it off? And are your photos used for anything beyond storing and sharing them?

If you can't find the answers, ask the company directly. From my experience, people are already doing that a lot more often.

Funnily enough, I never really thought about these questions being so important when choosing where to keep photos. It was the questions from customers that made me think about it more seriously.

And perhaps that's the bigger point. We are becoming more careful about who gets to see our photos, but we should probably be just as careful about what happens to them where we store them.

Disclaimer: The author is the founder of Yogile, a photo-sharing service that competes with some of the products mentioned.

Maurice Sikkink

Maurice Sikkink is the Founder of Yogile, a private photo storage and sharing platform. With over 15 years of experience in photo-sharing technology, he has seen firsthand how people's attitudes toward privacy have changed. He is particularly interested in how AI is reshaping the way we store, share and protect our personal memories.

This story was produced by Featured and republished with permission. Experts featured here are members of Connectively, Featured's expert network.

(Andre Granito) - We paid for and tested nine VPNs. The advertised monthly price rarely matched the bill.

By Andre Granito

Edited by Adrian James, Editor at Featured News · Oct 7, 2026

This year we paid for and tested nine VPNs (virtual private networks). From every angle. Today I am explaining my VPN research, focusing on the pricing aspects.

While starting our research, I was already surprised. One of the biggest VPN brands was offering a two-year plan at $3.49/month. However, on the checkout page, the total due jumped "a little bit": $94.23, paid upfront, because the deal actually covers 27 months (24 + 3 bonus months).

Monthly prices are one unit, not the full charge. The ads were offering $2.00-$4.99/month. But the real charge was from $56.94 to $119.76. Looking at the real monthly plan, without long-term commitment, the price is around $9 to $16.45.

And this doesn't happen only with this specific provider, or only in the VPN niche.

Personally, and professionally, it is common for me to have problems while registering or renewing domains with popular domain registrars. They offer amazing deals on yearly plans. This way you are locked in with the company, because transferring domain names is not something website owners usually do, I don't really need to explain why. However when the renewal charge comes, you are always surprised.

Median renewal increase of more than 85%. Four out of nine more than doubled, and the biggest jump was +232%.

On the other hand, this is not a general rule: one went up only 8%, and two didn't change at all.

Now, this is the point I want you to understand. Paying for at least five years of coverage cost $158 with the cheapest option and $511 with the most expensive.

This assumes renewal prices stay the same. Taxes and future promos aren't included.

In fact, bonus months and odd timeframes make comparison very difficult. One provider is a good example: $72 buys 36 months (12 paid + 24 free), so $2/month. But the renewal is $72 for only 12 months, which means $6/month.

My tip for you is: compare the effective monthly cost in the whole term.

I found very irritating the fact that many companies find themselves with the right to mislead and deceive their customers. Capitalism expects the seller and the buyer to complete a "transaction" when one offers something the other finds valuable at the price offered. If you need to deceive your customers to keep your company running, something is deeply wrong, and probably it is time to rethink it, move to a different niche, or simply shut it down.

Informed consumers are capable of driving this type of company out of the market.

Beyond the hidden catches in the pricing of these services, you need to keep an eye on other things.

During our research, we realized that most of them offer some kind of free trial, but almost all of those ask for a credit card and convert it into a subscription.

Only one of the nine offered a free trial with no credit card in the desktop app. Paid trials cost between $0.99 and $2.

All nine VPN services we tested refunded 100% of our money. Six on the same day, two the next day and the slowest after 7 days.

While researching one company, we had a few problems getting the refund. In fact, a support agent made our work very difficult from the first question we sent them. After some days of testing, we asked for a refund, and he would not approve it, even though our request was 100% within their terms, accusing us of "using services without paying for it" while testing and asking for refunds.

After disagreements and a heated argument, he instantly deleted our account; this automatically logged me out of the site, preventing me from contacting support again. I was stunned. Later that same day, I received the money back in my bank account.

One last tip about refunds: Purchases done through App Store or Google Play Store always follow the store rules, not the VPN provider's.

About half of the nine VPN providers we tested had a verifiable no-logs policy, the same percentage applies when we search for independent audits.

