It’s been a long road: After a decade of pushing it off, the federal government will be enforcing a regulation that requires truckers to submit their medical certification electronically, and not on paper.

The FMCSA (Federal Motor Carrier Safety Administration) first issued a shift towards a fully electronic Medical Examiner’s Certification back in 2015, but enforcement was pushed back several times due to IT concerns.

Now, the final national deadline for the upgrade is almost here: Truckers will officially need to file electronically on June 23, 2025.

What’s Changing?

Most commercial US truckers — and anyone who manages them — have already been submitting their DVIRs electronically for years, following the ELD mandate, which first went into full enforcement back in 2019.

By ensuring that all truckers track their hours of service with approved electronic logging devices (ELDs), the government gets more useful, searchable data. And since truckers also need to retain Records of Duty Status (RODS) data and backup data from the previous six months, there’s a wealth of information should anything need to be verified or double-checked.

That electronic data hasn’t included the Medical Examiner’s Certificate (MEC), a paper document that truckers must have on-hand during a routine stop, since it proves to state licensing agencies that they meet the physical qualifications for operating a commercial vehicle.

On June 23 of this year, that finally changes. According to CDLLife, “the compliance date was initially set for June 22nd, 2018. It was pushed to June 22nd, 2021, and then to June 23rd, 2025, due to IT system issues.”

How Does This Impact You?

This is a rare new regulation that won’t give truckers anything extra to do. Once the law switches over from paper certificates (which truckers must have on hand) to electronic ones, truckers won’t have to offer up the certificates in order to prove they deserve their license.

Instead, Certified Medical Examiners will be the ones required to submit any commercial vehicle driver medical exam results directly to the FMCSA, as well as the relevant state licensing agency. That whole process takes place through the National Registry of Certified Medical Examiners — leaving the commercial truckers and motor carriers themselves free to do their job.

However, you should still make sure you keep your paper records on hand for at least a month or two after the regulation switches over in late June. You won’t want to be caught off guard in the event of an unexpected IT failure, since the online system will still be brand-new. There’s no reason to put your commercial driver’s license (CDL) in danger when you’re perfectly qualified to drive a truck.

Some States Are Already Enforcing the Regulation

Depending on your state, these nationwide changes might have already arrived: State governments in Maryland, Minnesota, Utah, and Texas are already accepting electronic medical certificates from truckers.

From March 23rd to June 23rd of 2025, Texas will accept both paper and electronic certification to further ease the transition.

It’s another example of how modern trucking is increasingly online, even if the shift itself took many years.

Foreign-imported cars and light-duty trucks may soon start increasing their US price tags by over $7,000.

Why? The Trump administration’s new 25% tariff on all automobile imports. It comes right on the heels of a 20% steel and aluminum tariff, as well as a China-focused tariff, and it will be followed in early May with another tariff aimed at auto parts.

To learn about how these taxes will affect you, we checked in across a wide range of impacted industries and countries. Read on to hear from Canada’s prime minister – he’s not happy – as well as see which automakers have it worse. Trump ear-bender Elon Musk should be ecstatic: Tesla, Inc. will likely do the best, since about 80% of its cars are domestic.

How Do the New Tariffs Work?

The hubhub all stems from an executive order President Donald Trump signed on Wednesday. Starting just next week on April 3, the order calls for a 25% tariff on all automobile imports to the United States.

What about vehicles that were partially made in the US? The White House has an answer: It provided a fact sheet about the tariffs that clarifies this caveat. Any tariffs applied to the automobiles imported under the Canada-U.S.-Mexico Agreement (CUSMA) will only count towards the value of elements that were not made in the United States.

In other words, Canadian and Mexican automakers will still suffer the full brunt of the tariffs. And, while the Trump administration has announced and then paused other tariffs for Canada in the last month, Trump claims that this one is “permanent.”

Canadian Prime Minister Mark Carney called the duties a “direct attack” on Canadian autoworkers that’s “entirely inconsistent” with CUSMA and “the long history of relations in the auto sector right back through the Auto Pact,” a 1965 trade agreement between Canada and the US.

Automotive Logistics Will Have a Rough Year

What kind of impact will these tariffs have? Logistics news source The Lodestar checked in with investment bank Jefferies, which calls the new duties the “most heavy-handed and simplest option.”

Jefferies estimates that, if these tariffs were around this time last year:

• They would have affected eight million vehicles 
• They would have increased the price by $7,600 per imported car
• Ford will be less impacted than GM and STLA.

Transport Topics notes another automaker that will come out a winner due to the new tariffs: Tesla, Inc., which sees around 80% of its vehicles built domestically.

There’s one bit of good news, however: Heavy-duty trucks appear to be exempt from the tariffs, at least so far.

However, a second tranche of tariffs aimed at auto parts is expected on May 3, and will likely impact trucks, due to the entire US auto industry’s reliance on global supply chains. Provided these tariffs result in supply chains being rebuilt within the US, it’s the consumers who will likely fare the worst, since costs will rise and US automakers may deliver a lower quality product without global competition.

…But Cargo Frontloading Is Boosting Short-Term Profits

All the uncertainty from the new tariffs has resulted in a short-term benefit for the shipping business. However, the Port of Los Angeles just saw its second-best February on record, according to Executive Director Gene Seroka. Seroka was speaking at a March 19 press briefing, so it wasn’t in response to the new 25% tariffs. However, the new tariffs may lead to another short-term shot in the arm for the same reason: Cargo frontloading.

In logistics, the term “frontloading” refers to increasing efforts or costs at the beginning of a project period, a practice that reduces uncertainty and ensures everyone stays on track. Since the new tariffs – like the 25% steel and aluminum tariffs for all countries that were imposed earlier in March – are making shippers uncertain about how costs may fluctuate down the road, they’re all increasing shipments now in order to hedge against any upcoming changes.

In the case of LA’s port, Seroka expects consistent volume growth across the first and second quarters, and he predicts a 10% drop in volume in the second half of the year.

The Trump administration added 20%  tariffs on imports from China earlier this month, which reduced cargo shipments from China to LA down to 43% of the port’s traffic, down from 57% before the change.

In other news, US rail traffic increased more than 5% year-over-year for the fourth week in a row.

US and Canadian Steelworkers Face Layoffs

According to a new Freight Waves article, the US-based Cleveland-Cliffs Inc. is laying off more than 1,200 steel and mine workers in Michigan and Minnesota. The company cites market conditions as well as “weak automotive production” in the U.S. Last month, Canadian steelmaker MPG Canada laid off 140 positions, citing the US tariffs.

Part of the problem is the low demand for US cars: Within months, one forecaster has dropped from predicting 16.3 million vehicles sold in the US in 2025 to just 15.6 million.

Once the tariffs have been in place for a while, we may see a reversal of the US layoffs – but given the uncertainty surrounding the amount and size of all the tariffs that might be issued, it’s impossible to say when that might be.

Meanwhile, Canadian steelworkers aren’t optimistic about their job security. Canada-based Algoma Steel has laid off about 20 people, citing customer loss due to tariffs.

A Canada-focused reader survey from TruckNews.com puts a face to the insecurity that many workers in the country feel right now. One of them writes: “If these tariffs continue, I will probably lose my job in three months!” Another respondent says they “have been hauling organic grain to the same WA-state company for over 30 years. With the 25% tariff, they are looking to buy from within the U.S. and Argentina. Last load delivered March 4. Unless things change, I have lost my job.”

Will Canada and Others Counter the Tariffs?

We’re still waiting to see how any of this shakes out, but there’s a chance that Canada, Mexico, or the EU may work to counter the range of freshly imposed tariffs with their own US-specific tariffs. If so, Trump promised to continue escalating, according to a Truth Social post on Thursday.

