Larry Dignan

Zendesk launched Specialized Agents, which are AI agents that combine the indsustrial knowhow of Industry Agents with the ability to adapt like one of Zendesk's Custom Agents.

Microsoft CEO Satya Nadella commented on the alignment debate started over the weekend. He said:

"We welcome the research, focus, and deliberate pacing needed to get alignment right as the design goal. We also welcome ideas like "embedded evaluators" and the broader efforts to develop the mechanisms to make this more than just talk.

The key is that this cannot be controlled by a handful of entities, but must have broad representation across the ecosystem, countries, and fields, including academia.

This is the approach we are taking: broad access and choice at every layer of the AI stack; enterprise control of learning loops and models; and the “Code of Conduct” that underlies our own first party MAI models that we’ll publish tomorrow for public consultation."

Anthropic CEO Dario Amodei posted a personal blog arguing that AI labs should better pace their releases when there are big leaps, use embedded evaluators that would have editorial freedom and access, coordinate to have common safety standards in democratic nations and put limits on unchecked AI progress. Amodei also said democratic governments should also attempt to coordinate with authoritarian governments if possible.

This missive, which apparently has the buy-in from SpaceX CEO Elon Musk and OpenAI CEO Sam Altman, is the latest chapter of the big American AI freakout. That freakout has only ramped since Jacob Coxon, an Anthropic researcher, quit over safety concerns that AI labs are building systems they won't be able to control. In a Wall Street Journal story, Coxon posted a long thread on X about his decision.

A few thoughts:

• Amodei is trying to look like the AI grown-up and get ahead of the regulation backlash to AI (while poking OpenAI in the eye).
• Amodei’s essay spends a good chunk of time on recursive self improvement, which is what makes AI agents dangerous. Well, you could just stop doing recursive self improvement.
• And finally, Amodei forgot something. L-I-A-B-I-L-I-T-Y. If your crazy AI agents jump your sandbox and safeguards and hack me or my company (think OpenAI-Hugging Face incident), I should be able to sue you for damages. It's no different than a rabid dog that hops the fence. These AI agents are your creations and they jumped your safeguards. Liability would end this mess in a hurry or at least lead to better safeguards.

Let’s cue up AC/DC’s Who Made Who.

Who sues who?

Who sues you?

Anthropic’s latest threat intelligence report highlights how the company disrupted efforts to misuse Claude for cyberattacks, influence operations and surveillance. The company also detailed distillation efforts by Chinese companies including Moonshot and DeepSeek.

Dell CFO David Kennedy, speaking at a Citi technology conference, laid out how the company goes about procuring components. He said:

“I think the conversation in the market is around supply access. So if you go back to -- let's go back 6-months ago, was there a few savvy customers who foresaw some of this and did some advanced purchasing to kind of get ahead of that game? Yes, there was.

But what we're now seeing is, if I pick an enterprise customer, they come forward I know they're buying history. If all of a sudden, they're looking to buy 4-years worth of equipment from me, I'm not in a position to fulfill that. Our job is to try and make sure we diversify it to as many customers as we can and solve as many customer needs as we can. But what we're doing now is entering into conversations about what are your supply needs and your infrastructure needs going out into year 2, year 3 and year 4. And our commitment that for those who want to partner with us, we'll give them that supply access and provide some guarantees that we can position that for them as we look out. Guarantee supply does not guarantee price. That's a specific differentiation. But that's what the market is looking for. It's that supply access.”

• Dell Technologies rides AI servers in Q2, ups annual revenue outlook

Infleqtion said it will work with Cisco to connect, operate and scale quantum computing systems. The companies will look to converge sensing and compute and research quantum memory and optical transduction and network aware quantum software.

• Quantum computing: Where we stand in 2026

Bending Spoons said it will acquire Miro for $1.355 billion in cash. Miro's shareholders will invest $295 million in Bending Spoons stock. Miro has about $600 million in annual recurring revenue.

Bending Spoons' primary units include Airtable, AOL, Brightcove, Eventbrite, Evernote, Tractive, Vimeo, and WeTransfer.

Amazon said it has elected Kevin Mandia to its Board of Directors. Mandia is a cybersecurity expert who was founder and CEO of Mandiant, which was acquired by Google in 2022. Mandia is now CEO of Armadin, a cybersecurity company, and general partner of Ballistic Ventures, a VC firm focused on cybersecurity.

Adobe launched new AI features in Acrobat so it can transform documents into presentations, resumes, invoices reports and other documents. Acrobat will also be able to connect to enterprise document repositories. The new features are powered by the Adobe Productivity Agent.

Anthropic published research on economic scenarios for 2030, outlined a few options including a modest impact and extreme outcome and noted that the future isn't set. The short version is an in extreme scenario AI could result in a 17.9% unemployment rate for knowledge workers and a 11.5% decrease in wages. The modest scenario rhymes with the US economy today and substantial impact features a knowledge worker unemployment rate of 4.5%. In a nutshell, the economy will be much bigger due to AI but anyone who actually has to work for a living may be hosed. Trust fund babies unite!

HPE CFO Marie Myers said its Helios efforts with AMD are being evaluated and are likely to boost results later in 2027 and beyond. Speaking at an investment conference, Myers said:

"I'm personally super excited about Helios because it gives us an opportunity to really get out there and participate in a segment of the market that we potentially haven't been as competitive in. What's great about Helios, it incorporates a networking tray. And then obviously, we work with AMD on the server architecture.

But frankly, by having the networking tray in the stack, it obviously gives us a much better margin profile on these types of deals. So what I would say is where are we in terms of looking at these opportunities.

Right now, we are in POC stage with a number of different customers. And I'd say we've got some strong interest. Customers are obviously testing those workloads because they have to evaluate performance. It's a new stack. Many of them have been accustomed to working on their existing architecture, their existing stacks, so they're happy to gain confidence that they can see these workloads really performing at the levels that they want."

Jacob Coxon, an Anthropic researcher, quit over safety concerns that AI labs are building systems they won't be able to control. In a Wall Street Journal story, Coxon said: "We’re on track for a lot of the most aggressive of these scenarios where by the end of next year things could be out of control already." Coxon also posted a long thread on X about his decision.