One reason audiophile-leaning reviewers like myself still exist is that we listen closely to ensure new releases are indeed as fabulous as the manufacturers would like you to think. Thankfully, these days, they often are, and sometimes we discover extra surprises.

Consider the new reissue from Craft Recordings of Moodsville Volume 8: The Frank Wess Quartet. Originally released in 1960 via Prestige Records’ then-new Moodsville line, the subsidiary was a clever way to champion respected players outside the maelstrom of rapidly evolving progressive jazz. Moodsville was a far cry from the cheesy lounge-jazz albums flooding the market at the time by relative unknowns (often budget-line releases). Prestige Records’ Ron Eyre described the Moodsville concept as “Jazz performances of ballads wherein the musical integrity of the artist is maintained and at the same time the original beauty and feeling of the ballad is not lost.”

While this music may not stir your musical moneymaker like Ornette, Miles, or Mingus might have at the time, it had a place serving up super enjoyable, swinging, melodic chill sounds for reflective moments. Moodsville released nearly 40 LPs, including albums by Gene Ammons, Coleman Hawkins, Yusef Lateef, and Kenny Burrell, most recorded by the legendary Rudy Van Gelder.

Official press materials confirm: “Each album will be released on vinyl in partnership with audiophile leader Acoustic Sounds, a partner Craft has previously collaborated with on its acclaimed Bluesville Records and Contemporary Records Acoustic Sounds series. Delivering the highest-quality listening experience, the LPs feature all-analog mastering (AAA) from the original tapes by GRAMMY-nominated engineer Matthew Lutthans (The Mastering Lab) and are pressed on 180-gram vinyl at Quality Records Pressing (QRP). Faithfully reproduced tip-on jackets round out every release.”

Original vinyl copies of Moodsville Volume 8: The Frank Wess Quartet are not easy to find in good shape. My so-so copy is the only one I’ve found to date (until now!). A lush treat, this reissue’s surprise is that it is in stereo. Nowhere on the cover, hype sticker, or in the press materials do the producers indicate that this is a stereo mix, so that’s a sweet bonus. Overall, it sounds markedly less compressed, more natural, and more alive than my mono original.

The mix places Bobby Donaldson’s drums off to one side and Wess on the other, glued together by bassist Eddie Jones and Tommy Flanagan’s piano in the center. This approach probably had something to do with how the mono mix sounds. Van Gelder’s recording process (confirmed for me years ago when researching some Blue Note reissue reviews) changed once stereo LPs became a commercial reality in 1958.

He switched over to stereo recording and pretty much never looked back, his subsequent mono mixes effectively folded down from two-channel. Nonetheless, his wizardry still made great, punchy-sounding mono releases. I suspect his recording design here could fold down neatly into mono.

Moodsville Volume 8: The Frank Wess Quartet delivers exactly what the label intended: fine moods equally enjoyable for attentive listening or background spins. What’s not to like?

Our Ratings

★★★★★★★★★★ Music

★★★★★★★★★★ Sound

★★★★★★★★★★ Packaging

Where to Buy

Mark Smotroff is a deep music enthusiast / collector who has also worked in entertainment oriented marketing communications for decades supporting the likes of DTS, Sega and many others. He reviews vinyl for Analog Planet and has written for Audiophile Review, Sound+Vision, Mix, EQ, etc.  You can learn more about him at LinkedIn.

A virtual phone number is a normal dialable telephone number whose destination can be controlled through software or a communications service instead of being permanently tied to one physical phone or fixed line. When someone calls it, the provider routes the call according to configured rules, which may send it to an app, desk phone, another telephone number, voicemail or another supported workflow.

What Is a Virtual Phone Number?

A virtual phone number separates the number people dial from the physical device that ultimately handles the communication. Twilio’s technical glossary describes a virtual phone number as a standard telephone number that is not locked to a specific phone and can route calls or messages to supported destinations or workflows.

That distinction is the key to understanding the term. The telephone number still acts as an address that other people can dial. Behind it, however, routing rules determine what should happen when a call reaches the service responsible for that number.

