Contact information, direct messages and connected accounts were all potentially compromised, Meta said.

Hackers used Meta AI to hack into 20,225 Instagram accounts, Meta reported in a US local government data breach notice on 5 June.

According to the notice to the attorney general for Maine, the breach occurred on 17 April, but wasn’t discovered by the company until more than a month later, on 31 May.

The company explained that hackers exploited a now-resolved bug in its AI-assisted support tool designed to help Instagram users access their account after being locked out.

“HTS (high touch support) is an AI-assisted support tool designed to help users who are locked out of their Instagram accounts regain access,” said Amber Hannah, Meta’s associate general counsel for incident response.

“Users can request support from HTS and, as part of that process, can ask that a password reset link be sent to their email address.

“The tool itself worked properly and functioned as intended; however, due to a bug in a separate code path, the system did not properly verify that the email address provided by the individual requesting a password reset matched the email address associated with that user’s Instagram account.”

The bug allowed hackers to avoid triggering Instagram’s automated account protections, enabling password reset links to be sent to an email not connected to the account. Bad actors were then able to reset passwords to gain access to victims’ accounts if they did not have two-factor authentication enabled.

The hack affected prominent figures’ accounts, including the inactive Instagram handle for the Obama-era White House, beauty retailer Sephora and a senior US Space Force official.

Meta said that hackers could have potentially accessed sensitive data, including contact information, direct messages and communications, and connected accounts and linked services, such as email IDs. The company said that it would fix the bug before relaunching the AI tool.

In 2024, the Irish Data Protection Commission (DPC) fined Meta €251m for a 2018 data breach affecting approximately 29m Facebook accounts. The same year, the watchdog fined Meta €91m for improperly storing passwords.

In 2023, the company was fined €1.2bn by the DPC for violating GDPR guidelines by transferring users’ personal data outside of the EU.

AI-enabled cybercrime is fast becoming a sore point for companies, as attacks become more frequent and sophisticated. Just last month, hackers stole 8TB of data from the Taiwanese electronics manufacturer Foxconn, while medical equipment manufacturing giant Stryker was hit by a global cyberattack in March.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

A joint research collaboration between researchers at the University of Illinois at Urbana-Champaign (UIUC), UC Berkeley, and the open source AI-native vector database platform Chroma unveiled Harness-1, a 20-billion parameter open-source search agent built atop OpenAI’s gpt-oss-20B open source model that fundamentally redesigns how AI executes complex retrieval tasks.

Harness-1 achieves a massive leap in performance, scoring 73% average on its ability to recall relevant information correctly from a curated dataset, outperforming even GPT-5.4 (70.9%) and the next, most accurate open source search agent, Tongyi DeepResearch 30B, by 11.4 percentage points. (While GPT-5.5 has also been out for more than a month, the researchers didn’t test against this model as it wasn’t available when they were building theirs.)

Crucially for developers, the model and its environment are available immediately under the highly permissive Apache 2.0 license and model code/weights on Hugging Face.

Harness-1 also serves as proof-of-efficacy of another effort, Tinker, the distributed, web-based AI model training and fine-tuning API developed by Thinking Machines. Tinker was used specifically to train and run inference for Harness-1, highlighting how interactive infrastructure is actively enabling the next generation of autonomous models.

So how did the researchers do it?

Benchmarks Decoded (and Why Harness-1 Could Help Enterprises Tremendously)

To actually put these models to the test, the researchers evaluated Harness-1 and its competitors across eight highly complex search benchmarks. Rather than asking simple trivia questions, these tests required the AI to act like a real researcher sifting through diverse, dense data sources.

The benchmarks spanned several different domains, including open web searches, complex financial filings from the SEC, technical patent databases from the USPTO, and “multi-hop” question-answering tasks where the AI had to logically piece together scattered clues from multiple different documents to arrive at the correct answer.

When the results came in, Harness-1 dominated the open-source competition in its ability to successfully find and curate the right facts. Even more impressively, this relatively small 20-billion parameter model went toe-to-toe with massive, expensive proprietary AI systems. It actually outperformed heavyweights like GPT-5.4, Sonnet-4.6, and Kimi-K2.5 — thought to be the hundreds of billions or trillions of parameters. Only one giant frontier model—Opus-4.6 — managed to narrowly edge it out in overall average performance.

Harness-1 achieves its performance gains by offloading the exhaustive “bookkeeping” of a search session out of the model’s working memory and into a structured software environment.

As enterprise use cases grow more sophisticated, demanding that models autonomously sift through thousands of corporate documents or financial filings, these systems frequently succumb to “search amnesia”—forgetting their original queries, looping over rejected documents, or losing track of the specific claims they are trying to verify.