With one of them, a Domain Name System (DNS) leak test showed our real internet provider's DNS while the VPN was connected.

And here we get to the most important point of this article. When we talk about subscribing to a VPN provider - this can also relate to other niches - we have two possibilities:

1- You are indeed hiring a VPN service that will make your internet navigation safer and more private. Your data and path are actually protected. You are able to access streaming content from other countries, that are blocked in your region.

2- You are buying, paying for a service that is selling your data.

The latter - where you are the product - normally happens when we try to go for the cheapest option.

There are a few free plans I'd trust. With most free VPNs, ask how they make money. Often, the answer is your data.

In this case, there seems to be no difference between the VPN company and your internet service provider; you are simply swapping one company that sees your traffic for another. A VPN company selling your data might be even more dangerous than your internet service provider.

Without further ado, here are 5 simple things that apply to any subscription:

1- The amount charged today;

2- The actual term, including bonus months;

3- The renewal price and term;

4- Who processes the refund (provider or store);

5- Whether the trial requires a credit card.

If you are experienced and have purchased various online services, you have likely encountered some of these issues before and know the steps to take before committing, especially regarding annual or long-term plans.

However, the tips in this article can still be helpful, even to the most experienced internet users.

Often, pricing and renewal rates are the catch.

At other times, unusual terms and bonuses are introduced to confuse you.

And in some cases, depending on the niche, you might end up signing up for something that does you more harm than good.

The only way to avoid unpleasant surprises is to pay close attention to the five points mentioned above, plus the terms of service. Yes, it is tedious, but necessary.

Between May and July 2026, we paid for a plan from each of nine VPN providers (CyberGhost, NordVPN, Private Internet Access, PrivateVPN, Proton VPN, PureVPN, Surfshark, Turbo VPN and VPN.ac), tested them hands-on (baseline speed tests from Latin America), and then requested refunds. All were one-month plans: CyberGhost VPN (1 month plan), NordVPN (Complete Plan, monthly), Private Internet Access (1 month plan), PrivateVPN (1 month plan), Proton VPN (VPN Plus, monthly), PureVPN (Max, monthly), Surfshark (One+, monthly), Turbo VPN (Platinum, monthly), VPN.ac (1 month plan).

The long-term prices in this article come from each provider's U.S. checkout, in U.S. dollars, rechecked on August 14, 2026. These prices are for each provider's cheapest plan, not the higher tiers we bought for testing. All prices are before tax, because tax depends on where you buy. Renewal figures are the renewal prices shown at checkout or in each provider's terms; we did not wait years to watch them charge.

Disclosure: I'm the founder of Reviews Ally, which earns affiliate commissions from the nine VPN providers tested here. We paid for every subscription ourselves, and no provider reviewed or approved this article.

Andre Granito

Andre Granito is a software engineer, founder & lead reviewer at Reviews Ally, an independent site that buys and tests software with its own money. He paid for, tested and requested refunds from nine VPN providers, and published the pricing and refund data. He also tests other types of software.

This story was produced by Featured and republished with permission. Experts featured here are members of Connectively, Featured's expert network.

(NewsUSA) - Americans across the political spectrum agree on how the country should approach artificial intelligence, according to a new report by Gallup and the Special Competitive Studies Project (SCSP), a nonprofit and nonpartisan initiative with a goal of making recommendations to strengthen America's long-term competitiveness in artificial intelligence.

In the report, "Looking to the Future: American Priorities for AI in 2026," majorities of Democrats, Republicans, and independents similarly endorsed survey responses on the need for education and regulation regarding the use of AI. The report was based on a nationally representative survey of more than 3,000 U.S. adults conducted online in June 2026 using Gallup’s U.S. panel.

"The key finding in this year's data is that Republicans, Democrats, and independents want the same things: teach STEM, train workers on the job, put safety and security ahead of speed, and let independent experts test these systems before the public depends on them," said Ylli Bajraktari, president of SCSP. "America's edge in this competition will not come from adoption alone; it will come from a public that has reason to trust what it is adopting,” he said.

Americans’ concerns about the impact of AI on the job market persist, with 70% stating that AI will destroy more jobs than it creates over the next decade, respondents consistently supported the need for AI education. 