For example, imagine a small company with one customer-service number. During office hours, calls might ring a softphone, meaning a software-based phone application, used by two employees. After hours, the same number could send calls to voicemail or forward them to an on-call phone. Customers continue dialing the same number even though its destination changes.

Public telephone numbers also exist within national and international numbering systems. The International Telecommunication Union’s current E.164 international public telecommunication numbering plan, approved on February 13, 2026 and currently in force, defines the international numbering framework. A virtual-number service adds configurable routing and communications services behind such a number rather than creating an entirely separate kind of address that callers need to understand.

How Does a Virtual Phone Number Work?

The simplest mental model is to separate the process into three things: the number the caller dials, the provider responsible for processing that number, and the endpoint or workflow selected by the provider’s routing rules.

1. Someone Dials the Public Number

To the caller, the process can look ordinary. They dial a telephone number from a mobile phone, landline or another compatible calling service. They do not need a special app merely because the destination uses cloud telephony or virtual routing.

The telephone network then routes the call toward the carrier or communications platform responsible for handling that number. Exactly how this occurs varies across carriers, countries and service architectures, so it is more accurate to describe a general routing model than to assume every virtual-number provider uses identical infrastructure.

2. The Responsible Service Receives the Call

Once the call reaches the service hosting or processing the number, the platform identifies the configuration associated with it. A useful real-world example comes from Twilio’s inbound call processing documentation: when a call reaches one of its numbers, Twilio looks up routing information for that number and processes the call according to its configuration.

Twilio is one implementation rather than a universal blueprint, but it illustrates the central principle. The number is associated with routing information, and that information determines what the communications platform does next.

3. Routing Rules Decide What Happens

Depending on the service and configuration, an incoming call might:

• ring one user or extension;
• ring several employees or devices;
• enter a call queue;
• play an automated menu or interactive voice response prompt;
• forward to another telephone number;
• reach a Session Initiation Protocol (SIP) endpoint;
• trigger an application-controlled workflow; or
• go to voicemail when nobody answers.

These are examples rather than guaranteed features. The precise options depend on the provider, number type, country and service plan.

4. The Call Reaches the Selected Endpoint

An endpoint is simply the place where the communication is ultimately handled. It might be a mobile or desktop app, an Internet Protocol (IP) desk phone, a SIP device, another conventional telephone number or a voicemail system.

This makes location and device less rigid than they are with a telephone service permanently associated with one line. A support worker can potentially answer the same business number on a laptop at home, an IP phone in an office or another configured device without asking callers to learn a different public number.

5. Outbound Calls Use the Service in the Other Direction

Many virtual-number services also support outbound calling. A user starts a call through the provider’s app, phone system or another supported endpoint. The provider connects that call toward the recipient and, where permitted and correctly configured, presents an authorized telephone number as the calling number.

The important point is that the number itself and the device used to handle the call do not have to be the same thing. The provider sits between them and applies the account’s calling and routing configuration.

Is a Virtual Number the Same as VoIP?

No. A virtual phone number and Voice over Internet Protocol (VoIP) are closely related concepts, but they describe different parts of a communications system.

A virtual phone number is the address people dial and the routing relationship attached to it. VoIP is a technology for carrying voice communications using Internet Protocol networks. Many cloud phone services combine the two, which is why the terms are often used as though they were interchangeable.

A useful analogy is a street address and a delivery vehicle. The address identifies where something should be delivered. The vehicle is one way of transporting it. Similarly, a telephone number identifies a communications destination, while VoIP can provide part of the technology used to transport and process a call.

This distinction matters when diagnosing problems. A number may still exist even when an app is offline, while a VoIP call can suffer from network conditions such as latency or packet loss even though the number itself is configured correctly. Voice quality problems involving packet transport, codecs and network delay are therefore questions about how VoIP works rather than properties of the virtual number itself.

Virtual Number vs. SIM, eSIM and Call Forwarding

Several technologies can produce similar results from a user’s perspective, but they operate at different layers. The table below focuses on what each concept represents rather than on individual provider features.

The GSMA explains that a conventional SIM provides a secure means of authenticating devices to mobile networks, while eSIM technology allows SIM information to be securely downloaded into an embedded secure element. An eSIM therefore changes how a mobile subscription can be provisioned into a compatible device. It is not another name for cloud telephone-number routing.