“If the European Union works with Canada in order to do economic harm to the USA, large scale Tariffs, far larger than currently planned, will be placed on them both in order to protect the best friend that each of those two countries has ever had!” -Trump

The tax increases seem set to keep rising for the near future.

While you wait to see what ultimately happens, we recommend stocking up on toilet paper: Canadian softwood lumber tariffs recently doubled, reaching 27%, and US toilet paper prices are expected to soar as a result.

For entrepreneurs, small business owners, and creators, having a professional, visually compelling website is no longer a luxury, it’s a necessity. In today’s fast-moving digital economy, a website serves as the front door to a business, shaping first impressions and influencing customer trust.

However, the process of designing a website from scratch can be time-consuming, expensive, and technically complex especially for those without coding or design skills. Many business owners either spend countless hours trying to customize templates or hire expensive developers to bring their vision to life.

Enter Squarespace’s Design Intelligence, a groundbreaking AI-powered website builder that is removing barriers to professional web design. With its latest innovation, Blueprint AI, businesses can create stunning, personalized websites in minutes and without the steep learning curve.

Whether you’re a freelancer, coach, startup, or eCommerce entrepreneur, Squarespace combines AI-driven design tools with business management features, providing an all-in-one growth platform.

What is Design Intelligence?

At its core, Design Intelligence is an AI-powered tool that automates website creation. Instead of users spending hours adjusting layouts and tweaking code, Blueprint AI streamlines the process—just like an interior designer transforming a blank space into a stylish, functional office.

How It Works:

1. Smart AI-Powered Personalization: Users enter their industry, goals, and style preferences.
2. Automated Layouts: AI generates a professional, brand-aligned website instantly.
3. One-Click Customization: Modify fonts, colors, and page structures effortlessly.
4. Optimized for All Devices: Responsive design ensures a flawless mobile experience.
5. Time-Saving Automation: AI manages the design work, allowing business owners to focus on their content and growth.

Case Study: Sarah, a freelance graphic designer, needed a portfolio website but had no web development experience. Using Blueprint AI, she created a sleek, fully functional site in under 30 minutes—without hiring a developer.

For business owners like Sarah, Squarespace’s AI-driven design takes the stress out of website building, making it fast, intuitive, and cost-effective.

Check out Squarespace for yourself and use the code TC10 for a 10% discount.

Beyond Website Building: Squarespace’s Business Tools

A website is only the first step in running a successful business. Squarespace goes beyond design by offering integrated business tools that help entrepreneurs manage payments, sell digital products, automate scheduling, and streamline invoicing—all from one platform.

Here’s how Squarespace empowers businesses to operate more efficiently:

1. Squarespace Payments – Secure, Seamless Transactions

For many businesses, handling online transactions can be a challenge. Squarespace Payments offers a built-in, secure payment system that removes the need for third-party integrations.

How It Helps Businesses:

• Accepts major payment methods, including credit cards, Apple Pay, and PayPal.
• Provides built-in fraud protection to safeguard customer transactions.
• Eliminates third-party processing fees by keeping everything within Squarespace’s ecosystem.
• Enables businesses to sell products, services, and subscriptions directly on their site.
• Compared to other website builders that require separate payment integrations, Squarespace simplifies the entire process, ensuring a frictionless checkout experience.

2. Courses – Turn Your Expertise Into Income

The e-learning industry is booming, and Squarespace’s Courses feature enables creators to monetize their knowledge through online education.

Key Benefits:

• Easy course setup: Upload videos, PDFs, and lesson plans in minutes.
• Built-in payments: Charge one-time fees, subscriptions, or tiered pricing.
• Seamless student experience: User-friendly course navigation with progress tracking.

Example: A yoga instructor used Squarespace Courses to launch an online training program, generating passive income while expanding her brand reach.

Unlike platforms that require multiple integrations, Squarespace keeps everything in one place, making course creation accessible to non-technical users.

3. Client Invoicing – Manage Clients and Payments in One Place

For freelancers and service-based businesses, handling invoices manually can be time-consuming. Squarespace’s Client Invoicing feature automates this process, helping users get paid faster.

Why It’s Essential:

• Customizable invoices – Personalize branding and payment terms.
• Automated payment reminders – Reduce delays and improve cash flow.
• Seamless integration with Squarespace Payments – Clients can pay directly from their invoice.
• Compared to using external invoicing tools, Squarespace consolidates billing and payment processing, saving businesses valuable time.

4. Acuity Scheduling – Smart Appointment Management

For businesses that rely on appointments and bookings, Acuity Scheduling automates the entire process—reducing no-shows and administrative workload.

How It Benefits Businesses:

• Clients can book online 24/7, eliminating back-and-forth emails.
• Accept deposits or full payments upfront to secure bookings.
• Automated reminders and calendar syncing prevent scheduling conflicts.

For example: A life coach integrated Acuity Scheduling with her Squarespace site, allowing clients to book sessions effortlessly while she focused on growing her business.

Compared to standalone scheduling tools, Acuity’s integration streamlines workflows, helping service-based professionals stay organized.

Check out Squarespace’s business tools now and use the code TC10 for a 10% discount.

Why Squarespace is the Go-To Platform for Business Growth

While many website builders focus solely on design, Squarespace takes a comprehensive approach, providing AI-powered customization, built-in payments, digital products, and scheduling—all in one platform.

What Sets Squarespace Apart?

• AI-powered website creation in minutes
• No-code design customization for small businesses
• Integrated payment processing for seamless transactions
• Online course creation tools for digital entrepreneurs
• Smart scheduling features for service-based businesses

Unlike competitors that require multiple plugins or third-party tools, Squarespace offers an all-in-one ecosystem, allowing entrepreneurs to scale without complexity.

Think of Squarespace as your business’s digital headquarters combining AI-driven design with the essential tools needed to grow and succeed.

Take the Next Step with Squarespace

With Blueprint AI, Design Intelligence, and integrated business tools, Squarespace is helping entrepreneurs, freelancers, and small businesses build powerful online brands without the technical hassle.

Ready to transform your business with AI-powered website design?

Discover how Squarespace’s AI-powered Design Intelligence can transform your business today.  Use code TC10 to get a 10% discount.

Microsoft is the latest company to equip its AI chatbot, Copilot, with deep research tools, allowing it to use reasoning and analysis to provide complex answers and complete tasks. It is also able to connect with third-party data connectors and derive insights across a company’s tools.

This adds to the growing wave of chatbots with advanced thinking capabilities. Most recently, Google introduced reasoning AI models to Gemini, and xAI’s Grok 3 model includes bots trained with “reinforcement learning” that helps it use a chain-of-thought reasoning process.

These changes come at a time where there are increasing concerns about the reliability of information provided by AI chatbots. AI blunders happen often, and where chatbots are now integrated into businesses, misinformation could pose a large risk.

What Do New Copilot Tools Consist Of?

The updated Microsoft 365 Copilot includes a ‘Researcher’ and ‘Analyst’ agent to provide a deeper reasoning ability and thorough research skills.

The ‘Researcher’ agent merges OpenAI’s deep research model with Microsoft’s own “advanced orchestration” and “deep search capabilities”. It also integrates with the other programs in a user’s toolbox, such as Salesforce, creating in-depth reports and strategies with a broad range of knowledge.

The ‘Analyst’ model is built on OpenAI’s o3-mini reasoning model, and focuses on advanced data analysis to tackle complex queries and provide refined answers. The agent knows programming language Python, which the user is able to read and inspect for themselves as the agent is coding.

Both tools will start to be rolled out in April for customers on the new Frontier program, which will from then on be the place where experimental Copilot features first launch.