Until now, the prevailing solution to this amnesia has been brute force. Engineers typically force models to constantly reread an ever-expanding, append-only transcript of their own actions, piling every search, read, and thought back into a massive context window.

Harness-1 introduces a paradigm shift away from this method, proving that the bottleneck for true artificial autonomy isn’t necessarily the size of the model, but how efficiently its working environment manages state. It highlights once more, as Anthropic’s Claude Code has also done, that the raw model is arguably less important than the harness — or set of conditions — through which it runs.

Technology: Doing the Paperwork in the Environment

To understand the technical leap of Harness-1, consider a real-world analogy.

Imagine hiring a brilliant research assistant and placing them in an empty room without a desk, notepads, or filing cabinets. You ask them to write a comprehensive report on a highly complex topic, which requires them to read dozens of books while keeping every single quote, citation, and dead-end search perfectly memorized in their own head. Eventually, no matter how intelligent the assistant is, their cognitive load will max out, and they will start dropping facts or losing the thread of the assignment.

This is exactly how traditional search agents operate today. They are trained as policies over growing transcripts, meaning the model searches, reads, searches again, and appends everything into its own context window.

As lead researcher Patrick (Pengcheng) Jiang of the University of Illinois noted on X: “At some point the model is not just ‘searching’ anymore. It is also being asked to be a memory system, a note taker, a verifier, and a librarian.”

Harness-1 solves this by giving the AI a desk and a filing cabinet—what the research team calls a “state-externalizing harness.”

This harness is an active, surrounding environment that takes over the routine bookkeeping, maintaining a recoverable working memory that includes a candidate pool of documents, an importance-tagged curated evidence set, compact evidence links, and verification records.

By separating semantic choices from structural state management, the AI is freed up to do what it does best.

The policy still decides what to search, determines which documents to keep, and knows when to stop, while the environment simply holds the state.

Here is a subsection breaking down the training methodology and how it differs from prior agentic search models:

Training Harness-1: A Masterclass in Data Efficiency

The training pipeline for Harness-1 represents a fundamental shift in how the AI industry approaches agentic learning.

Historically, developers have treated search agents as policies operating over massive, ever-growing transcripts, forcing reinforcement learning (RL) algorithms to simultaneously optimize both semantic reasoning and the raw memorization of a search state.

Harness-1’s creators took a radically different approach: because their custom “harness” handles all the routine bookkeeping—like maintaining evidence links, candidate pools, and verification records—the training process only needed to teach the model how to operate this structured interface.

This division of labor drastically simplified what the underlying 20-billion parameter model actually needed to learn.

The process began with a remarkably narrow Supervised Fine-Tuning (SFT) stage. Rather than scraping petabytes of new behavioral data, the team generated just 899 filtered trajectories using a GPT-5.4 teacher agent that was plugged into the exact same harness environment the student model would eventually use.

The goal of this SFT phase was not to inject vast amounts of domain knowledge into the model, but simply to teach it the mechanical rhythms of a good researcher: how to format tool calls, how to tag documents by importance, and the discipline of verifying a claim before promoting it to the final curated set.

Following SFT, the model underwent Reinforcement Learning (RL) using an algorithm called CISPO, applied over full search episodes capping at 40 turns.

The team designed a highly specific terminal reward function that explicitly separated discovery from selection. The model was rewarded not just for finding a relevant document, but for successfully promoting it into the final answer set, while being penalized if it found the answer but failed to curate it.

The researchers also instituted a “tool diversity” bonus; without this specific incentive, they found the policy would quickly collapse into a lazy, search-heavy strategy where it spammed queries but bypassed the harder work of reading and verifying the text.

What makes Harness-1 truly innovative compared to prior work is its unprecedented data efficiency. The entire model was trained on roughly 4,400 unique items—899 SFT trajectories and 3,453 RL queries.

In stark contrast, competing open-source models required vastly larger datasets to achieve worse results: Context-1 utilized over 17,200 training items, while Search-R1 relied on a staggering 221,300 items to learn search behaviors.

By proving that a smarter external cognitive architecture can replace brute-force data scaling, Harness-1 suggests that the future of agentic AI lies in building better environments for models to work within, rather than just training larger models on more data.

Product: Enterprise Applicability and Generalization

From a product perspective, Harness-1 is delivered as a highly capable 20B agent merged into the openai/gpt-oss-20b base architecture.

For enterprise tech stacks, the applicability is massive because businesses need AI to execute multi-step research across proprietary databases without hallucinating or running up exorbitant compute bills.