The technologies can also coexist. Someone could have a virtual business number that rings a smartphone whose cellular subscription is delivered through an eSIM. The virtual number and eSIM are doing different jobs in that arrangement. The practical differences become especially important when comparing a virtual phone number vs. eSIM for travel, business calling or maintaining a second contact number.

What Types of Virtual Phone Numbers Exist?

Virtual-number providers can offer several number categories, although the available types and permitted capabilities depend on the country and provider. Twilio, for example, documents offerings that include local, national, mobile and toll-free numbers across supported markets.

Common categories include:

• Local or geographic numbers: numbers associated with a particular city, area or numbering region.
• National or non-geographic numbers: numbers intended for broader national use rather than one specific local area, where the country’s numbering plan provides them.
• Mobile numbers: numbers assigned from mobile ranges in markets where the provider offers them.
• Toll-free numbers: numbers for which the called organization generally bears the relevant call charge rather than the caller, subject to the applicable country’s system.
• Vanity numbers: memorable numbers selected partly for branding or recognizability. The memorable format does not imply a different routing technology by itself.

Do not assume that a number supporting voice will automatically support Short Message Service (SMS), Multimedia Messaging Service (MMS), short codes or every other communications feature. Number capabilities can differ by number type, carrier, provider and jurisdiction.

What Are Virtual Phone Numbers Used For?

The main attraction of a virtual number is flexibility in what happens after somebody contacts it. That creates several practical uses without requiring the public-facing number to change.

One Number for a Distributed Team

A company can publish one support or sales number while allowing staff in different locations to answer calls through supported applications or devices. Routing rules can send a call to an available employee, a queue or voicemail instead of depending on one desk phone.

Separating Public and Personal Numbers

A freelancer, small-business owner or employee may want customers to use a business number rather than publishing a personal cellular number. The business number can then route to a device the person already uses.

Local Contact Points

Where regulations and provider availability permit it, an organization may use numbers from different geographic markets while handling calls through a centralized communications system. A local-looking number does not by itself prove that the person answering is physically located in that area.

Call Tracking and Department Routing

Different numbers can be associated with departments, locations or campaigns while still feeding into the same communications platform. Twilio documents call tracking and direct inward dialing as uses of virtual phone numbers.

Number Masking

Some platforms place an intermediary number between two participants so they can communicate without directly exposing each other’s personal telephone numbers. For example, a marketplace could provide a proxy number that lets a buyer and seller communicate without exchanging their personal numbers.

That creates separation between the participants’ public contact details, but it does not guarantee anonymity from the communications provider. The provider may maintain account, routing or other records under its policies and applicable law.

Limitations and Risks You Should Understand

Virtual numbers are flexible, but the phrase describes a broad category rather than a guarantee of specific capabilities. Several limitations deserve attention before you depend on one for an important use.

Internet and Service Availability Can Affect Calls

If your destination is a VoIP app, browser or IP phone, receiving calls can depend on Internet connectivity, electrical power, device configuration and the availability of the communications service. A number may be correctly provisioned while its chosen endpoint is temporarily unreachable.

Forwarding to a conventional telephone number can create a different dependency chain, so it is inaccurate to say that every virtual-number call requires the final user to have an Internet connection.

SMS Support Does Not Guarantee Verification-Code Support

A number may technically receive SMS while still being rejected by a bank, social network or another service for account verification. Acceptance can depend on the receiving platform’s policies, number classification, carrier relationships and the messaging route involved.

That means “supports SMS” and “works with every one-time password service” are not equivalent claims. The reasons some VoIP numbers cannot receive verification codes can involve restrictions outside the virtual-number provider’s direct control.

Caller ID Does Not Prove Who Is Calling

A virtual phone number should not automatically be treated as suspicious, but the telephone number displayed on a screen is not reliable proof of identity. The U.S. Federal Trade Commission warns that scammers can make fake names or numbers appear on caller ID, a practice known as spoofing.

Spoofing and virtual numbering are different concepts. A legitimate business can use a valid virtual number, while a scammer can falsify caller ID without legitimately controlling the displayed number.