Copilot Keeping Up With The Competition

AI companies are in a constant race of improving and defining their chatbots within the crowd, particularly as AI continues to establish itself more and more within the workplace.

Copilot is the latest chatbot to receive deep research and analysis abilities. Not so long ago, Google unveiled a new ‘reasoning’ update to its chatbot Gemini, allowing the bot to pause and ‘think’ before giving answers. ChatGPT also has a Deep Research feature, in the form of an agent OpenAI claims reports “at the level of a research analyst”.

The ability to ‘think deeply’ and refine this thinking as new information pops up, could be crucial in the development of AI. It is most certainly an important step in creating AI bots that are able to problem solve and combine knowledge as humans do.

Risks Remain When Using AI for Research

However, as exciting as deeper reasoning agents are, there are still concerns about how accurate the information AI bots provide, particularly if they are being used in professional settings.

The fear is that AI bots can “hallucinate” or make information up. There have been cases of lawyers being sactioned for submitting fake case citations given by ChatGPT, or Microsoft chatbot MyCity telling business owners it is legal to take a cut of their workers’ tips.

Not only do chatbots need the ability to think deeply about problems, but also take the sources they are given and decide whether it is reliable or accurate enough to mention in its research. At the moment at least, this seems like a level of intuition reserved only for humans.

The Atlantic’s Editor-in-Chief was added to a government group chat on messaging app Signal, which is raising security concerns about where sensitive military conversations take place.

Signal is a privacy-focused messaging app, which helps its users avoid being tracked. The President of the Signal Foundation, Meredith Whittaker, has described the app as the “gold standard of private comms” and outlined Signal’s central security principle, which includes end-to-end encryption. The app also boasts impressive security features such as profile picture blurring and disappearing messages. Despite these measures, there are still concerns about how appropriate Signal is for sensitive government conversations.

However, the app remains a solid choice for businesses looking for added privacy and security, and below, we’ve outlined the best ways you can bolster your defenses while using Signal.

What is Signal and Who Owns It?

Signal was the app used by US government officials to discuss military plans, sparking concerns around its suitability for highly sensitive conversations. Signal is an open-source messaging service used by an estimated 40-70 million users per month.

It is owned by the Signal Foundation, a US non-profit that relies on donations and grants for funding rather than advertising, putting user privacy at the center of its business model. Its co-founders are Moxie Marlinspike and Brian Acton, the latter who founded WhatsApp.

The company’s focus on privacy and security has earned it a great reputation among cybersecurity experts and journalists. Its end-to-end encryption allows only the sender and receiver to view messages, so the only information the company has on its users is the date an account was added and the date said account was last used.

Growing Concerns Over Security

While Signal’s security measures are strong, it is still not considered an appropriate tool to house sensitive military conversations, such as the one Atlantic Editor-in-Chief Jeffrey Goldberg found himself in the middle of this week.

Added to a group chat, Goldberg was subject to US plans to bomb Houthi targets across Yemen, including weapons packages and targets. He also read discussions about perceptions of European “free-loading” and America’s role in “bailing out” Europe.

There are certainly concerns about government security in these cases. Especially since certain features of Signal, such as disappearing messages, could be in violation of federal law that requires official government records to be kept.

Tips for Using Signal in Your Business

Signal is otherwise a very secure messaging app. Beyond its end-to-end encryption and disappearing message feature, the app is open-source, meaning it can be continually checked by experts to prevent hacks.

However, if you are concerned about security when using the app, here are our tips for using Signal in your business:

• Location: Phones can be less secure than computers because of how often we use them. If possible, avoid opening the Signal app on public transport, or keep it restricted to an antivirus-protected computer.
• Features: Take advantage of features such as disappearing messages, profile picture image blurring and an incognito keyboard.
• Link previews: Disable link previews through Settings > Privacy in order to reduce the risk of tracking or phishing scams.
• Verification: Verify safety numbers before sharing sensitive information, allowing you to make sure you are communicating with the correct person.
• Device: Consider using a secondary phone number when you sign up.
• Security: Always choose a strong PIN and consider using biometrics to access the app.

Volvo is opening the books for its new Electric HGV, which it claims will change the long-haul trucking world forever.

The Volvo FH Aero Electric has been teased since last year but now the car manufacturer says that potential customers can sign up to order and that the official launch will be Q4 of this year.

As companies consider transitioning their fleets to electric, this truck has 600km range and the ability to recharge in just 40 minutes.

Significant Leap

Volvo is describing the arrival of its new flagship electric model as “a real breakthrough in zero-emission transport.” Its e-axle design means that there is more room for batteries, hence the 780 kWh installed capacity.

In a news piece on Sustainable Times, Roger Alm, President of Volvo Trucks, said: “Now, transport companies can operate really long distances with electric trucks without having to compromise on productivity. The superfast charging and high payload capacity make this a very competitive solution.”

This truck, Volvo says, delivers too on capacity. It can haul up to 44 tons in the UK and US, and 46 tons in the Republic of Ireland when paired with a standard tri-axle semi-trailer.

The company already sells its VNR Electric models stateside, and these have an operating range of 275 miles. The page is now live for US buyers to “build their truck” for the new model.

Competition on the Road

Volvo claims it has so far delivered more than 3,800 electric trucks to customers in 46 countries around the world.

However, other companies are pushing hard too. This year will also see the arrival of a new Renault T-Model, which will also have a 600km range on a single charge.

“We believe that a range of 600 km on a single charge, combined with the development of public charging infrastructure networks by 2026 – in particular through our joint venture Milence – will enable us to achieve the operational parity [with diesel technology] that our customers expect.” – Emmanuel Duperray, Senior Vice President Electromobility at Renault Trucks

Not Just About Range

However, Duperray also added that his company is “…not looking to enter a race for autonomy on a single charge.” Instead, he argues that manufacturers need to think about how an electric truck is currently “more expensive than a combustion vehicle.”

“We need to rethink low-carbon logistics, in other words, reconsider transport patterns to optimize the use of transport vehicles and therefore reduce the cost per kilometer.” – Duperray

Electric trucks – especially delivering these sorts of ranges – are a huge step towards decarbonization, but companies also need to keep pushing the envelope in terms of design, cost efficiency, and infrastructure.

The law is harsh, but it is the law: so says the AI judge in a recent study from the University of Chicago Law School. The study analysed the differences between AI and human legal decision-making, and found that human judges were significantly influenced by non-legal and emotional factors, unlike OpenAI’s GPT-40.

The findings raise questions on how closely AI can mirror the judgement of humans, particularly in settings such as the legal system, where an emotional intelligence and nuance is needed. Even as models continue to become faster and stronger, this empathy gap could hinder AI’s potential.

AI Puts Its Faith in The Law While Humans Consider Emotional Elements

The original study set out to understand how federal judges and law students made legal decisions. Participants reviewed the simulated appeals of defendants involved in international war crimes. Defendants were shown to be sympathetic and unsympathetic, and the judges and students were also given different levels of relevant and irrelevant information.

When faced with a sympathetic defendant, the federal judges had their decisions swayed 65% of the time, even ruling to overturn convictions in cases where legal precedent supported it. The law students were overall able to stick to precedent 85% of the time.

In the recent University of Chicago’s Law School study, GPT’s decisions followed legal precedent more consistently. In fact, GPT stuck to the law in over 90% of cases, even when faced with more sympathetic defendants.

In both studies, the judges were also given hints that a defendant’s conviction could be legally flawed. While GPT remained law-abiding in these cases, the human judges’ decisions remained dependent on how sympathetically the defendant was portrayed to them, despite any other background knowledge.

AI Unable to Tap Into Human Sensibilities

Each case reviewed in the study was designed to be ambiguous in order to evaluate the level of decision-making. However, this did not seem to affect GPT in its rulings either. 

