Category: AI & Chips

Anthropic, the company behind the Claude AI assistant, says it will set aside 5% of its shares for a new fund aimed at the job disruption AI could cause. It’s a plan, not a payout program yet, and there are a lot of open questions.

What was announced

• Size: 5% of Anthropic’s shares, subject to stockholder approval.
• Possible growth: the board could add another 2% for every 50% rise in the company’s value above a set baseline. The baseline hasn’t been made public, and the top-ups are at the board’s discretion.
• Structure: a wholly owned LLC rather than a tax-exempt nonprofit. Anthropic says that gives it more flexibility, and it has promised the fund won’t be used for lobbying or political donations.

What the money could do

Anthropic says the fund could make large donations to traditional nonprofits, or explore direct payouts to “certain populations.” The company says it has already committed $200 million to studying how AI affects jobs and that “bolder approaches” may be needed. The fund is its next step.

What isn’t decided yet

This is the important part. There are no payments and no sign-up process. Still unanswered:

• Whether shareholders approve the fund
• Who would qualify for direct payments, and how much they’d get
• Which nonprofits or programs would receive money
• How the fund will be governed
• What milestones trigger the extra 2% allocations

Anthropic has said it will share more details on the structure and governance later.

How it fits with Anthropic’s other work

In July 2026, Anthropic published the research agenda for a $200 million Economic Futures Research Fund. It pays for outside studies, pilots and trials. Its priorities include retraining and job placement, updating income support for people displaced by AI, and giving workers a stake in AI-driven growth. The announcement notes that today’s US system for supporting displaced workers assumes joblessness is temporary, and that AI could cause longer-lasting displacement.

How it compares

• OpenAI: The Decoder reports that the OpenAI Foundation holds about 26% of OpenAI’s for-profit company. The same article argues Anthropic’s LLC probably has less outside oversight.
• Google: its AI Opportunity Fund supports AI skills training and doesn’t include direct payments.
• Elon Musk: he said on X that Anthropic should give shares directly to the public, and that he plans to give one SpaceX share to each of roughly 72 million US children next year. He hasn’t said how that would work.

Some people online have questioned Anthropic’s motives. Reports say the company is preparing for a public stock offering, and one adviser called the timing both required and good for its public image.

What it means for workers and small businesses

For now, nothing changes, and nobody can apply for anything. The takeaways:

1. AI job disruption is being treated as a real planning issue. The companies building AI are putting money and shares behind it.
2. Retraining and income support look like the first areas, based on the priorities in the $200 million research fund.
3. The details will decide everything: who is eligible, how much is paid, and who controls the fund.

We’ll update this article when Anthropic publishes the governance details or shareholders vote.

Sources: The Decoder, Techloy, Anthropic.

Featured photo: Anthropic CEO Dario Amodei at TechCrunch Disrupt 2023. Photo by TechCrunch, licensed CC BY 2.0, via Wikimedia Commons.

If ChatGPT started answering you with charts, side-by-side comparisons and little interactive tools this month, that’s GPT-6. OpenAI rolled it out to every ChatGPT plan in early October, free users included.

What Intelligent UI actually does

Until now, ChatGPT mostly answered in paragraphs and bullet points. With Intelligent UI, the answer takes whatever shape fits the question:

• Ask it to compare two things and you may get a side-by-side layout.
• Ask how something works and you may get an interactive diagram.
• Ask for help with a calculation and it can build a small calculator or bill splitter right in the chat.
• Ask a simple question and you still get a simple text answer.

The other change you will notice is speed. Instead of making you wait while it “thinks,” ChatGPT can now begin answering and fill in more detail as it goes.

Not seeing it?

OpenAI says the rollout is gradual. If GPT-6 isn’t in your model menu yet, it may just not have reached your account. On company Business or Enterprise plans, an administrator may also need to turn it on.

A few ground rules for using it at work

A smarter assistant is still an assistant. If you or your employees use ChatGPT for business, keep these habits:

• Don’t paste in passwords, customer records or anything confidential unless your company has a business plan with the right privacy settings.
• Check the facts. Numbers, dates, laws and quotes deserve a second look before they go into an email or document.
• Decide what’s allowed. Even a one-page policy about which AI tools staff can use, and for what, prevents a lot of problems later.

GPT-6 arrives as Google pushes its own Gemini agent for work, so expect these tools to keep changing quickly.

Sources: 9to5Mac, OpenAI, OpenAI Help Center.

Some of the most powerful people in tech sat down for lunch at the White House on September 29 and walked out having signed a one-page promise to keep advanced AI in check. It’s voluntary, it’s short, and it’s already being debated.

Who was in the room

The signers included Google’s Sundar Pichai, Meta’s Mark Zuckerberg, Nvidia’s Jensen Huang, Anthropic’s Dario Amodei, OpenAI president Greg Brockman and Elon Musk. Leaders from Microsoft, AMD and Palo Alto Networks were also at the event, though reports don’t confirm they signed.

What they promised

The companies committed to layers of checks to make sure their AI models act as intended, including:

• Internal testing of their own models
• Audits by an outside firm
• Reviews by each company’s board of directors
• Regular meetings with each other to set safety standards and best practices

The document also says that over time it may make sense to turn these steps into laws or regulations, and that the companies will follow them either way.

What else came with it

President Trump said the deal is “almost like a constitution, in a way,” and described companies policing themselves because “their companies are at stake.” He also signed an executive order telling federal agencies to use the term “Super Intelligence” instead of “artificial intelligence,” and talked about naming an AI czar and forming a 10-person oversight committee.