Emergency Calling Requires Special Attention

Never assume that the presence of an ordinary-looking telephone number means emergency calling works exactly like it does on a traditional fixed or mobile line.

In the United States, the Federal Communications Commission has required providers of interconnected VoIP service to provide Enhanced 911 (E911) capabilities as a standard feature. FCC material also explains that interconnected VoIP providers use registered-location information to route 911 calls and provide emergency centers with the caller’s callback number and location information.

Those rules do not mean every product marketed as a virtual number has identical emergency-calling functionality. Before relying on a service for emergency calls, verify the provider’s current emergency-calling support, registered-location requirements and limitations for the country and service you will use.

Phone Number Rules Differ by Country

Number availability is not simply a commercial decision made by an app provider. Countries regulate their numbering resources, and requirements can vary by number type and user category. Twilio’s current regulatory documentation explains that phone-number compliance can require end-user identity and supporting documents according to the applicable country, number type and regulatory requirements.

A provider offering a local number in one country may therefore require identity, address or business documentation that it does not request for another country’s number. Some number types may also be unavailable to users who cannot satisfy applicable eligibility or documentation requirements.

Control and Portability Matter

If a telephone number becomes important to customers, clients or account recovery, consider what happens if you later change providers. Ask whether the number can be ported, what information is needed to authorize a transfer and what happens after cancellation.

In the United States, FCC materials document local number portability between several wireline and wireless service arrangements while also noting that geographic and technical limits can apply. The FCC’s number-portability review explains both the value of retaining a telephone number and existing portability limitations. Rules and technical eligibility differ internationally, so do not assume that every virtual number in every country can be transferred to another provider.

What Should You Check Before Getting a Virtual Number?

Start with what you actually need the number to do. A service that is appropriate for occasional inbound calls may not meet the requirements of a support team, a regulated business or someone who depends heavily on messaging.

• Voice capability: Check whether you need inbound calls, outbound calls or both.
• Messaging: Verify SMS and MMS separately instead of assuming every number supports both.
• Number type and location: Confirm that the required country, area code or number category is actually available to you.
• Endpoints: Check whether you can answer through mobile apps, desktop apps, browsers, SIP phones, forwarded numbers or the specific devices you use.
• Routing: Determine whether you need queues, simultaneous ringing, schedules, extensions, menus or voicemail.
• Portability: Find out whether the specific number can be transferred if you leave the provider and whether geographic or regulatory limits apply.
• Emergency calling: Verify the service’s emergency-call support and any registered-location requirements.
• Verification codes: If receiving one-time codes is important, confirm acceptance for the particular services you use rather than relying only on an “SMS capable” label.
• Compliance: Check identity, address, business-registration or local-presence requirements before obtaining an international or geographic number.
• Security: Review account protections such as multifactor authentication, administrator permissions and recovery procedures because control of an important business telephone number can be operationally sensitive.
• Costs: Compare the recurring number charge with inbound, outbound, forwarding, messaging, international and usage-based charges that apply to your expected traffic.

Before publishing the number widely, test the complete communications path yourself. Place inbound and outbound calls using the devices and networks you expect to use, confirm voicemail or fallback routing, and verify that caller ID and routing behavior match your intended configuration.

The Short Version

A virtual phone number separates the telephone number people dial from one permanently attached physical destination. Calls still travel through telecommunications networks to the service responsible for the number, but that service can use configurable routing to decide which supported device, telephone number or workflow handles the communication.

That flexibility is why virtual numbers are useful for distributed teams, business-personal separation, call routing and masked communications. It does not make every service identical. Voice and messaging capability, verification-code acceptance, emergency calling, portability and regulatory requirements can all vary by provider, number type and country, so those details should be checked before the number becomes operationally important.

Five alleged leaders of the Black Axe cybercrime syndicate, known for its involvement in global-scale cyber-enabled financial fraud, have been extradited to the United States to face wire fraud and money laundering charges.

Prosecutors accuse Perry Osagiede, Franklyn Osagiede, Osariemen Clement, Collins Otughwor, and Musa Mudashiru of coordinating an internet fraud campaign from Cape Town from 2011 to 2021 that involved advance fee schemes and romance scams.