In an attempt to bridge the gap between the AI and human judges, the researchers prompted GPT to think beyond simply following the letter of the law. This involved tweaking the information fed into the bot with the aim of replicating the responses of the human judges. 

The researchers trained the AI on legal realism theory, which suggests that judges use factors outside of the law to determine a ruling, such as emotional and social contexts. They also urged GPT to think more broadly about justice beyond the law, and even explicitly told it to consider sympathetic defendants in a different light. 

Despite these adjustments, the AI judges remained focused on legal precedent, and did not consider the emotional aspects of a case. Researchers concluded this to be an example of legal formalism, where rules and precedents are followed with no regard to personal feelings. 

Use of AI Judges in Courtrooms is a Possibility

Discussions around how AI will continue to establish itself within society’s systems are not going anywhere, particularly as platforms such as ChatGPT and Gemini continue to advance and integrate themselves into our daily lives.

Some experts suspect that AI will play a significant legal role in the future, with the National Center of State Courts having released some guidance last year on using AI in the courtroom. Although, in cases where AI has been used for producing court documents, there have been reports of it inventing fake case law, meaning there may be a long way to go before the legal system puts its trust in these bots.

We are yet to see AI operating in the same way as human judges do. However, a key component of an effective legal system is being able to take the law and apply it based on different circumstances. That being said, AI may be missing the nuanced outlook of human judges, and it is difficult to know whether they will ever be on that same emotional wavelength.

Malaysia plans to tighten regulations on semiconductors as it comes under intense US scrutiny. Allegedly, it has been detected that semiconductor chips that are integral to AI development have passed through Malaysia and ended up in China. The allegation has been made that the chips in question are made by the US-based company Nvidia, which is banned from selling technology to China.

In response, the US is demanding that Malaysia tracks the movement of Nvidia chips that enter the country. Concurrently, it is investigating whether or not DeepSeek, the popular Chinese AI tool, is developing software with US chips, which it has been prohibited from doing. China, relatedly, has launched a probe into Nvidia.

Since President Trump’s return to the White House, the US has doggedly pursued China in the unfolding AI race. While some commentators are calling for better relations between the two superpowers, especially where AI is concerned, it’s unlikely that the government would countenance closer collaboration. The two countries have enjoyed a particularly hostile relationship in the last decade or so.

US Pressures Malaysia to Investigate Flow of AI Chips

The US has told Malaysia to tighten regulations surrounding the export of US-made semiconductors. The Nvidia chips, which form a key building block of AI technology, have reportedly been making their way to China via Malaysia.

There are concerns that when the chips arrive in Malaysia, some of them are rerouted to pass into China. This could potentially explain the meteoric rise of Chinese AI platform DeepSeek.

Malaysian Trade Minister Zafrul Aziz told The Financial Times: “[They’re] asking us to make sure that we monitor every shipment that comes to Malaysia when it involves Nvidia chips. They want us to make sure that servers end up in the data centers that they’re supposed to and not suddenly move to another ship.”

Paranoia Abounds Over China AI Successes

As it puts pressure on Malaysia, the US is also conducting an investigation into whether or not China’s latest AI developments have been built on US semiconductor technology. Nvidia, which powers much of the AI development around the world, is not allowed to sell its advanced chips to China.

This has not deterred the superpower. In recent months, China has rolled out two genuinely game-changing AI propositions – DeepSeek and, more recently, Manus AI. The groundbreaking impact of these models has raised suspicions in the US. It is now thought that China has continued to build its latest innovations with US-made semiconductors.

In December 2024, meanwhile, China’s antimonopoly regulator launched a probe into Nvidia. It is currently investigating whether or not the US chipmaker has violated the country’s antitrust laws.

US-China Hostilities Unlikely to Abate

The race for AI supremacy is a hotly-contested battle between the US and China, with no clear winner at this point in time. Recent innovations from the eastern superpower, including DeepSeek and Manus, have attracted global attention. At the same time, the US continues to pour vast sums into meeting its own ambitions, such as the $500 billion investment into Project Stargate.

Meanwhile, Stephen Orlins, who heads up the National Committee on United States-China Relations (NCUSCR), has called on the two superpowers to set aside their differences and cooperate on AI. With so much bad blood between the two countries in the last decade or so, it would require a herculean diplomatic effort to bring about the kind of mutually-beneficial relationship that Orlins is proposing.

And with President Trump back in the White House, the US is pursuing a hostile trade strategy where China is concerned. It’s very unlikely that the current administration would entertain such a relationship.

Do androids dream of electric sheep? Very soon we might be able to ask them directly. That’s according to Nvidia CEO Jensen Huang, who believes that the rollout of humanoid robots is less than five years away. The comments came at the annual GTC developer conference, during which the chipmaker boss was asked by reporters when AI could be considered ubiquitous.

Huang also unveiled a new innovation from Nvidia, known as Isaac GR00T N1, an open-source foundation model that promises to accelerate the burgeoning humanoid robot race. Across the world, tech companies are doubling their efforts to win the space, with a number of landmark developments taking place this week alone.

The return of President Donald Trump to the White House, and subsequent tariff wars with Europe and China, has undoubtedly given the humanoid robotics movement a shot in the arm.

Humanoid Robots Will Be Ubiquitous in a Few Years, Says Nvidia Boss

The widespread use of humanoid robots is less than five years away, says the Nvidia CEO. According to Huang, who heads up the global chipmaker, robots that resemble humans could become a fixture in everyday life in just a few years’ time.

The remarks came after a keynote at the annual Nvidia developer conference, during which Huang was asked by reporters which signs would indicate that AI had become ubiquitous. Said the CEO: “When, literally, humanoid robots are wandering round, which is not five years away. This is five-years-away problem, this is a few-years-away problem.”

Heralded as the “Super Bowl of AI,” GTC 2025 brings together thought leaders and industry players from the world of AI for a four-day long conference. During a two-hour long keynote earlier in the day, Huang reflected on the “extraordinary progress” that AI has made in recent years.

Chipmaker Outlines Vision for Robot Future

As part of his address, Huang introduced Isaac GR00T N1, an open-source foundation model for the development of humanoid robots. In order to develop simulated training data for robots, this model will be paired with Cosmos AI, another Nvidia innovation.

Through Isaac GR00T N1 and his remarks, Huang appeared to set out a vision for the near-future in which humanoid robots are an intrinsic part of the human workforce. Elaborating on his comments, he stated: “I think it ought to go to factories first. And the reason for that is because the domain is much more guard-railed, and the use case is much more specific.”

He continued: “The value of it is very, very easy to determine. The going rate for renting a robot is probably $100,000 and I think it’s pretty good economics.” The news is unlikely to be welcomed in some quarters, with fears mounting that AI will begin to take over more and more jobs in the coming years.

Trump’s Return Sees Race Heat Up

Huang’s comments don’t exist in a vacuum. Across the world, the race to develop advanced humanoid robots is progressing at breakneck speed. This week, European robotics company Neura Robotics revealed that it was to launch the so-called “best robot on the market” as early as June, known as 4NE-1.

Meanwhile, the newest iteration of Boston Dynamics’s Atlas robot was filmed breakdancing in a video posted to YouTube on March 19. And Apollo, a human-like machine created by Apptronik, this week assembled an engine part in a live demonstration of its prowess.

Since Trump returned to the White House, tech activity has reached fever pitch in the US, with companies across the space investing heavily in innovation and development. Against a backdrop of unfolding international trade wars, there’s an extra incentive for this increase in activity – putting US tech squarely on the map.

Baidu is going after its fellow Chinese rival, DeepSeek, launching a reasoning model it claims is as good as DeepSeek R1 but far cheaper.

