Named after the Greek god of messengers, Hermes watches the education landscape: spotting new opportunities, pressure-testing the ventures we're building, and tracing every read back to the real-world signals behind it.
The evidence library: the raw signals the pipeline is watching across the education ecosystem. Every idea is built from these.
Districts nationwide are grappling with increased special education demands amid persistent staff shortages and compliance pressures.
In just one academic year, Marietta City Schools in Georgia saw the percentage of elementary English learners (ELs) working in or above grade level rocket from 11 percent to 67 percent.
Sixty-five percent of educators use AI to bridge resource gaps, even as platform fatigue and a lack of system integration threaten productivity, according to Jotform's EdTech Trends 2026 report.
Time-intensive home visits show promise.
A fictional space station orbiting the moon is turning into a real-world digital success story. Spacegate Station, a STEM series created in 2022 by Duval County Public School (DCPS) to support daily instruction, has unexpectedly taken off on YouTube, drawing sustained engagement from viewers far beyond the district.
Health sciences professor Humberto López Castillo urges students to use AI to help with science research, but never to lose sight of the human element.
New edtech products that have caught our attention this month
When districts adopt evidence-based practices like Structured Literacy, it’s often with a surge of excitement and momentum. Yet the real challenge lies not in the initial adoption, but in sustaining and scaling these practices to create lasting instructional change.
With the new school year now rolling, teachers and school leaders are likely being hit with a hard truth: Many students are not proficient in reading.
Article URL: https://acadialearning.org Comments URL: https://news.ycombinator.com/item?id=46829300 Points: 1 # Comments: 0
Article URL: https://www.governance.fyi/p/why-singapore-and-estonias-edtech Comments URL: https://news.ycombinator.com/item?id=46825033 Points: 6 # Comments: 3
Imagine trying to teach a student how to navigate the city of New York in 2026 using a map from 1950. The streets have changed names, new bridges have been built, and the traffic patterns have completely changed and are unrecognizable.
Article URL: https://github.com/nyr-github/ai-3d-learning Comments URL: https://news.ycombinator.com/item?id=48330451 Points: 2 # Comments: 0
Across the country, educators, parents, and policymakers are struggling with a question that schools can no longer afford to avoid: What role should cell phones play in today’s classrooms?
I was once asked during an icebreaker in a professional learning session to share a story about my last name. What I thought would be a light moment quickly became emotional.
Chronic absenteeism has stabilized at historically high levels, signaling a long-term engagement challenge rather than a short-term pandemic disruption, according to a new national white paper released by Concentric Educational Solutions.
Without a doubt, career and technical education (CTE) is priceless for high school students wanting to get real-world, hands-on job skills before they graduate and turn their interests into career paths.
Adopting the state-approved Bluebonnet Learning curriculum for elementary students will come with a funding boost of more than $3 million for the district, which also approved a $2 billion operating budget on Thursday. The post Houston ISD board unanimously approves controversial Bible-infused curriculum appeared first on District Administration .
Fayette County Public Schools has been mired in controversy over its financial situation in recent years, with district officials recently saying FCPS has misstated its finances since at least 2008 and has much less money than previously thought. The post Fayette County superintendent files ‘whistleblower’ complaint after being put on leave appeared first on District Administration .
In one of the opening scenes of “Toy Story 5,” Jessie — a cowgirl doll — tries to find out why the twins who live across the street never want to play with her owner Bonnie. What she finds, when she peers through the window of the neighbor’s home, is the two young children on […]
The Texas State Board of Education will vote Friday on a set of new social studies standards that have drawn fire and fervor for espousing pro-American views and Christian values. If approved, the vote would typically mark the beginning of a long, and probably divisive, process to design curriculum based on the standards. But The […]
Sentara Health's virtual nursing partner program is helping give time back to night shift nurses, says this CNO. HealthLeaders spoke to Amber Price , senior vice president and enterprise CNO at Sentara Health , about the challenges nurses face on the night shift and how bringing in virtual nursing partners can help. Tune in to hear her insights. Click here to read the accompanying article. Pillar: CNO Image: Tags: innovation nurses nursing staff technology telemedicine training Secondary Pillars: CNO Article Type: Analysis Published Date: Monday, June 15, 2026 Hide sidebars: Render small main image:
Teacher stress declined modestly in 2026, but teachers were still far more likely than similar working adults to report higher stress, worse well-being and greater financial strain, extending a pattern that has persisted since 2021, according to new RAND research.