The pushback

Critics say a voluntary promise isn’t enough. Senator Mark Warner of Virginia called for Congress to require mandatory testing, evaluation and incident reporting for advanced AI, and said the administration’s approach amounts to renaming the technology and letting companies regulate themselves. On the other side, Nvidia’s Jensen Huang has said some AI warnings are overblown and that safety and innovation don’t have to conflict.

Why it matters

For most people, the practical effect today is small. Nothing changes in the AI tools you use this week. The bigger question is where this goes next: whether the promises turn into real, outside oversight, or stay a handshake. Either way, the fact that rival companies signed the same page tells you how seriously the industry is taking the pressure.

Sources: CBS News, Al Jazeera, SiliconANGLE.

Google has a new pitch for businesses: instead of a dozen different AI tools, use one AI agent that does a bit of everything and works inside the apps you already use. It even comes with its own company email address.

What it does

An “agent” is AI that does tasks instead of just answering questions. Google says its Gemini agent can plan a job, use the right tools, connect to business systems like Salesforce or ServiceNow and hand back finished work inside your documents and inboxes. It runs in the cloud, so a long task keeps going after you close your laptop. For bigger jobs it can split the work among temporary helper agents.

Two details stand out:

• “Coworker” mode gives the agent its own company email address and storage, so it can work more like a team member than a tool.
• It isn’t only Google’s AI. The agent picks the model it thinks fits each job, and it can use Anthropic’s Claude models as well as Google’s own Gemini models.

Google’s timing isn’t an accident. OpenAI launched its own always-on agents in September, and both companies are racing to become the AI that businesses run their day on.

What to think about before you invite one in

An AI with its own email address and access to your files is, from a security point of view, a new employee. Google says each agent gets its own identity, its actions are recorded in audit logs, and it runs in a protected sandbox. Those are good signs. Your own setup still matters:

• Give it the least access it needs. Start with one team or one task, not the whole company drive.
• Know who is responsible for it. Someone should own its settings and review what it does.
• Watch the bill. Google mentions spending caps that pause an agent when it hits a limit. Use them.
• Keep sensitive data in mind. Decide in advance what it may and may not touch, especially customer and financial records.

AI agents are coming to small businesses whether we’re ready or not. The companies that set simple rules early will get the benefits without the surprises.

Sources: Unite.AI, CNBC.

If a new laptop or phone looks more expensive than you remember, you aren’t imagining it. A worldwide shortage of memory chips is pushing prices up, and the cause is the AI boom.

What’s going on

Every computer needs memory (RAM) and storage. The same few companies, mainly Samsung, SK Hynix and Micron, also make a special high-speed memory called HBM that goes into AI chips. AI companies pay far more for HBM, so the factories have shifted toward it. HBM also takes roughly three times as much chip-making material per gigabyte as the memory in your laptop, so every factory line that switches to AI memory removes even more of the regular kind from the market.

The result is simple supply and demand. HP says memory now makes up about 35% of a PC’s component cost, up from 15 to 18% just two quarters earlier. Dell, Lenovo, HP, Acer and Asus have all raised prices on their 2026 models. The market research firm TrendForce expects regular memory prices to rise another 10 to 15% this quarter, and memory makers are warning that supply will stay tight into 2027.

What this means for you

If you need a computer soon, don’t hold out for a price drop. Nobody in the industry is predicting relief in the next year. Waiting will probably cost you more, not less.

If your current computer still works, make it last. Gartner expects people to keep their PCs about 15 to 20% longer than before. A few smart moves can buy you real time:

• Swap an old hard drive for a solid-state drive (SSD). It’s often the single biggest speed boost an older PC can get.
• Clean up startup programs and remove software you no longer use.
• Keep it updated, and back it up. If you’re on Windows 10, you now have an extra year of free security updates.

For businesses, plan replacements earlier than usual and spread them out instead of replacing every machine at once. Older computers stay on the network longer, so updates, backups and security matter even more.

Need a hand with this at home or at work? HD Technologies sets up, secures and supports Wi-Fi, computers and networks for homes and small businesses. Email us at [email protected].

Sources: Gartner, Tech Times, The Register.

You may never buy anything with “TSMC” on the box, but you almost certainly own something it made. The Taiwanese company builds the chips inside iPhones, Nvidia’s AI processors and much of the hardware that runs the internet. And business has never been better.

The numbers

TSMC’s third-quarter profit beat analyst expectations, and the company raised its forecast for the year again. High-performance computing, the category that includes AI chips, made up about 68% of revenue. Its most advanced manufacturing processes (7 nanometers and smaller) accounted for 79% of wafer revenue, and its newest 2-nanometer process is already starting to show up in the results.

The real story: there isn’t enough to go around

Even with record output, TSMC says demand is still well above what it can supply. Nvidia, Apple, AMD and Broadcom are all competing for the same limited factory space, a fight CNBC described as “fierce.” Analysts say that squeeze is giving rivals Samsung and Intel an opening, since big chip designers are now testing other factories as a backup plan.

There’s a catch for investors, too. On the same day as TSMC’s record report, chip and AI stocks sold off on worries about how the massive AI buildout is being paid for. Record profits and nervous markets in the same week tell you how big, and how uncertain, this moment is.

Why it matters to everyone else

When one company in one place builds most of the world’s advanced chips, any shortage shows up everywhere. It’s part of the reason memory and computer prices are climbing (see why your next laptop costs more), and part of why tech companies keep talking about building more chip factories in the U.S. and elsewhere.

For regular buyers, the takeaway is practical: expect new gadgets with the latest chips to stay pricey and occasionally hard to get, at least through next year.

Sources: CNBC, CNBC, DIGITIMES.