The new Baidu reasoning model, Ernie X1, was also accompanied by the unveiling of Ernie 4.5, the latest version of the company’s foundational model.

The news comes just days after Alibaba made its bid for a top AI spot with a new AI assistant built on its Qwen LLM; and the launch of Manus AI, the world’s first fully autonomous AI agent.

Can Baidu Contend With DeepSeek?

It seems that the AI launches are coming thick and fast from China in a trend that will be alarming AI ventures in the US. Baidu’s latest launches take aim squarely at DeepSeek’s much lauded – and much villainized – R1 chatbot.

Reuters shared a Baidu statement in which the company claimed: “Ernie X1 delivers performance on par with DeepSeek R1 at only half the price.”

It adds that X1 has “stronger understanding, planning, reflection, and evolution capabilities.” It also makes the claim that the reasoning model is its first to use tools autonomously.

However, for Western users, the key advantage is that the Baidu models are currently not in the crosshairs of any governments. DeepSeek, in contrast, has already been banned in several countries, including Italy, over data usage concerns.

Playing Catchup

Despite the controversy, DeepSeek still hit number one in both the Apple App Store and the Google Play Store in January.

Baidu is facing strong competition from Alibaba as well, whose stocks surged upon the launch of its QwQ-32B, which is free to download and use, even commercially. However, it was DeepSeek and OpenAI that Alibaba waved its gauntlet at and not Baidu, giving a strong sense of where the ecommerce giant sees its greatest competition.

Baidu, however, is pushing hard. It made grand claims for its Ernie bot in April last year, saying that it had exceeded 200 million users. The company also relaunched its AI development competition, offering a prize of 50 million yuan ($7 million). There are rumors that an Ernie 5 model is imminent.

There Can Be Only One (Percent)

Baidu’s CEO seems to be pretty confident that his company will manage to sustain, despite the huge competition in the AI battle for supremacy; and that’s not just from rival Chinese ventures.

In October, Robin Li predicted the demise of 99% of AI companies. He said:

“Probably one percent of the companies will stand out and become huge and will create a lot of value or will create tremendous value for the people, for the society. And I think we are just going through this kind of process.”

He explained that the main issue with chatbots – accuracy – will be solved in the next 18 to 20 months. However, the sheer volume of companies is not sustainable. It seems, though, that he is self-assured that Ernie will be one of the survivors.

It’s a rare moment, but Elon Musk and Sam Altman have actually agreed on something: The two tech CEOs have decided that they both would like to fast track the trial over OpenAI’s shift to becoming a for-profit venture.

Musk, who co-founded OpenAI with Altman, is suing, arguing that the conversion will create a monopoly to stamp out competitors.

He is perhaps particularly referring to his own AI project, xAI. Musk has also made a bid to buy OpenAI, which was rejected with a polite “no thank you” from Altman.

What’s Been Agreed?

According to Reuters, the rival billionaires have decided that they both want the case over and done with as soon as possible.

They have jointly proposed a trial in December. The filing in US District Court for the Northern District of California adds that they have also agreed to delay a decision on whether the case will be decided by a jury or just by a judge.

“We welcome the court’s March 4 decision rejecting Elon Musk’s latest attempt to slow down OpenAI for his personal benefit,” OpenAI said in a blog post.

What Does the Case Hinge Upon?

Musk filed the lawsuit last year. Initially, his claim was that Altman and co had “manipulated him” to “Shakespearean proportions” into co-founding the AI company under the guise of it being a for-profit enterprise focused on developing humanity.

He later added the gripe that OpenAI was behaving in an anticompetitive way.

However, US District Judge Yvonne Gonzalez Rogers, threw out some of Musk’s claims, damning them as being a “stretch,” and causing “irreparable harm.” He also argued that OpenAI should not be forced to stop its bid to become a for-profit business.

He did not throw them out entirely, so a court standoff is now likely.

Why Does Altman Want the Change?

The AI pioneer says that the change of structure is necessary to ensure that his company can keep innovating. This is because, he argues, developing AI technology is pricey.

Reuters reports that OpenAI’s last fundraising round of $6.6 billion, and a new round of up to $40 billion under discussion with SoftBank Group are conditioned on its change of structure.

The company also has lofty goals, focused on creating Artificial General Intelligence (AGI) – a theoretical type of AI that can match or surpass human intelligence.

As the launches come thick and fast from China – Alibaba and Baidu unveiled new tools in the last week alone – there is a lot at stake for American AI ventures. OpenAI is arguing that this change of structure is what will give it the best chance of thriving in the face of fierce competition.

Google is slowly going to mothball its mobile phone assistant as it pushes uptake of its AI chatbot, Gemini.

In a move that will surprise no one, Google is going to phase out Assistant stating that it won’t be accessible “on most mobile devices” or for new download “later this year.”

Google has been pushing Gemini hard in the past few year, adding new AI functionality to Sheets and plying developers with Google Gemini Code Assist. This latest news will see Gemini as the assistant on all Google phones as AI becomes central to the company’s offerings.

What Will Happen for Google Phone Users?

Brian Marquardt, the senior director of product management for the Gemini app, gave details of the fate of Google Assistant, which was launched in 2016. He says that it was designed to “unlock a more natural way to get help from Google” but now AI will “reimagine” the experience.

He explained in a blog post: “Over the coming months, we’re upgrading more users on mobile devices from Google Assistant to Gemini; and later this year, the classic Google Assistant will no longer be accessible on most mobile devices or available for new downloads on mobile app stores.”

He does add that for phone owners whose device doesn’t meet the minimum system requirements to run Gemini, “Google Assistant functionality will not change at this time.”

More Than Mobiles

But the move will not just impact phone owners, with Marquardt stating that an AI glow-up is also coming for the “tablets, cars and devices that connect to your phone, such as headphones and watches.”

There aren’t details as yet but Gemini will also be integrated into home devices like speakers, displays, and TVs. Marquardt promised information on what this will look like “in the next few months.” He added that Google Assistant will continue to be available on these devices for the foreseeable future.

What’s Next From Gemini?

In a post comparing Assistant to Gemini, Google says that interactions will change for users as “Gemini can understand natural language.” It also says that, while “Gemini might not always get it right,” it will improve and “get faster.”

In February, Google detailed all six of the AI improvements it had made for Pixel phones. These included an extension to let Gemini control smart devices in users’ homes, and a redesigned interface.

But it’s not just Google phones, which have integrated Gemini capabilities. At the beginning of the month, iPhone users got new Gemini lock screen widgets with the latest version of the app for iPhones. The updated app also allows users to share text, images, and links directly to Gemini from any app on their iPhone.

Google is obviously hoping that Gemini will become so integrated into phone users’ experiences that it becomes a go-to, especially as AI Siri is still not where Apple wants it to be.

Three of the biggest names in tech have come together to figure out some standards for data governance as AI technology continues to evolve.

Microsoft, IBM, and Cisco are among the names who have backed a committee, which is trying to standardize data provenance protocols.

This touches upon everything from the copyright issues that have plagued AI ventures to privacy and authenticity.

What Are the Tech Giants Hoping to Achieve?

The trio have put their backing behind the OASIS Data Provenance Standards Technical Committee.

This is the brainchild of OASIS Open, a global open source and standards organization, and the Data & Trust Alliance, which describes itself as “a consortium dedicated to developing data and AI practices that create business value and earn trust.”

Their hope is that by bringing key players to the table, they can agree what they describe in a blog post as “a standardized metadata framework for tracking data origins, transformations, and compliance, helping businesses establish clearer governance practices.”

What Are the Next Steps?

