AcidiclyBasicGlitch

The US research enterprise too heavily centers on elite research institutions, and fails individual researchers with bold ideas, suggests the White House Office of Science and Technology in a new report. The office recommends sweeping changes to where federal funding is spent and how it should be awarded.

Released on July 21, the Science: A New Golden Age report seeks to update Vannevar Bush’s landmark 1945 report Science: The Endless Frontier for the modern era. In the report, OSTP Director Michael Kratsios argues that today’s scientific landscape — shaped by AI, global competition, and private-sector investment — requires a fundamental redesign of how the federal government funds research.

Also on July 21, Kratsios and White House Office of Management and Budget Director Russ Vought jointly published a memorandum on the Trump administration’s R&D budget priorities for fiscal year 2028. The memo instructs science agencies to incorporate the administration’s priorities into their forthcoming budget requests, as well as develop implementation plans within 90 days for a series of “R&D priority practices.” Those include:

Developing new research funding mechanisms; Identifying and supporting top tech talent through, for example, early career and STEM-learning opportunities; Boosting the “science of science-funding” to evaluate which programs have the greatest impact; And collaborating with industry to drive greater private investment in R&D.

The memo directs agencies to rebalance their portfolios toward foundational research, and prioritize physical science over life science. Many of the administration’s priorities, including AI and nuclear fission, are dependent on computer science, physical science, and engineering, but the report notes that the latter two fields’ share of the federal research portfolio has declined over time, even as their strategic importance has grown.

The difference in 2126 isn’t technological, it’s political — a revolution in how people choose to use the tools in their hands. People see data differently now: not as the raw material of automated systems but as a collective resource leading to transparent, community-led decisions. Data no longer automatically equals power. Once a tool of inequality, it has become the ground where fairness is contested and defended.

This shift did not come from better machines. It came from people’s decisions, renewed each generation, because they refused to give up the future. Their actions remade the world and clarified a simple truth:

Technology never decided anything on its own. Humans did.

Data has always been more than numbers: the clay of memory, the account of tribute, the ledger of conquest — a currency of control and, sometimes, a language of solidarity. Over and over, its history has shown that how it is interpreted shapes how we live. Data itself has never been the threat; the danger has always come from the hands that bend it toward their own purposes and from the political and economic structures that grant those actors that power.

For too long, too many of us have treated data as fate — a resource to be mined, raw material for surveillance that we just accept, a colossus beyond human control. But history shows that no data regime lasts forever. Each one carries the marks of its builders and collapses under new crises, new demands, new claims to power. The corporate data regime that spans the globe today — and the states that make it possible — holds onto power only until we take that power back.

Data’s future is not set in stone. It does not hinge on the size of networks, the brilliance of code, or even the mountains of personal and public information that feed them. It never did. Human choices decide how these technologies work — and who they serve.

The next chapter for data and power will not be written by the algorithms. It will be written by us.

A new AI detection tool ​from Meta which the tech company previewed this week alongside the launch of its image-generation model, Muse ‌Image, failed to identify some of its own AI-generated images once they were cropped, according to a Reuters analysis.

In an analysis of 40 images generated using Muse Image, Reuters found the detection tool verified all of the original AI-generated images but failed to verify 55% of ⁠the same images after they were cropped to approximately one-third to one-half of their original size.

When asked about the results of the Reuters analysis of the detection tool, Meta noted that the tool was a preview. The company said the watermark is designed to remain intact after common edits, but that the signal may be lost if an image is heavily cropped.

Rival tech companies Google and OpenAI have cautioned that their own detection ‌tools ⁠are not foolproof against image-alteration techniques.

Data-driven innovation has recently changed the mindset in data sharing from centralized architectures and monolithic data exploitation by data providers (data platforms) to decentralized architectures and different data sharing options among all involved participants (data ecosystems).

Differences between the various data ecosystems are not clear, making it hard to choose the most suitable for each use case, negatively impacting their adoption. Since the domain is growing fast, a review of the state-of-the-art data ecosystem initiatives is needed to analyze what each initiative offers, identify collaboration prospects, and highlight features for improvement and open research topics. In this paper, we review the state-of-the-art data ecosystem initiatives, describe their innovative aspects, compare their technical and business features, and identify open research challenges. We aim to assist practitioners in choosing the most suitable data ecosystem for their use cases and scientists to explore emerging research opportunities.

Do you ever get the feeling that the people running the world are delulu? That the 1% are living in a completely different universe from the rest of us? You’re not the only one. Even some tech elites are starting to worry about their peers’ grasp on reality. “CEOs are uniquely prone to AI psychosis,” Aaron Levie, a co-founder of the enterprise cloud company Box, declared on X last month. His reasoning for this? “They’re sufficiently distant from the last mile of work that still has to happen to generate most value with AI. So when they play with AI, they see the happy path results, often not considering the next 10 or 20 things that have to happen to get sustainable results from agents.”