Harness-1 manages its frontier-level performance at what the creators describe as “Context-1-level cost and latency.” Because the context window is strictly managed by the budget-aware harness rather than continuously expanding, enterprises can deploy this agent autonomously without incurring the exponential token costs typically associated with long-horizon AI tasks.

Even more impressively, Harness-1 proves it can generalize well beyond its training data. According to the research team, it was incredibly cheap to train, utilizing just 899 filtered supervised fine-tuning (SFT) trajectories and a mere 3,453 reinforcement learning (RL) queries.

“Instead of training the model to survive a giant append-only transcript, we train it to use a structured search interface: search, curate, revisit, verify, and submit,” Jiang explained.

This leanness proves a critical point for the AI industry: developers do not necessarily need petabytes of new behavioral data if they build a better cognitive framework for the model to operate within.

Licensing: The Power of Apache 2.0

One of the most significant aspects of the Harness-1 release is its licensing. In plain language, Apache 2.0 is a highly permissive, enterprise-friendly software license that fundamentally enables commercialization.

Unlike “copyleft” licenses (such as the GPL) that can force companies to open-source their own proprietary software if they integrate the code, or “research-only” licenses that ban commercial use entirely, Apache 2.0 gives businesses the green light to freely build, modify, and monetize the technology.

For developers and startups, this means Harness-1 can be seamlessly integrated into commercial enterprise search products, internal data retrieval tools, or customer-facing AI applications without fear of legal reprisal.

The only major requirement is that users must include the original copyright notice and explicitly state any significant modifications they make to the source code, positioning Harness-1 as a highly viable foundational building block for the enterprise.

Community Reactions: A Resounding Validation

The announcement has clearly struck a nerve within the developer community, validating the very real pain points engineers face when building agentic systems. Jiang’s multi-part announcement thread on X quickly garnered massive traction, pulling in over 256.1K views, 3.7K likes, 2.9K bookmarks, and nearly 300 reposts within a matter of days.

This high engagement underscores a growing consensus in the AI space that brute-forcing context windows is a losing battle.

When Jiang posted on X, “I’ve been wondering: maybe search agents are bad at search partly because we make them do all the paperwork in their head,” the resonance was immediate.

For developers who have spent the last year wrestling with AI agents that confidently forget their primary instructions halfway through a database search, the Harness-1 approach feels like a desperately needed course correction.

Ultimately, the community sentiment highlights a shift in industry priorities. Developers are moving away from asking how large an AI model’s context window can get, and instead asking how efficiently an AI model’s environment can manage that context for it. By offloading the paperwork, Harness-1 is proving that smaller, smarter systems can outmaneuver the giants—provided they have the right desk to work at.

The ‘disappearing into the bushes like Homer Simpson’ strategy is a bold choice.

from the ncose-is-a-threat-to-free-speech dept

In totally sane and not-crazy anti-pornography activism news, the National Center on Sexual Exploitation (NCOSE) considers online pornography a national security threat. This may be the stupidest thing NCOSE has ever claimed in its decades’ long fascistic fight against sexuality. 

The group’s president and chief executive officer, Marcel van der Watt, wrote for the Washington Times about cases of sexual exploitation that could potentially harm individuals who are a part of the military-industrial complex. However, he offers no clear example of such cases and simply relies on the organization’s standard talking points that all sexual expression is bad. 

He writes:

Adults are often an underreported victim group because of shame and fear of social repercussions, yet they are deemed high-value targets by exploiters because of their financial resources.

This claim is meant to apply to military and government employees with security clearances who could be subject to coercion, extortion, and other legitimate forms of exploitation if they post nudes consensually or if they were legitimately victimized by criminals. All of these are real issues, but, in true NCOSE form, van der Watt falsely conflates activities that are illegal with those that are lawful. Van der Watt likens this exploitation as a “symptom” of pornography’s ubiquity in national culture, despite the resurgence of white Christian nationalism.

The argument can be summed up that military personnel have private lives, and sometimes those private lives involve pornography, and because of that, they may be coerced into sharing nudes, which means that the government must outlaw porn as a national security threat.

There are a few logical leaps in there.

He calls for the U.S. Justice Department to stand up its long-dormant obscenity task force that once went after legal pornography producers who released hardcore content. Van der Watt’s calls echo a recent attempt by Sen. Jim Banks of Indiana urging acting Attorney General Todd Blanche to reinstate the task force while singling out the $3.15 billion-valued OnlyFans.com. 

And it’s clear that the real threat here is just… van der Watt doesn’t like the idea that some people enjoy pornography:

Yet the national security threat posed by pornography is incubated on a far more granular level.