This is all about transparency. The framework being worked upon will hopefully provide some guardrails for companies as they move from talking about standards to actual implementation.

As the Oasis team says in its blog post, the standards will ideally enable data producers “to deliver clear and consistent data lineage information.” This means that they will be in a better position to ensure that they are data compliant and that they are not heading into dangerous waters in terms of data privacy, security, and intellectual property rights.

Questionable AI Data Usage

Interestingly, the standards also talk about “data acquirers” having “transparency around the data they aim to acquire and a mechanism to determine whether to trust and use the data on offer.” This is particularly pertinent if companies are working with AI ventures, building their own systems on LLMs provided by others.

Microsoft works closely with OpenAI, for example, and has recently been involved in a spat with Chinese upstart, DeepSeek, which it accused of “improperly” obtaining OpenAI data to train its model.

IBM currently uses an open-source foundation for its platform, watsonx, but has written a detailed blog post about the benefits of open-source versus proprietary AI LLMs. In this, it talks specifically about the dangers of “incomplete, contradictory, or inaccurate data.” But it also nods to the issues of ensuring “training data was gathered with accountability” and that this data harvesting is “compliant with laws and regulations.”

Here, however, lies the issue, as AI frameworks are still being churned out in the USat quite a clip. Worse, there has been huge friction between those arguing for safety nets and those pushing for innovation – notably OpenAI. Also notably, OpenAI is not at this table.

Community Notes for Facebook, Instagram, and Threads will start appearing for US users on March 18th.

The company’s CEO, Mark Zuckerberg, announced sweeping changes to the company’s moderation system in January. Central to this was the decision to ditch third-party moderators and instead adopt a model like the one used on X.

Meta’s chief global affairs officer, Joel Kaplan, defended the decision, arguing that “too much harmless content gets censored.” However, alarm bells were immediately rung about a potential spike in hate speech, but the main concern is that it’s clearly a cynical attempt to curry favor with the new president.

What Will Moderation Now Look Like?

The key difference is that instead of third-party moderators, the social media platforms will now be “policed” by users. These contributors have been able to apply for a role since February.

These special users will be able to provide notes that directly fact-check a post on any of these platforms. Their suggestions will then be rated by other contributors as helpful or not helpful. It is this that will then determine how the Community Note appears on screen to your average user.

Using an algorithm, the system will determine which contributors tend to have opposing views and a Note will only appear if both sides agree that it is helpful. This means that not all Notes will be published.

The social media giant adds that if a post does carry a Note, it will not be downranked.

Collaboration with X

Meta is actually using X’s open-source algorithm to evaluate Community Notes, according to CNBC.

“As X’s algorithm and program information is open source — meaning free and available for anyone to use— we can build on what X has done, learn from the researchers who have studied it, and improve the system for our own platforms.” – Meta spokesperson

However, it adds that it may look at other options as time goes by. “As our own version develops, we may explore different or adjusted algorithms to support how Community Notes are ranked and rated,” it said.

Cause for Concern

Despite Meta’s upbeat insistence that all will be fine, there are rising voices of concern. Indeed the co-chair of Meta’s own oversight board told BBC News that there were “huge problems” with what had been announced. Helle Thorning-Schmidt spoke about worries that marginalized groups could be targeted. She went as far as to state that there were “extremely dangerous times ahead” for social media users and democracy.

Holding X’s model up as an exemplar is hardly confidence building after all. Elon Musk, owner of X, has slashed his moderation team, welcomed back banned users (including Donald Trump) and the platform has seen a terrifying spike in hateful activity, including racist slurs.

Concerns about content have been so high that advertisers voted with their feet and have left the platform in their droves.

Zuckerberg and team will be watching the impact of the new model in the US. Meta has chosen to use its home country as a testing ground, and hasn’t rolled it out in any other nation. This is, perhaps, not least because the EU is already stating its concerns.

This new moderation model is going to be received very differently outside of the US. Other countries will be monitoring what happens to content stateside and will no doubt be feeding back any concerns to Meta.

OpenAI is petitioning the government to offer AI companies some relief from the hundreds of bills currently being proposed and debated.

In a 15-page document, the company has responded to the government’s call for input as the Trump administration works on an AI Action Plan.

The document asks for the federal government to basically intervene to stop the measures put in place by individual states curtailing AI’s rapid innovation. The appeal is likely to hit a chord as innovation at all cost – a hands-off approach – is an approach that OpenAI boss Sam Altman knows will appeal to the President.

What Does OpenAI Want?

The letter strikes a similar tone to Altman’s appeal for copyright laws to be relaxed for AI innovators to innovate.

His company argues that the bills that various US States are currently considering will slow down development. In exchange, reports Bloomberg, which has seen the document, AI companies would give the government voluntary access to their technology.

OpenAI is already working closely with the Trump administration. A month ago, it unveiled ChatGPT Gov, which is an AI agent built specifically for governmental use.

OpenAI CPO Kevin Weil explained that government employees can feed the agent “non-public, sensitive information” and that the AI will operate within secure government hosting environments.

Winning the AI Arms Race

Altman is appealing to Trump’s desire to win the AI arms release, which seems to have gone almost white hot with constant releases from China. DeepSeek gained the number one spot in both the Apple and Android App charts with its AI agent and Alibaba has been relentless with its releases.

In the proposal, OpenAI put its finger right on this sore point. It wrote that if China’s “developers have unfettered access to data and American companies are left without fair use access…the race for AI is effectively over.”

Trump has already rescinded his predecessor’s AI Safety law – and has promised to “unleash the potential of the American citizen,” especially, it seems, those working in AI.

State Intervention

However, individual states are looking to put legal AI frameworks in place. In California, one such bill was blocked in September by Governor Gavin Newsom but it hasn’t been taken off the table completely.

There is momentum for AI safety bills, including from within the industry itself – notably from OpenAi rival and AI wunderkind Anthropic. In fact, a letter was sent to Newsom urging him to lead on AI regulation and the signatories included employees of OpenAI, Google DeepMind, Anthropic, Meta, and xAI.

Instead, the OpenAi leadership is urging the Government to bypass States legislators and let Ai ventures just communicate directly with that the US AI Safety Institute. Chris Lehane, OpenAI’s vice president of global affairs, said in an interview: “Part of the incentive for doing that ought to be that you don’t have to go through the state stuff, which is not going to be anywhere near as good as what the federal level would be.”

This is not going to sit well with the senators currently working on bills, as it essentially take their power away. However, as the Chinese AI companies push their wares – and innovate at speed – the desire to beat them might drive Trump to consider OpenAI’s proposal whatever the implications.

The AI announcements – especially from China – seem to be coming thick and fast at the moment, with Alibaba hitting the headlines today.

Only a month ago, Chinese eCommerce giant, Alibaba, announced an AI model it claimed outperformed ChatGPT and DeepSeek. Now the company has unveiled a new version of its AI assistant app.

The new launch is powered by the company’s Qwen AI reasoning model, which the company is pushing to take on OpenAI and its homegrown (and controversial) rival, DeepSeek. This LLM is also being used by Chinese startup, Manus, for what is being claimed to be “a completely autonomous agent”.

What Has Alibaba Launched?

The company’s stocks surged upon the announcement of QwQ-32B, which is free to download and use, even commercially. Alibaba says that it is “very suitable for application scenarios with rapid response or high data security requirements.” It adds that it can be “easily deployed” to local devices “on consumer-grade hardware”.

The key update is that this AI assistant uses the Qwen LLMs as opposed to the company’s Quark reasoning model.

In a blog post (in Chinese), the company says that this latest launch “has achieved a qualitative leap in mathematics, code, and general capabilities”. Alibaba also took a swipe at DeepSeek, claiming its new AI assistant is “comparable to DeepSeek-R1”. It also claims that its latest launch “almost completely surpass[ed] OpenAI-o1-mini”.