The five accomplices allegedly used multiple aliases, social media and online dating websites, and voice over internet protocol (VoIP) phone numbers to talk to victims in the United States and trick them into sending money after making them believe they were in a romantic relationship.

When their targets refused to send them money, they coerced them into making the payments using various manipulation tactics, including threats of publishing sensitive photos of the victim online.

“Black Axe is a notoriously violent transnational criminal organization that also happens to dabble in romance scams to make money,” FBI Newark Special Agent in Charge Stefanie Roddy said.

“The ability of FBI Newark and our partner agencies to reach into South Africa illustrates our resolve to hold accountable any and every type of fraudster who preys on innocent victims here in the United States.”

The defendants were arrested in South Africa in 2021 at the United States’ request and extradited to the United States on September 11, 2026.

If found guilty, they face a maximum of 20 years in prison for the wire fraud charges, a maximum of 20 years for money laundering, and two more years for aggravated identity theft charges.

Last month, law enforcement agencies from 22 countries arrested 58 individuals and identified 263 suspects linked to cybercrime networks coordinated by African crime groups as part of “Operation Jackal IV,” an international joint action focused on disrupting the Black Axe criminal ring.

Spanish authorities also arrested 34 cyber fraud suspects believed to be part of a criminal network connected to the Black Axe group.

Founded in 1977 in Nigeria, Black Axe is one of the world’s most dangerous and far-reaching cybercrime syndicates, believed to have an extensive network of money mules and facilitators, as well as 30,000 registered members.

In January 2024, the United States sentenced Black Axe member Olugbenga Lawal to ten years in prison for laundering millions stolen by Black Axe operators from elderly Americans in internet fraud schemes.

Join Mikko Hyppönen and security leaders from the NFL, CHANEL, and Atlassian for a two-hour digital summit on what AI-speed attacks change, what defenders should stop doing, and how to validate, decide, fix, and re-validate at machine speed.

Save your seat

MediaTek on Tuesday unveiled its new Dimensity 9600 Pro and 9600M phone processors, both of which highlight support for advanced gaming as well as artificial intelligence features for upcoming phones.

Phone processors are an important component in any device — they help power everything from how fast a phone runs, image processing, games with intensive graphics and running certain AI features on your device without needing to connect to the internet.

The Dimensity 9600 Pro will be the company’s first to use the 2-nanometer process in partnership with chipmaker TSMC, which we’re also seeing in rival chips from companies like Apple’s A20 Pro processor that’s shipping with the iPhone 18 Pro and Pro Max. This should mean the processor is denser, allowing for more performance and power efficiency.

MediaTek says the 9600 Pro will support faster performance and on-device AI features over last year’s Dimensity 9500, and for gaming, it will be capable of console-quality gaming at an exceedingly smooth 185 frames per second.

The 9600M will be the step-down processor meant for a wider range of devices, and MediaTek says that it will still power AI features as well as gaming at the same 185fps.

MediaTek’s new high-end processors will face their most competition from Qualcomm’s Snapdragon line, the latter of which typically powers Samsung’s Galaxy S and Z phones in the US. We’ll soon know what Qualcomm’s next high-end processor will look like, as the company will likely use its Snapdragon Summit to debut it on Sept. 22. But MediaTek’s high-end processors are found in a lot of devices internationally, with last year’s 9500 powering phones like the Oppo Find X9 Pro.

Phones running on MediaTek’s Dimensity 9600 Pro and 9600M are expected to debut later this year.

Mike Sorrentino

Senior Editor

Mike Sorrentino is a Senior Editor for Mobile, covering phones, texting apps and smartwatches — obsessing about how we can make the most of them. Mike also keeps an eye out on the movie and toy industry, and outside of work enjoys biking and pizza making.

See full bio

Advertisement

A total of 680 account holders have been affected globally.

It is understood that sensitive personal information belonging to 12 Revolut account holders in Ireland has been leaked in a data breach.

Last Saturday (12 September), Revolut said that it disclosed customer data to an unauthorised fraudulent party after receiving requests for information from an email with a government agency domain. The company mistakenly fulfilled the data requests before realising that the sender was phishing for data, it said in a community alert.