Pornography corrodes the exact social foundation that supports national strength. It normalizes sexual objectification and fosters impersonal, consumeristic attitudes toward sex. Over time, its use may lead to habituation, where users require increasingly extreme content to achieve the same level of sexual arousal.

Such rhetoric is part of the far-right’s project to eliminate legal and consensual pornography. This has long been a core tenant of NCOSE’s mission. This is the organization that used to be called “Morality in Media.” Remember, this is the same organization that called for the magazine Cosmopolitan to be removed from Walmart checkout lines because the publication was “pornographic.” NCOSE once went after academic database provider EBSCO for not blocking students from accessing anatomically correct sexual education materials. NCOSE is also one of the central groups to argue the pseudo-scientific claim that pornography is addictive like a drug and serves as a public health crisis. NCOSE persists in these claims, despite no evidence of such. And now it wants you to believe its a “national security” threat?

Michael McGrady covers the tech and legal sides of the online porn business.

Filed Under: marcel van der watt, national security, pornography

Companies: ncose

The subscription now starts at $5 per month.

Global data on math achievement is revealing a dismaying trend: Girls are doing worse than boys — and the margins are huge.

In 2023, fourth-grade boys outperformed their female peers in a vast majority of schools, growing the gender gap that existed prior to the pandemic, according to an international study released last week.

Among eighth-graders, the rate of boys scoring higher than girls increased exponentially since 2019, rolling back gains in math equity that had been shaping up for more than a decade. Matthias Eck, a program specialist for UNESCO’s Section of Education for Inclusion and Gender Equality, tells EdSurge that prior data showed girls were catching up with boys in math achievement.

“But in the latest data, we see that the gap is widening again between girls and boys, and that’s at the detriment of girls, which is quite concerning,” he says.

This international trend echoes what U.S. analysts saw when data from the Nation’s Report Card was released last year.

The latest analysis is based on data from the Trends in International Mathematics and Science Study (TIMSS), a global study published every four years that measures math and science achievement among fourth- and eighth-grade students. The International Association for the Evaluation of Educational Achievement performed the analysis in partnership with UNESCO.

Widening Achievement Gaps

The new data is part of the first set of TIMSS results that measure student performance following the onset of the pandemic. The analysis shows that among top performers in fourth grade, 85 percent of counties’ results skewed toward boys. Slightly over half of the countries and territories from which data was collected have an advanced math achievement gap that favors eighth-grade boys, while none are lopsided toward girls in either grade.

Eck, one of the report’s authors, argues the data shows a correlation between longer school closures and higher rates of learning loss in math, with some variation among countries and territories. “One of the hypotheses is really that those disruptions during the pandemic may have exacerbated existing disparities and have reduced learning opportunities for girls, and potentially those that were at risk of low achievement have been more affected,” Eck says. “The fact that girls were out of school and were not in the learning environment, it could have impacted their confidence, but that’s just the hypothesis.”

But the numbers contain other alarming signals.

For example, the share of regions with a gender gap among fourth-grade students who are failing to reach basic math proficiency is on the rise, and most of them have a higher proportion of struggling girls, according to the report. And while the gender gap in underperformance among eighth-graders is shrinking, the proportion of countries and territories where girls have a higher failure rate spiked.

Researchers are being cautious when it comes to drawing conclusions about the causes behind the results, but girls’ experience of gender stereotypes and confidence in their math abilities can play a role.

“Boys and girls are equally able in mathematics, but these learning outcomes can be shaped by a range of factors,” Eck explains, “and that can be persistent gender stereotypes, but also teacher expectations — and they’re based, of course, on those gender stereotypes.”

Targeted Solutions

UNESCO is pushing education systems across the globe to take a hard look at whether their gender equity strategies are working, especially efforts aimed at younger students.

Eck notes that the consequences of girls’ achievement in math can have far-reaching effects in their lives — and very real consequences in societies writ large. “We know that mathematics is quite foundational to learning across the school subjects, it’s also critical for pathways into science, technology, engineering, mathematics careers,” he says. “These sectors are at the center of innovation, technology advancement, inclusive growth and sustainable development, so that’s quite concerning in terms of those sectors.”

But there’s no widely accepted solution to this problem.

Increasing girls’ math performance will take work at the national policy level, local communities, within families and the culture of classrooms, Eck says. And changes have to include challenging gender stereotypes that limit how far girls think they can go in mathematics, he adds.

“I think what is really critical is that we see those large gaps emerging early, at the fourth grade level when students usually are around 9 or 10 years old,” he says. “That means that whatever we do, the action we take to address the issue must start quite early and must be very targeted.”