Notably, 72% respondents support promoting STEM education as a channel to train workers to use AI in their future jobs, and 66% support creating on-the-job training programs that teach AI skills. However, respondents overall were split on adding AI skills to core K-12 classes such as reading or math, with 45% in support and 44% opposed.

In terms of security, 79% of respondents say the government should maintain safety and data-security rules even if such rules slow AI development, compared to 7% who would prioritize speed. Approximately two-thirds (60%) of respondents agreed that the U.S. government should create AI regulations, and 75% favored testing and evaluation of new AI systems by independent experts before such systems are widely disseminated.

On preparing the workforce, 72% of Americans support promoting STEM education to equip workers with the skills to use AI in their future jobs, and 66% support creating on-the-job training programs that teach workers AI skills. Republicans, Democrats, and independents express similar levels of support for both. Opinions remain split on adding AI skills to core K-12 classes like math or reading, with 45% in support and 44% opposed.

Half of the respondents (51%) say they use AI at least sometimes at work, school, or in their personal lives, up from 39% in 2025. Two in three (68%) describe themselves as knowledgeable about AI, up from 60% last year. The new report also shows increased trust among regular AI users since last year, but increased skepticism among non-users. 

"The public is becoming more experienced with AI while simultaneously becoming more cautious about its economic implications," said Gallup Principal, Justin Lall. "That combination of firsthand use and lingering concern makes workforce preparation more important than ever."

The data suggest room for compromise where attitudes towards AI diverge among parties. Republican respondents were more likely to favor mandatory age requirements, increasing AI in manufacturing, increasing energy production for AI systems, and reducing AI regulation. Democrat respondents were more likely to favor increasing government funding for AI research at colleges and universities. However, a majority of respondents across parties showed similar support for allowing companies and universities to use government resources for AI research. 

Visit scsp.ai to view the full report and learn more. 

(NewsUSA) - BIX, the neo-finance app that pairs a self-custody crypto wallet with a Visa card, is accelerating its push into the United States, appointing David Imonitie as head of US sales as it works toward near-nationwide availability.

The company currently supports eligible residents of 29 US states, subject to identity verification and programme requirements, and says it is targeting expansion to 49 states within six to eight months, subject to applicable approvals and partner readiness.

"We are excited about the states we are open in and have great expectations for the rest. That will not dampen our enthusiasm or our work ethic in growing customer acquisition. The other states will open, and in the meantime we will grow," Imonitie said.

BIX is not a new entrant. The app and its virtual card programme are already operating, and the company describes the current phase as an expansion of an established business rather than an initial launch.

The product combines a self-custody wallet with a Visa card programme. Users hold USDC or USDT across supported combinations of Ethereum, BNB Smart Chain, Base, Solana and Optimism, and spend against that balance at merchants worldwide. Purchases run through Visa and settle to the merchant in fiat.

Standard membership is free and a virtual Visa card costs $9.99. The company says there are no funding fees, with a 1% FX fee applied to eligible foreign-currency purchases under the US programme. Spending limits are, in BIX's description, flexible and customizable based on the user's deposited collateral, subject to programme requirements.

The business operates as Bix Ltd, registered in the Cayman Islands with operating offices in Dubai. BIX is a finance app rather than a bank, with card issuing and other regulated financial services delivered through third-party institutions under their respective arrangements.

Several additions are in development. BIX Plus, at $119 a year, will bundle access to Visa Signature cards with eligible benefits such as airport lounge access, rental-car insurance and travel discounts, with packages depending on programme and country. Lounge access does not necessarily include complimentary visits.

Plus will also introduce virtual accounts, supporting incoming payments and outgoing fiat transfers connecting stablecoin balances with ACH, SEPA and SWIFT, including details such as IBANs. Physical cards are planned at $99, separate from membership, with a 0.6% BIX ATM withdrawal fee.

The most substantial change in the pipeline is direct wallet-linked spending, under which eligible card purchases will draw against supported wallet balances without a separate top-up. Until that rollout completes, card funding uses a separate smart contract.