Blended learning websites help teachers combine traditional instruction with online learning.
In my Systems Analysis and Design course, students are not handed the requirements for building a software application. They have to uncover them by asking the right questions within an AI-based learning activity. The post The real work of AI and instructional technology is creative appeared first on eCampus News .
The Big Impact of Small Telescopes Elizabeth Redden Fri, 06/26/2026 - 03:00 AM In an era of big data and big telescopes, college observatories remain essential. Byline(s) Alex Gianninas
Female Academics ‘Increasingly Delay Motherhood’ Until Age 35 Susan H. Greenberg Fri, 06/26/2026 - 03:00 AM “Pronounced penalties” for those early-career staff with children may explain why Ph.D.s postpone becoming parents, a recent study finds. Byline(s) Jack Grove for Times Higher Education
Readers Respond on ‘Noncredit’ Sara Brady Fri, 06/26/2026 - 03:00 AM Unofficial names and alternate terms. Byline(s) Matt Reed
Texas Law Dean Pushes Socratic Teaching Amid Rise of AI gianna.jakubowski Fri, 06/26/2026 - 03:00 AM Byline(s) Gianna Jakubowski
Florida Universities Consider Banning Undocumented Students Sara Weissman Fri, 06/26/2026 - 03:00 AM And the board overseeing state colleges is eyeing a similar ban. Together, the policies could make Florida the fourth state to limit noncitizens’ enrollment in public colleges and universities. Byline(s) Sara Weissman
Dear Colleague Letter Asks Colleges to End Affinity Housing Emma Whitford Fri, 06/26/2026 - 03:00 AM The Trump administration alleges that housing that caters to minority students violates the Fair Housing Act. Experts say the guidance is unlikely to withstand legal challenges. Byline(s) Emma Whitford
New HBCU Partnership Speeds Path to Law School Joshua.Bay Fri, 06/26/2026 - 03:00 AM Grambling State has teamed up with Southern University Law Center to allow students to earn a bachelor’s and a law degree in six years, lowering costs and strengthening Louisiana’s attorney pipeline. Byline(s) Joshua Bay
California Adjuncts Sue for ‘Uncompensated Work’ kathryn.palmer… Fri, 06/26/2026 - 03:00 AM Byline(s) Kathryn Palmer
Corequisite Math Might Be Less Effective Than Previously Thought Johanna Alonso Fri, 06/26/2026 - 03:00 AM Byline(s) Johanna Alonso
The politically created academic centers have drawn fierce criticism from faculty, who say they expand state intrusion into higher education.
From a large district’s consolidation plan to a report on states meeting special education requirements, what did you learn from our recent stories?
The Key Podcast: Historians and American Exceptionalism sara.custer@in… Thu, 06/25/2026 - 11:41 PM Byline(s) IHE Staff
arXiv:2605.19576v2 Announce Type: replace-cross Abstract: Self-evolving skill libraries face a silent failure mode we term \emph{library drift}: unbounded skill accumulation without outcome-driven lifecycle management causes retrieval degradation, false-positive injections, and performance stagnation. Recent evaluation confirms the symptom (LLM-authored skills deliver +0.0pp gain while human-curated ones deliver +16.2pp (SkillsBench)), yet the underlying mechanism has not been isolated. We provide (1) a \textbf{reproducible trigger}: ablations that isolate drift: one disables skill injection (flat floor, +0.002), one imposes premature retirement (active harm, $-$0.019); (2) \textbf{trace-level diagnostics}: an append-only evidence log with per-skill contribution scores, attribution verdicts, and router engagement metrics that make the failure visible before it reaches end-task scores; and (3) a \textbf{verified fix}: a minimal governance recipe (outcome-driven retirement + bounded acti
arXiv:2604.26136v2 Announce Type: replace-cross Abstract: Preserving a speaker's voice identity while generating speech in a different language remains a fundamental challenge in spoken language technology, particularly in specialized domains such as scientific communication. In this paper, we address this challenge through our system submission to the International Conference on Spoken Language Translation (IWSLT 2026), the Cross-Lingual Voice Cloning shared task. First, we evaluate several state-of-the-art voice cloning models for cross-lingual speech generation of scientific texts in Arabic, Chinese, and French. Then, we build voice cloning systems based on the OmniVoice foundation model. We employ data augmentation via multi-model ensemble distillation from the ACL 60/60 corpus. We investigate the effect of using this synthetic data for fine-tuning, demonstrating improvements in intelligibility (WER & CER) and speaker similarity (SIM), with gains varying across languages.
