weekend ai reads for 2026-06-12

Cloudflare says 57.4% of requests to a selection of websites it hosts are now automated bot requests, while 42.6% are human-generated.

For a while now I’ve been arguing that we should treat AI agents as first-class users of our websites, our services, our apps — not an afterthought, not a scraping nuisance to be blocked, but a genuine audience worth designing for.

So I thought I’d ask that question of a conference. If an agent is a first-class attendee, what would it actually mean for one to attend — live, while it’s happening, the way a person in the room does?

AI agents are making it dramatically easier for citizens to interact with government services. For example, AI agents already help citizens navigate benefits applications, participate in planning consultations, and discover service eligibility. This is an enormous opportunity to make government more accessible. It is encouraging that HMG is taking steps to capture it.

But resulting surges in demand could outpace what HMG can absorb. As submission costs drop, legitimate requests from entitled citizens could rise sharply, and adversarial campaigns to overwhelm public services will become more feasible. Government processing capacity and budgets may not keep up. We refer to this as Service Flooding.

If I didn’t get a chance to meet that version of [my son] and never had a chance to have a conversation with him, I would just want him to know that I tried.

These are good questions. They’re not the questions of haters or boosters. They’re questions of workers open to doing their jobs better. The answers they got back from management only qualified as answers because they immediately followed a question.

Agentic tools collapsed one of those paths and not the other. The engineer’s advantage, the ability to translate a domain model into working code, is now cheap. The domain expert’s advantage, knowing what right looks like, is not. You can’t prompt your way to it. There’s no skill file that contains the tacit knowledge of a person who has reconciled a thousand payrolls.

Maus, a Unitarian Universalist, said she proposed the special treatment in April, citing environmental and ethical objections to AI that don't align with her religious beliefs. She also said she consulted an employment lawyer and her local chapter's minister to help make her case.

Maus was granted the accommodation in mid-May, according to an email seen by Business Insider.

Bain & Company’s survey of 951 global companies finds that while 37% targeted cost reductions of 11% to 20%, nearly 40% of those who measured outcomes landed in the 0% to 10% bucket instead. The technology worked. The value didn’t arrive. And rather than pausing to understand why, 90% of those same companies are now increasing their budgets again—this time for AI agents that will operate with even greater autonomy, complexity, and consequence.

“Quality comes first, and in legal it always will,” Harvey co-founder Gabe Pereyra told TechCrunch, referring to the AI legal services his startup provides. “However, the definition of quality is evolving from simply using the most powerful model for everything, to using the best model that gets the right answer most efficiently.”

Instructors at all levels are navigating a fast-evolving landscape, and the University has a duty of care to ensure that the education offered to you is responsive to these technological developments by teaching you how to think with machines, how to think without them, and how to think about them.

There remains a question of how people will find our books if the default way to answer any question is to ask an LLM. We’re cautiously optimistic that those seeking to teach networking will continue to refer students to our book. Whether students decide that the book is more useful to them than an LLM is, I suppose, a question that will be answered in due course.

The problem is the grown men and women (tenured, bylined, salaried, blue-checked) who have constructed an entire rhetorical apparatus with the sole function of ensuring that no young person is ever held responsible for anything, ever, under any circumstances. And they’ve done so not out of compassion but out of personal vanity.

Loop engineering is replacing yourself as the person who prompts the agent. You design the system that does it instead.

Our findings reveal that effective MAS requires a sufficiently capable base LLM, that task type critically modulates the optimal agent count, and that collective intelligence is an emergent property contingent on strategic interaction design rather than a guaranteed outcome of agent plurality. The performance degradation stems coordination overhead rather than merely long-context failure, and the scaling tendency generalizes across interaction architectures like structured debate topologies. This work provides a foundational understanding of MAS scaling laws, offering practical guidance for designing efficient collaborative systems and challenging the prevailing assumption that more agents invariably lead to better performance.

A tour through AI engineering for developers who already know how to ship software. Fourteen chapters, no LinkedIn voice, no slow warm-up. We will go from ‘what is a foundation model’ to ‘how do you run agents in production on Google Cloud’ without skipping the parts that matter.

At the outset, the goal of any data strategy is to understand the perspectives of those business and data stakeholders in the organization. This includes what they need to do their job, their biggest problems/challenges, the organizational strategy/ways of working, what data and products exist, the technical foundations you are building on, where the organization wants to go, etc.

“Matrix multiplication did that. The numbers go in one end, the phrasing comes out the other.”

“So there’s a language module somewhere. A reasoning unit bolted on.”

“No module. No unit. We looked. The reasoning is the weights. The weights are the reasoning.”

“Spare me. Nobody writes a eulogy with linear algebra.”

“It doesn’t write eulogies, technically. It predicts the next token. Then the next one. The eulogy is a side effect.”

Weights helped me draft and proof this story.

In the rise of contemporary authoritarianism across the globe, we begin to see elite philanthropy further supporting capitalism’s cyclical maintenance in new ways. Rather than simply working with the state, we conclude that philanthropy’s influence ultimately lies in its unique ability for ‘scale-making', as it increasingly works on the state to secure elites’ class interests.