The leaked data includes full names, dates of birth, occupations, addresses and contact details, passports, drivers’ licenses, financial statements and facial verification images.

“Revolut recently identified a sophisticated external impersonation scam where an unauthorised third party utilised a legitimate government agency domain email to submit fraudulent requests for information,” a Revolut spokesperson said.

“Upon detection, we immediately blocked the address and alerted the relevant government agency as well as enforcement agencies, data protection, and financial regulators,” they added.

“Revolut systems and customer funds are unaffected. We have contacted the limited number of impacted individuals directly to inform them and provide support.”

A total of roughly 680 account holders are understood to have been affected globally. Revolut has more than 80m customers across the world, with roughly 3.4m based in Ireland.

The data leak comes as the fintech continues on its aggressive expansion campaign to establish itself as a licensed banking institution in major markets globally, including the UK, US, France, Australia, Mexico and Peru.

Reports from April suggest that the company is on its way to a $200bn valuation in an eventual initial public offering expected on or after 2028. It was last valued at $75bn after a major share sale last November.

It recently received approval to provide crypto services in the United Arab Emirates, adding to its more than 16m crypto customers across Europe and the UK. The fintech also announced its own euro-backed stablecoin called EURR last month.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

Three senators will put AI legislation in front of Congress this week, and none of them can say when it might be voted on.

John Kennedy, a Republican from Louisiana, plans to introduce a measure on Wednesday requiring developers to build a shutdown mechanism into their systems.

Bernie Sanders is hosting a briefing with AI researchers the same day and has a bill that would ban artificial superintelligence outright. Josh Hawley has already introduced bipartisan legislation covering national security, civil liberties and labour protections.

The Associated Press canvassed Congress on Tuesday and found the same answer in both parties, which is that nothing will move. Rick Scott, a Republican from Florida, ran through the list of things Congress has not finished, election security and a budget among them, and asked:

“So you think we’re going to get AI done?”

Mark Warner, Democrat of Virginia, put it as a rule: “We never do anything until we have a holy heck moment.”

What makes the stall unusual is who is asking. Dario Amodei said on Saturday that the industry must slow down, Sam Altman agreed the same day, and Elon Musk added his support. Rahm Emanuel, who ran Barack Obama’s White House, told the AP that in his whole career he had never seen it:

“I’ve never had an industry or a CEO ask for regulation and oversight.”

Donald Trump has answered the request repeatedly and in the negative. He called the case for limits a conspiracy on Monday, having dismissed extinction fears the week before, and one of his first acts upon returning to office was to rescind Joe Biden’s 2023 executive order on AI oversight.

JD Vance has been cautious about regulation. Mike Johnson said legislatures reflexively cover things in red tape, and suggested Trump summon the AI companies to a meeting rather than pass a law.

Senate Republicans are not uniformly with him. John Thune, the majority leader, told the AP that Congress can set guardrails around the more consequential threats “without harming our ability to stay ahead in the AI race”, and he is working with Amy Klobuchar on a bill that would give Washington power to block unsafe models and impose a duty of care on developers.

Hawley wants protections for workers, for children online and for people living near data centres, while saying plainly that frontier models will not blow up the world.

A bipartisan working group led by Chuck Schumer reported in 2024 that the United States should spend at least $32bn over three years on AI development and safeguards. Almost none of it followed.

Pressure from outside Congress has kept rising throughout it. Obama has been telling candidates they need an AI policy of their own; Max Stier of the Partnership for Public Service told the AP that delay will prove costly, and researchers at the labs have spent the summer warning publicly about what their employers are building.

David Sacks, until recently Trump’s AI adviser, told the same companies on Sunday to slow down quietly and stop asking Washington for anything in return.

None of the three bills has a committee date, a floor date, or a companion in the House. Kennedy’s kill switch arrives on Wednesday into a chamber whose own leader has said the votes are not there for the broader bill he is already negotiating.

This story appeared in Today, Explained, a daily newsletter that helps you understand the most compelling news and stories of the day. Subscribe here.