Crypto cards have existed for years, but the category has been held back by two problems: volatile assets unsuited to everyday spending, and custody arrangements that required users to hand their assets to a provider. Dollar-pegged stablecoins addressed the first. Self-custody models are an attempt at the second.

Imonitie said it was the simplicity of the product that drew him to the role.

"What made me decide to take this role was very simple. I wanted to be at the forefront of the future of payments. The simplicity that BIX provides was undeniable and exciting for all people, whether you understand crypto or not."

Availability remains the sector's binding constraint. No provider in the category currently offers uniform nationwide US coverage, and expansion timelines depend on approvals and partner institutions rather than on product readiness.

More at bix.fi and on Instagram.

David Imonitie is head of US sales at BIX. Find him on Instagram.

(NewsUSA) - For decades, companies trying to reach consumers followed a familiar formula: create a compelling story, earn or secure media exposure, and make sure the information could be found through search engines.

Artificial intelligence is rapidly changing that equation.

Millions of consumers are increasingly asking platforms such as ChatGPT, Google’s AI tools, Gemini, Claude and Perplexity for recommendations and answers instead of simply typing keywords into a traditional search box. For businesses, that raises a new question: What information will AI systems find when they try to understand a company, product or organization?

It is a question that News USA has been thinking about from an unusual vantage point.

Founded nearly four decades ago, News USA built its business around content syndication—the distribution of editorial-style articles through a large network of news and information publishers. The company has worked with more than 32,000 organizations over its history, helping companies and nonprofits place informative content in media outlets around the country.

Today, that established distribution infrastructure is finding a new role.

“Companies used to think primarily about whether people could find them on Google,” said Rick Smith, founder and CEO of News USA. “Now they also have to think about what artificial intelligence knows about them, where that information comes from and whether their competitors are more visible.”

News USA can distribute a single editorial article across thousands of news and information websites, creating a broad digital footprint rather than relying on a single media placement.

That distinction could become increasingly significant in an AI-driven information environment.

AI platforms evaluate enormous amounts of information published across the web when generating answers. While no company can guarantee what an AI system will say or cite, having accurate, useful information about an organization appearing across independent third-party websites can expand the body of information available online about that organization.

News USA is now combining its traditional syndication capabilities with a growing focus on AI visibility—examining how brands appear in AI-generated answers and how third-party media coverage may contribute to their broader digital authority.

The change reflects a larger transformation underway throughout marketing and public relations.

Traditional publicity is no longer simply about getting an article published. Companies increasingly need to consider how information travels from publishers to search engines and ultimately into AI-powered discovery tools.

For Smith, the technology may be changing, but the underlying principle has not.

“Whether someone is reading a newspaper, searching Google or asking an AI assistant,” he said, “credible third-party information still matters. Our job is to help organizations get their stories into places where people—and increasingly machines—can discover them.”

After nearly 40 years of syndicating stories, that may give an old communications model a surprisingly important role in the newest generation of information technology.

(NewsUSA) - Quantum information science, engineering, and technologies (QISET) is in its second phase, now rapidly evolving from lab-based projects to real-world applications that are impacting the strategies and finances of the federal government and the private sector, according to experts at the Special Competitive Studies Project (SCSP), a nonprofit and nonpartisan initiative with a goal of making recommendations to strengthen America's long-term competitiveness in artificial intelligence.

In the first phase of QISET development, known in the industry as Quantum 1.0, quantum mechanics and science fostered the development of lasers and transistors. Today, Quantum 2.0 activity centers around developments in areas including photonics, microelectronics, and specialized materials. 

As the United States seeks to up its game in quantum information science, SCSP highlights the history, foundation, and future of place-based quantum innovation in a new limited newsletter, "Quantum States," which examines how different states and regions are leading the way.

States were assessed across a range of metrics, including cited quantum information science research, patents, the volume of both 'pure play' and quantum-enabling  companies, the number of military research facilities, breadth of the full quantum stack (defined as computing, sensing, and networking), and the number of quantum-related job openings. 

"Developing a robust ecosystem of quantum technologies is not a one-size-fits-all approach, in which local leadership is guaranteed by hosting the most companies within a region. Rather, it requires a cohesive, strategic effort leveraging strengths within that state," according to the SCSP experts. 