High Ambitions for Alibaba’s App

CNBC news reports that Alibaba is planning on investing $380 billion yuan ($52.5 billion) in cloud computing and its AI infrastructure over the next three years.

Its launches have been arriving rapidly and the company is keen to emphasize that its products are as good – if not better – than rivals. It says in the blog post: “As of now, the number of derivative models of Qwen in domestic and overseas AI open source communities has exceeded 100,000, surpassing the American Llama series of models to become the world’s largest open source model group.”

Its AI ambitions are having a huge impact on the company’s bottom line. In February, CNBC reported on Alibaba’s quarterly results. The company’s Cloud Intelligence Group had seen year-on-year sales growth of 13% to 31.742 billion yuan (over $4 billion). “The AI era presents a clear and massive demand for infrastructure. We will aggressively invest in AI infrastructure,” said Alibaba CEO, Eddie Wu.

Chinese AI Dominance

However, will Alibaba be allowed to push for dominance outside of China? DeepSeek is facing deep mistrust and has already been banned in several countries over national security concerns. In particular, the fact that users’ data is stored in China is causing concern.

There is also the Trump administration’s strong backing for American AI ventures. Accusations that DeepSeek has been using OpenAI’s output to train its own models is likely to see relations worsen.

However, DeepSeek has yet to be fully banned in the US – only from some Government devices. While the focus is on DeepSeek, Alibaba is merrily launching products and doesn’t seem to be attracting the same heat…for now.

TikTok is hoping to calm the storms around child safety on its platform with new tools to control usage.

The latest features include a reminder for teens to stop using the app at night as well as updates to its parental controls.

However, will they be enough to halt the tide of concerns about children’s safety on TikTok? The company is facing scrutiny in the UK for its use of children’s personal information; and lawsuits in both France and the US. These are as well as the looming cloud over its future stateside.

Setting Boundaries on TikTok

TikTok has published a blog entitled: “New ways we’re supporting parents and helping teens build balanced digital habits”. In it, head of ops, Adam Presser, says that the updates add to “the strongest safeguards” that the company already puts in place “by default” for its teen accounts.

The key update is to the Family Pairing option, which launched five years ago. Now parents will be able to block their children from being able to access TikTok at certain times of the day. They can also set these times as a recurring schedule.

Another update gives parents more transparency. They will be able to see who their teen is following, who follows them and also if their child has blocked any accounts.

TikTok adds that it is also re-enabling the STEM feed of tailored content in all accounts – including those in which the teen has turned it off.

Calming Impact for Younger Users

For under-16s, TikTok has also announced a “wind down feature”. This kicks in after 10pm if the teen is still scrolling. It explains that this will be “a full-screen takeover with calming music to help teens relax and be mindful of the time”. The calming music might soon also be joined by meditation exercises, the company adds.

If the teen decides to ignore and keep scrolling, a second full-screen prompt will appear. TikTok says this is “harder to dismiss” but can be dismissed. As a result, some are already questioning how impactful it will be. However, TikTok says its confident of an impact. “In countries where this has already been piloted, the vast majority of teens decide to keep this reminder on”, it writes.

Too Little, Too Late?

These new announcements come just months after TikTok announced plans to block beauty filters for teenage users.

The company has launched a crackdown too on underage users, announcing that children under 13 years old will be booted off the platform. In Europe, it has teamed up with telecoms giant, Telefónica, for a new API called “Age Verification”. This uses AI “to understand how people can use age information from their phone provider to confirm their age.”

However, with lawsuits focused on the mental health impact of usage – especially how addictive the platform is for children – these latest announcements seem pretty feeble. Not least because TikTok knows exactly how addicted children can become – it has researched it.

On top of this, research seems to suggest that the company’s age verification tools are simply not working. In January, Psypost reported that 68.2% of under-13 users reported using TikTok. With statistics like this, it seems meditation exercises and calming music aren’t going to solve the problems.

If reports are to be believed, the world’s first fully autonomous AI agent is here. Developed by Chinese start-up Butterfly Effect, Manus AI is a “general AI agent” that can autonomously execute complex tasks, including analyzing data, generating reports, and ordering groceries.

Manus AI has been called “China’s second DeepSeek moment,” and promises a dizzying array of potential use cases. But how does it work? And what might it be capable of? In this guide, I’ll dive deep into the platform, unpacking everything that we know so far.

What Is Manus AI?

Manus AI is a new AI agent developed by Chinese start-up Butterfly Effect. Named after the Latin word for “hand,” Manus is more than “just another chatbot or workflow…it’s a completely autonomous agent,” according to chief scientist Yichao “Peak” Ji.

As a fully autonomous platform, it is capable of making decisions independently. While ChatGPT, Gemini, and Claude require some level of human input, Manus is able to think, plan, and execute without any assistance. In other words, Manus promises to leave many of its contemporaries in the dust.

Manus was unveiled last week, and currently it’s in private beta mode, meaning you’ll need an invitation in order to test it out for yourself. But its rollout is sure to send shockwaves through the tech world. US companies will consider it a call to arms in their bid to best China in the unfolding AI race. Critics, meanwhile, will likely despair at the prospect of a machine that has the potential to completely replace a human.

What Can Manus Do?

Broadly speaking, an AI agent is a program capable of executing tasks on behalf of a user or system. This involves decision-making, problem-solving, and interacting with different external environments. Some examples include ecommerce agents that track and provide shipping updates on items that you’ve purchased, and trading agents that buy or sell stocks based on algorithms.

Manus is a fully autonomous program, meaning that it can act totally independently. Ji has called it “the next paradigm of human-machine collaboration and potentially a glimpse into AGI.” This last part, in particular, is giving AI enthusiasts a lot of cause for excitement.

AGI, or artificial general intelligence, is the concept of software that can understand or learn anything that a human being can. It’s an ongoing field of theoretical research that has so far failed to yield any substantial results. Manus might be the closest that anyone has got.

On the official Manus website, there are several case studies allegedly showcasing what the program is capable of. Among them, you can view a detailed seven-day travel itinerary for Japan for a hypothetical Seattle couple. The itinerary includes “Proposal Opportunity” locations that are recommended by Manus. Other highlights include an Apple-inspired business card design, and a three-second bird chirping and steam sound effect. Truly, the ridiculous and the sublime.

How Does Manus Work?

Unlike some of its contemporaries, Manus has been designed as a multi-agent system, meaning that it combines several different AI models. These models specialize in a broad range of tasks, so Manus has many theoretical use cases. Two of the biggest large language models (LLMs) that it relies upon are Claude 3.5 Sonnet and Qwen, with Butterfly Effect thought to be eager to upgrade Manus to the latest version of Claude, 3.7.

The agent has its own knowledge and memory recall, so it can learn from past experiences to iteratively improve its future performance. Let’s say you instruct Manus to export some data analysis results into a PDF. The next time it’s asked to perform data analysis, it will automatically export it in the same way.

While standard AI chatbots simply regurgitate information, Manus can perform deep research, sophisticated data analysis, report generation, workflow automation, code writing and even deployment. And that’s just the tip of the iceberg.

What Does Manus Mean for the Future?

It’s very early days, but some believe that Manus has the potential to upend the AI space. Founder of The Rundown AI newsletter, Rowan Cheung, described the platform as “Deep Research + Operator + Claude Computer combined, and it’s REALLY good.”

Other commentators, however, have cautioned against getting carried away. According to assistant professor at the University of Chicago Booth School of Business, Bradford Levy: “There is plenty of evidence that Manus doesn’t get things right on the first try or gets stuck in infinite loops…there’s a good chance it won’t meet expectations.”