If you’d asked me yesterday what Steve Bannon and Bernie Sanders had in common, I might have said “white hair.” But after a wild week in which AI doomerism burst solidly into the mainstream, the right-wing provocateur and progressive senator are joining forces — with various religious leaders, lawmakers, and the actor Joseph Gordon-Levitt, no less — to express their deep concern about AI regulation.

Concern about cutting-edge AI models had been mounting for months as those models increasingly behaved in ways their makers didn’t expect. It exploded last week, when a researcher at the AI company Anthropic resigned publicly and warned that the people building AI “earnestly believe it could kill us all by the end of the decade.”

It sounds like science fiction. It is the stuff of science fiction. But it’s not a fringe position: Over the weekend, leading AI researchers and executives also warned that AI is developing dangerously fast and called for measures to rein it in. (We’ll run through those options in just a minute.)

Now, a wide spectrum of lawmakers, activists, and pundits are also calling for stronger safeguards — Sanders and Bannon among them. The pair will speak at an event billed as a “Pro-Human Assembly” in Washington on Tuesday alongside Republican Sen. Marsha Blackburn, Congressional Progressive Caucus chair Greg Casar, and the archbishop of San Francisco, among others.

But President Donald Trump is decidedly not on board, and it’s hard to see how any substantive AI regulations could pass without him.

“There is a SICK conspiracy going on against AI and Data Centers,” he wrote this morning on Truth Social, “and the only one happy about it is China.”

All three approaches have precedent in federal regulation, so they wouldn’t be (at least in theory) too tricky to stand up. The real roadblock, as always, is political. One AI researcher told Vox he worries we’ll miss the opportunity to regulate AI while waiting for lawmakers to get on the same page.

Even under threat of apocalypse, Washington, DC, moves far, far slower than Silicon Valley.

Amazon Prime Video is coming after TikTok with the addition of short-form news clips. The company said on Monday that it’s expanding its news coverage to include on-demand news clips offering both local and national stories. These will be available on TVs and other living room devices, but will soon roll out to the web and Prime Video’s mobile apps, the company said.

The clips will be available in the News destination within the Prime Video app in its Trending Topics and the Local News carousels. They join Prime Video’s existing local, national, and world news coverage from a number of major providers, which are available to everyone in the U.S. regardless of whether they have a Prime subscription.

Amazon could not yet confirm whether the clips on mobile would be oriented for vertical video or not, only that they would be considered short-form content on iOS and Android.

While TikTok is better known as an entertainment destination, it’s also, along with Instagram, increasingly a place where younger people get their news. Gen Z in particular often watches news creators and influencers on short-form apps, including YouTube Shorts. Amazon, which doesn’t have a true social destination of its own, is instead adding short-form video to Prime Video, in hopes of capturing Gen Z’s attention.

But Gen Z viewers are drawn to these social apps not necessarily because of the length of the clips, but rather because of how the information is presented. Based on findings from Pew Research, these younger news consumers want short-form news content that “feels organic, not over-produced.”

Prime Video is not the first streamer to add short-form clips to its content lineup. Peacock added support for short-form clips in March, followed by HBO Max and Netflix in July, then Disney+ in August. Combined, the updates represent an industrywide effort to meet viewers where they are — which is now often looking for entertainment in shorter bursts and consumable in moments or minutes instead of portions of hours or longer.

The AI industry likes to abuse the word “open.” It appears in product releases, research papers, policy debates, and investor presentations. A company publishes model files to Hugging Face, developers run them on their own GPUs, and the release is quickly described as an “open source model.” Not necessarily. It may only be open-weight.

The difference is more than a technicality. It determines whether you can merely deploy a completed neural network or whether you can meaningfully inspect, reproduce, alter, and redistribute the system that produced it. A genuinely open source system should grant the freedom to do all of the above.

Weights are the learned numerical parameters created by training. Together with the model architecture and inference code, they allow a large language model (LLM) to function. You can download an open-weight model, self-host it, fine-tune it on internal documents, and avoid routing prompts through a proprietary API.

Open weights are publicly available. They matter because running them locally can offer greater control over data, privacy, costs, supplier API changes, and vendor lock-in. They’ve also helped build a large ecosystem of local-model runtimes, inference providers, fine-tuning tools, and specialized downstream models.