Currently, California leads in most aspects of Quantum 2.0, including industrial capacity, talent pipelines, and market ecosystems, but other established hubs include New York, Illinois, Colorado, Maryland, and Massachusetts. 

However, several other states are poised to become players, including, Texas, North Carolina, and Florida given their high number of PhDs awarded, and the presence of Quantum research centers in many of their academic institutions. 

"The most successful quantum states build strong synergies between pillars of a quantum ecosystem: academia, NIST, national labs, startups, and private industry," according to the SCSP experts. Ultimately, true quantum hubs emerge where industry, academia, and government actively connect.

Visit scsp.ai to learn more and for future editions of the Quantum States newsletter.

(NewsUSA) - Proposed rule changes at the Office of Management and Budget (OMB) could weaken the research enterprise that underpins American technology leadership, according to experts at the Special Competitive Studies Project (SCSP), a nonprofit, nonpartisan initiative that makes recommendations to strengthen America's long-term competitiveness in emerging technologies.

OMB has proposed the most sweeping revision in years to the regulations governing federal grants and research awards. SCSP has filed formal comments urging the agency to revise the rule before finalizing it.

Accountability for taxpayer dollars is not optional, and SCSP supports OMB's stated goals of transparency, research integrity, and stewardship, said Ylli Bajraktari, president of SCSP, in a recent article. Political accountability and scientific expertise are not enemies. But applied without clear limits, the rule risks damaging the innovation engine that has been the backbone of America's scientific success, he said.

The new regulations contain several concerning provisions, said Bajraktari. Specifically, "Senior political appointees would review every discretionary award, reviewers would be directed to weigh presidential policy priorities, peer review would be explicitly downgraded to 'advisory,' and agencies could terminate active research awards simply because priorities changed after the work began," he said.

The stakes are measurable, according to SCSP's 2026 Tech Competition Scorecard, which found that China holds a decisive overall lead in robotics for advanced manufacturing, and that the narrow U.S. lead in quantum is eroding under Beijing's coordinated, state-backed strategy. Meanwhile, the federally funded share of national research and development dropped by nearly one-third between 2010 and 2019, and federal AI research spending remains far below the $32 billion annual level recommended by the National Security Commission on Artificial Intelligence. America already faces a funding gap, Bajraktari said, and should not compound it with a confidence gap.

The proposed changes take their cues from a termination-for-convenience model of government contracting, but research grants are different, Bajraktari emphasized. A research grant supports multiyear experiments, doctoral researchers, and laboratory partners using custom equipment and generating data over time. Stopping that work midstream destroys value that reimbursement cannot recover. The losses also reach companies, investors, national labs, and startups that make decisions based on whether federally funded research is stable and merit-driven.

"When award decisions look political rather than technical, or when a grant can vanish because priorities shifted, private partners hedge, talent looks elsewhere, and the whole geometry of innovation weakens," said Bajraktari.

SCSP's comments ask OMB to consider four points before finalizing any changes to how research grants are handled.

• Keep merit at the center. Scientific and technical merit, evaluated by qualified experts, should remain the primary basis for selecting research proposals. Senior appointees have a legitimate oversight role in compliance, security, and program fit, but should not substitute their judgment for expert evaluation of scientific quality, and any override should be documented in writing against published criteria.

• Make awards durable. Competitively awarded research grants should not be terminable simply because policy priorities changed after the fact. Reserve termination for legal violations, security concerns, or performance failures, and give recipients a chance to respond and to wind down responsibly.

• Account for the private sector. Before changing how research awards work, agencies should ask whether the change will deter co-investment, interrupt commercialization pathways, or push globally mobile talent toward America's competitors.

• Assess the competitive impact. OMB's own analysis counts only paperwork costs and calls them modest. Forgone discoveries and deterred capital are real national security costs even when they never appear as a line item, so major changes to federal research policy should come with a technology-competitiveness impact assessment.

SCSP also urged OMB to slow down. An October 1 effective date would impose new selection and termination frameworks on fiscal year 2027 awards before agencies have built the procedures to implement them well, Bajraktari said.

Visit scsp.ai to learn more.