arXiv:2604.15877v2 Announce Type: replace-cross Abstract: As LLM agents scale to long-horizon, multi-session deployments, efficiently managing accumulated experience becomes a critical bottleneck. Agent memory systems and agent skill discovery both address this challenge, extracting reusable knowledge from interaction traces, yet a citation analysis of 1{,}136 references across 22 primary papers reveals a cross-community citation rate below 1\%. We propose the \emph{Experience Compression Spectrum}, a unifying framework that positions memory, skills, and rules as points along a single axis of increasing compression (5--20$\times$ for episodic memory, 50--500$\times$ for procedural skills, 1{,}000$\times$+ for declarative rules), directly reducing context consumption, retrieval latency, and compute overhead. Mapping 20+ systems onto this spectrum reveals that every system operates at a fixed, predetermined compression level: none supports adaptive cross-level compression, a gap we term
arXiv:2602.14814v3 Announce Type: replace-cross Abstract: Over the last years, state-tracking tasks, particularly permutation composition, have become a testbed to understand the limits of sequence models architectures like Transformers and RNNs (linear and non-linear). However, these are often sequence-to-sequence tasks: learning to map actions (permutations) to states, which is incompatible with the next-token prediction setting commonly used to train language models. We address this gap by converting permutation composition into code via REPL traces that interleave state-reveals through prints and variable transformations. We show that linear RNNs capable of state-tracking excel also in this setting, while Transformers still fail. Motivated by this representation, we investigate why tracking states in code is generally difficult: actions are not always fully observable. We frame this as tracking the state of a probabilistic finite-state automaton with deterministic state reveals and
arXiv:2602.08275v3 Announce Type: replace-cross Abstract: Elucidating the language-brain relationship requires bridging the methodological gap between the abstract theoretical frameworks of linguistics and the empirical neural data of neuroscience. Serving as an interdisciplinary cornerstone, computational neuroscience formalizes the hierarchical and dynamic structures of language into testable neural models through modeling, simulation, and data analysis. This enables a computational dialogue between linguistic hypotheses and neural mechanisms. Recent advances in deep learning, particularly large language models (LLMs), have powerfully advanced this pursuit. Their high-dimensional representational spaces provide a novel scale for exploring the neural basis of linguistic processing, while the "model-brain alignment" framework offers a methodology to evaluate the biological plausibility of language-related theories.