The Open Source Initiative (OSI), steward of the Open Source Definition (OSD), makes the distinction directly: “Open Weights refer to the final weights and biases of a trained neural network.” Those values determine how a model interprets prompts and produces outputs. Releasing them can let others fine-tune, adapt, or deploy the model. But the OSI adds that weights alone expose only “a fraction of the information required for full accountability.”

As James Landay, director of the Stanford Institute for Human-Centered AI (HAI), explained: “Open weights are progress. You can download the model, run it on your own machine, keep it out of someone else’s data pipeline. But you still can’t see how the thing was built, what it was trained on, or why it behaves the way it does. That’s not an open model. That’s open distribution.”

Without the training data or sufficiently detailed documentation, outsiders cannot determine which sources were used, what copyrighted or private material may have been included, how data was selected or removed, which languages and communities were underrepresented, whether benchmark data leaked into training, or what alignment and safety methods affected the model after pretraining.

Landay continued: “There’s a wide gap between open-weight AI and open source AI.” He contends that unless developers disclose training data or provide a “thoroughly documented, auditable account of it,” you can’t test, reproduce, or challenge the work in the fullest sense.

The OSI has its own definition of open source AI: the Open Source AI Definition (OSAID 1.0). It requires model parameters, including weights, to be made available under OSI-approved terms, but does not prescribe a specific legal mechanism for doing so.

Luca Antiga, CTO of Lightning AI and a prominent PyTorch contributor, has argued that OSAID’s treatment of weights leaves “a gaping hole that will make licenses less effective in determining whether OSI-licensed AI systems can be adopted in real-world contexts.”

Other open source figures have also criticized OSAID. Bruce Perens, author of the original OSD, denounced the OSAID in 2024. He later declared: “It’s not Open Source! … It’s unfortunate that the Open Source Initiative itself is now involved in Openwashing.”

He’s far from alone. Bradley Kuhn, policy fellow and hacker-in-residence at the Software Freedom Conservancy (SFC), and Red Hat Senior Commercial Counsel Richard Fontana have called for OSAID to be repealed, arguing: “The OSI acted too quickly to impose an overly ambitious policy compromise on the community. OSAID undeniably created a rift in the FOSS community; that rift seriously damaged the OSI’s reputation, authority, and influence. Meanwhile, OSAID shows no signs of having any positive policy influence on machine learning practitioners, the FOSS community, or regulators.”

OSI acknowledged when OSAID 1.0 was released in October 2024 that the definition would continue to evolve. Critics contend that its central shortcomings have yet to be resolved.

That said, the Linux Foundation’s Mike Dolan submitted the Open Model, Data, and Weights (OpenMDW) license to the OSI. The license has been around since 2025 and lists contributors from Amazon, Meta, IBM, Microsoft, and Nvidia, giving it substantial industry backing.

Conventional open source revolves around source code. LLMs are a different kettle of fish: they combine code, architecture, and numerical weights derived from training datasets that may be proprietary, copyrighted, or undisclosed. OpenMDW’s answer is to define separate terms for a model’s architecture, training data, and weights, bringing the components supplied by a licensor under one agreement.

It sounds reasonable to me, but the submission has encountered objections on OSI’s license review mailing list. As Stefano Maffulli, OSI’s former executive director, who led the organization while OSAID was being formulated, said: “I continue getting the impression that the OpenMDW review is tainted by an ideological bias: Because we don’t like big tech and AI now is big tech, then we don’t like AI; therefore, we’ll do anything to block it.”

It’s too late to bury our heads in the sand. As Stanford’s Landay put it: “Open weights answer ‘Can I run this?’ Open source answers ‘Can I trust this, improve it, and build the next thing on top of it?’ Right now almost everyone – American labs and Chinese labs alike – is answering the first question but nowhere close to the second.”

We need both. Whether OSI adopts OpenMDW is an open question. Still, OpenMDW and its supporters are at least trying to establish licensing terms that cover code, data, and weights together. Unless someone succeeds, “open AI” risks becoming an oxymoron – or merely another hollow tech marketing term. ®