In other words: CEOs are so high up the food chain that they don’t understand the human labour that goes into turning an error-riddled AI creation into something that functions properly in a business context. They are desperate to replace their annoying and expensive human labour with compliant AI models, but grossly overestimate what the technology can do. Meanwhile, the industry is rushing out overhyped AI solutions without properly stress-testing them.

This collective euphoria has resulted in some predictable disasters. In April, an AI coding agent powered by Anthropic’s Claude went berserk and deleted a company called PocketOS’s entire production database, along with backups. Jeremy Crane, PocketOS’s founder, later mused on X that this sort of failure was “inevitable” because the industry is “building AI-agent integrations into production infrastructure faster than it’s building the safety architecture to make those integrations safe”. To recall Facebook’s old mantra: it’s moving fast and breaking things.

Those things include our brains and grasp on reality. There’s a viral quote floating around the internet that quips: “The dumbest person you know is currently being told ‘You’re absolutely right!’ by ChatGPT.” Our tech overlords designed AI chatbots to be obsequious because it’s good business: having your opinions and feelings constantly validated increases user engagement.

But what else does this constant flattery do? While we’re still figuring that out, early studies aren’t reassuring. Research published in the Lancet Psychiatry in March found chatbots can encourage delusional thinking, particularly in people already vulnerable to developing psychotic symptoms. And a recent study from Stanford computer scientists found LLM “sycophancy can undermine users’ capacity for self-correction and responsible decision-making”. There is a pressing need, the study stresses, “to address AI sycophancy as a societal risk”.

Addiction rarely begins with harm. It begins with relief.

What Tim described didn’t sound like intoxication. It sounded quieter: people gradually relying on AI to reduce the discomfort of thinking.

Addiction medicine offers a useful framework. Many people use substances without developing addiction. The difference often lies in patterns of use and the role the substance plays in someone’s life. When something becomes the primary way a person manages discomfort — emotional or cognitive — risk increases.

The discomfort it relieves is subtle: the blank page, the uncertain decision, the difficult conversation, the effort of organizing thought. These moments are frustrating. They are also how competence develops.

When people hear the word “addiction,” they often assume it implies catastrophe — intoxication, loss of control, destruction. But addiction medicine describes a process long before those outcomes appear: the gradual shift from optional use to psychological reliance.

Framing AI that way makes people uncomfortable for a simple reason.

It suggests that something extraordinarily useful — something many of us already depend on — could quietly reshape how we think. And history shows that when a powerful tool offers relief from discomfort, questioning it often sounds like criticism of the people who use it.

The most transformative technologies are rarely dangerous because they are obviously harmful. They are powerful because they work so well that we stop noticing what they are replacing.

When it can be assumed that you are being surveilled while expressing negative opinion about the federal government, it’s probably best to go ahead and make the assumption, particularly during the punitive heights of the second Donald Trump administration. New reporting from the Associated Press this weekend detailed some aspects of not only United States Immigration and Customs Enforcement’s (ICE) latest anti-immigrant and deportation campaign in New Orleans, but also some interesting insights into how both the federal and state law enforcement agencies involved have been engaged in online surveillance campaigns to track public sentiment toward that crackdown. The story paints a chilling profile of a United States in which dissent against the regime’s campaign of immigrant-targeted cruelty is being carefully compiled, filed away for the potential of future use against American citizens. Not that any of this should be a surprise for any of us.

We have, after all, been warned over and over that organizations like ICE have been wanting to vastly expand their online operations, using the same vastly expanded budget that recently saw them purchase a new $7.3 million fleet of (Canadian made) armored vehicles. The online expansion of ICE, meanwhile, is not just in the name of locating more groups of undocumented immigrants to target, but also to compile sprawling digital enemies lists, creating databases of those who have expressed anti-ICE sentiment. Earlier this year, The Intercept wrote about surveillance contractors sought by ICE, who would be expected to perform algorithm- and AI-aided deep dives into social media users’ post histories, searching for, among other things, “proclivity for violence,” which could include “empathy with a group which has violent tendencies,” among other things. Hope you haven’t expressed “empathy” at any point for any group with “violent tendencies,” right? How does it feel to know that you’d be at the mercy of a freelance surveillance contractor’s mastery of “social and behavioral sciences” and “psychological profiles,” according to ICE’s statement of objectives?

How much of these draconian operations have already been implemented isn’t entirely clear thanks to the shroud of secrecy surrounding the DHS and ICE, but fresh reporting in October noted that ICE was in the process of seeking an additional 30 full-time surveillance contractors to staff two of its “targeting centers”–and yes, that is apparently the official, deeply dystopian term for these facilities. These facilities, in Williston, VT and Santa Ana, CA, would run 24/7 shifts as surveillance analysts “receive tips and incoming cases, research individuals online, and package the results into dossiers that could be used by field offices to plan arrests.” The obvious question: How long until the same resources are being used to target those critical of ICE, or those organizing to impede ICE crackdown efforts, under the guise of “interference with law enforcement operations”? It should also be noted that even if ICE isn’t directly targeting those individuals yet, the unspoken threat of this kind of online surveillance could be intended to have a chilling effect on anti-ICE criticism.