arXiv:2601.11061v2 Announce Type: replace-cross Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) is highly effective for enhancing LLM reasoning, yet recent evidence shows models like Qwen 2.5 achieve significant gains even with spurious or incorrect rewards. We investigate this phenomenon and identify a "Perplexity Paradox": spurious RLVR triggers a divergence where answer-token perplexity drops while prompt-side coherence degrades, suggesting the model is bypassing reasoning in favor of memorization. Using Path Patching, Logit Lens, JSD analysis, and Neural Differential Equations, we uncover a hidden Anchor-Adapter circuit that facilitates this shortcut. We localize a Functional Anchor in the middle layers (L18-20) that triggers the retrieval of memorized solutions, followed by Structural Adapters in later layers (L21+) that transform representations to accommodate the shortcut signal. Finally, we demonstrate that scaling specific MLP keys within this circuit allows fo
arXiv:2510.00586v3 Announce Type: replace-cross Abstract: Existing data poisoning attacks on retrieval-augmented generation (RAG) systems scale poorly because they require costly optimization of poisoned documents for each target phrase. We introduce Eyes-on-Me, a modular attack that decomposes an adversarial document into reusable **Attention Attractors** and **Focus Regions**. Attractors are optimized to direct attention to the Focus Region. Attackers can then insert semantic baits for the retriever or malicious instructions for the generator, adapting to new targets at near zero cost. This is achieved by steering a small subset of attention heads that we empirically identify as strongly correlated with attack success. Across 18 end-to-end RAG settings (3 datasets $\times$ 2 retrievers $\times$ 3 generators), Eyes-on-Me raises average attack success rates from 21.9 to 57.8 (+35.9 points, 2.6$\times$ over prior work). A single optimized attractor transfers to unseen black box retrieve
arXiv:2506.07031v5 Announce Type: replace-cross Abstract: Emerging Large Reasoning Models (LRMs) consistently excel in mathematical and reasoning tasks, showcasing remarkable capabilities. However, the enhancement of reasoning abilities and the exposure of internal reasoning processes introduce new safety vulnerabilities. A critical question arises: when reasoning becomes intertwined with harmfulness, will LRMs become more vulnerable to jailbreaks in reasoning mode? To investigate this, we introduce HauntAttack, a novel and general-purpose black-box adversarial attack framework that systematically embeds harmful instructions into reasoning questions. Specifically, we modify key reasoning conditions in existing questions with harmful instructions, thereby constructing a reasoning pathway that guides the model step by step toward unsafe outputs. We evaluate HauntAttack on 11 LRMs and observe an average attack success rate of over 70\%, achieving up to 13 percentage points of absolute imp
arXiv:2606.21649v2 Announce Type: replace Abstract: Existing embedding models are inherently static: they encode text segments in isolation, ignoring their surrounding context and temporal order. This paper introduces EvoEmbedding, a novel embedding model that generates evolvable representations for retrieval. It is tailored for long-context scenarios, where information is dynamic, sequential, and requires continuous state tracking. Our design is simple: EvoEmbedding maintains a continuously updated latent memory as it sequentially processes inputs, and uses it alongside the raw content to jointly generate evolvable embeddings. Consequently, for the same query, our model adapts its representation to retrieve distinct targets based on the evolving context, going beyond static semantic search. To equip the model with this capability, we construct EvoTrain-180K, a diverse dataset for the joint optimization of latent memory and retrieval. Furthermore, we introduce a memory queue to prevent
arXiv:2606.21097v2 Announce Type: replace Abstract: Deploying highly capable personalized conversational agents in resource-constrained or privacy-sensitive environments remains a significant challenge. We identify a fundamental bottleneck in the existing approaches: current training paradigms treat personalization and grounding as a single monolithic learning problem. Under these paradigms, language models are forced to simultaneously address what to say (content grounding) and how to say it in a user-specific way (personalization), which introduces significant computational and optimization challenges. Consequently, contextual grounding is often sacrificed for persona adherence, or vice versa, resulting in responses that are either weakly grounded in the conversational history or insufficiently personalized. In this work, we propose the Generic Response-Augmented Generation (GRAG) framework that decouples these competing objectives by leveraging offline, generic responses from high-c
arXiv:2606.19852v2 Announce Type: replace Abstract: Information extraction from pathology reports is essential for cancer staging, tumor registry population. Yet key data remains embedded in narrative reports, making manual extraction labor-intensive and error-prone. Traditional supervised Natural Language Processing pipelines address this through fully supervised Named Entity Recognition and Relation Extraction, but require expensive manual annotation and suffer cascading failures when upstream entities are missed. In this study, we developed a zero-shot, agentic workflow, and evaluated five open-source generative Large Language Models (LLMs) to populate 13 College of American Pathologists synoptic fields from lung resection pathology reports. We compared them against a state-of-the-art supervised GatorTron NER-RE baseline using a novel, registry-aligned evaluation framework. The baseline achieved Micro-F1of 0.960, while the best zero-shot model (GPT-OSS-20B) achieved Micro-F1 of 0.89