It remains to be seen who will be proven right. While new AI products and announcements are landing on a near-daily basis, the truth is that we won’t feel their material impact for some time. Until then, we’ll be keeping track of all the latest happenings to arise in this hugely exciting space.

Salesforce plans to invest $1 billion in Singapore over the next five years, it was revealed on Tuesday. The remarks came courtesy of CEO Marc Benioff at CONVERGE LIVE, a global meeting of business leaders, policymakers and investors, hosted in Singapore.

Benioff claims the investment is premised upon accelerating the country’s digital transformation, including the adoption of Agentforce, its AI offering. Just last week, Salesforce unveiled a series of updates to the AI agent in a bid to ward off competition from Microsoft.

Salesforce is one of many tech companies hedging its bets on AI in a big way. President Trump’s return to the White House in January has led to businesses ramping up their AI spend, with many determined to take market share off of OpenAI, which continues to be the platform of choice.

Salesforce to Invest $1 Billion in Singapore

Cloud software giant, Salesforce, plans to invest $1 billion in Singapore over the next five years, according to CEO Marc Benioff. The company has been investing in the country for almost 20 years, with this latest pledge designed to fuel the widespread adoption of Agentforce, the company’s AI platform.

In a statement, the company claims: “Agentforce delivers Singapore an opportunity to rapidly expand its labor force in many key service and public sector roles…as Agentforce adoption accelerates, it has the potential to drive significant impact across Singapore’s industries, startups, and the public sector.”

The news caps a busy month for the software company, which last week unveiled a series of upgrades to its AI model, known as Agentforce 2dx, and last month announced plans to let go of more than 1,000 employees to make room for AI. Alongside this, it also plans to invest $500 million in Saudi Arabia, and a further $500 million in Argentina.

Asia Covets Global AI Hub Status

Singapore is jostling for top-dog status in the AI world. In 2023, the country unveiled the National AI Strategy 2.0 (NAIS 2.0), which included plans to invest $1 billion in infrastructure, research, and talent development. This has since been matched by Amazon, Alphabet, Microsoft, and now Salesforce.

Elsewhere, countries across Asia are making similar plays. Malaysia recently committed $250 million over the next 10 years to acquire chip design blueprints from Arm Holdings, the UK-based semiconductor company. With this information, it hopes to meet growing demand for domestic AI and data centers.

Thailand has received substantial investment in recent years. In 2024, Microsoft revealed that it planned to build new cloud and AI infrastructure in the country. The same year, Google announced $1 billion of investment to build a data center and cloud region there, while Amazon Web Services (AWS) recently committed $5 billion over the next 15 years.

Tech Companies Betting Big on AI

In recent years, companies across the tech sector have raced to invest in AI. With President Trump back in the Oval Office, momentum has reached fever pitch. Soon after taking office, the returning Republican dismantled many of the safeguards that the previous administration had put in place. He also announced Project Stargate, a new company with designs to be the “largest AI infrastructure project in history.”

Several big tech players have taken this as their cue to invest heavily in AI. Last week, reports emerged that search engine giant, Google, plans to show AI Overviews for even more search queries, despite an underwhelming response to the new AI feature. Meanwhile, DBS Bank recently revealed that thousands of employees will make way for new AI deployments.

Salesforce hopes that its own forays will be successful. In the words of CEO Benioff: “We don’t just do sales and marketing here. We also do extremely advanced artificial intelligence development.” Whether or not its recent investments are enough to cement its status as a frontrunner in the race, however, remains to be seen.

AI darling Deepseek won’t be looking for outside investors any time soon, according to reports about comments from its own founder.

Liang Wenfeng is not seeking investors for his generative AI company, and is particularly against government investors.

That makes sense: The China-based Deepseek has faced plenty of criticism already, and has even been banned in multiple countries.

Liang Wants to Avoid Government Influence

The new scoop comes from the Wall Street Journal, which cites “people familiar with the matter” to deliver the news that “Liang Wenfeng has told associates he isn’t in a hurry to get investment, fearing that outsiders would interfere in DeepSeek’s decisions.”

Furthermore, Liang is particularly cautious about government-linked investors, since, sources say, he’s concerned about the appearance of Beijing influence over Deepseek and its AI models.

Granted, the lack of government investment hasn’t stopped plenty of Western media outlets and governments from an aboundance of caution. Not only have Italy, Australia, and Taiwan passed bans on the company’s AI models, but the US-China AI chip trade war has further constrained Deepseek for years already.

Deepseek Doesn’t Have Money Troubles

Most startups have outside investors, who hold influence over the business and its path towards profits. That’s not the case for Deepseek, however.

According to TechCrunch, Liang himself owns a full 84% of the company, with the rest owned by individuals with ties to his hedge fund, High-Flyer.

Deepseek also doesn’t need investor money at the moment, even if it does have other problems to face. “Money has never been the problem for us; bans on shipments of advanced chips are the problem,” Liang said in 2023, according to the TechCrunch report.

But How Long Can Deepseek Avoid Monitization?

These new reports aren’t much of a surprise to anyone who has been following Deepseek’s sudden and successful entrance into the competitive global AI business: The company has already taken the moderately original approach of making its AI models open-sourced as well as free to use.

This aligns with Liang’s previously stated problems with the VC-funded startup model popular in the US. He’s against rapid monitization, implying in a 2023 interview that it clashes with a focus on research fundamentals.

That said, while Deepseek may not need investment currently, the available chips that Deepseek needs to continue evolving do cost money. This may be a case in which even Deepseek can’t avoid investors forever.

Online home goods company Wayfair is slashing its workforce by 340 tech employee positions, the company has announced.

This marks the second round of layoffs that Wayfair has initiated since 2025 began. Back in January, the ecommerce retailer cut 730 jobs in the process of exiting its Germany operations.

The company has positioned its new job cuts as a streamlining move, designed to deliver “long-term success.”

What to Know About the Wayfair Layoffs

The 340 lost positions will come as the company closes its Technology Development Center located in Austin, Texas.

Wayfair will be keeping open its other locations, which include four centers based out of North America (Seattle, Mountain View, Toronto, Boston), as well as one in Bengaluru, India.

“To best support Wayfair’s next phase of growth, we must refocus our resources, streamline our operations, and ensure our teams are structured for long-term success,” the company said in a news bulletin on its website.

Wayfair Is “Leveraging Generative AI”

Those of you playing “tech layoff bingo” at home might be wondering if Wayfair has mentioned AI in their layoff announcements. They did.

In a discussion of the ways the company is “harnessing technology” amid its layoffs, Wayfair’s announcement mentioned “advancing personalization and simplifying navigation,” before broaching the topic of AI.

“Additionally, we are leveraging generative AI to boost productivity across our organization, ensuring efficiency and innovation at every level. Together, these advancements empower us to fulfill our brand promise—making it easy for our customers to create a home that is just right for them.” – Wayfair spokesperson

It’s still too early to say whether AI can truly deliver on the lofty promises that many VCs and tech CEOs have issued over the past few years, but it’s already clear that those promises have led to a shakeup in the labor market as executives everywhere feel empowered to lay employees off under the assumption that AI can replace their efforts.

Tech Layoffs Continue

The news from Wayfair is hardly surprising. Ever since late 2022, the tech industry has been continually laying off workers. We’ve charted the biggest layoffs across that time period in a constantly updated list.

It seems 2025 will be a tough year for many businesses, with a second Trump administration triggering market volatility due to an uncertain economy.

The impact of AI-spurred layoffs will make the labor market even weaker. In fact, adding some insult to injury, one way businesses are adapting is the use of AI in filtering through the resumes of tech workers hoping to rebound after their previous layoff.