MIT just published a study today that should scare everyone AI is making us worse at detecting fake news The same way GPS made us worse at reading maps GPS effect: You stopped navigating → your brain forgot how AI effect: You stopped thinking critically → your brain is forgetting how When GPS arrived we all thought: great, I never have to get lost again Nobody asked: what happens to the skill I just outsourced? Same thing is happening with AI and your ability to think You let AI fact-check You let AI summarize You let AI decide what's credible And every time you do — that muscle gets a little weaker The danger isn't that AI spreads misinformation The danger is that AI makes you too comfortable to notice when it does We built a tool that makes us smarter at tasks And dumber at the judgment that matters most #MIT #AI #FakeNews #CriticalThinking #Misinformation #ArtificialIntelligence #DigitalLiteracy #TechResearch #Cognition #AIRisks

TL;DR: New paper on the economics of fast data delivery. It analyzes how different coding strategies (like RLNC) translate into real economic value for latency-sensitive tasks like blockchain syncing. 📄 arxiv.org/pdf/2603.20426 #TechResearch #Blockchain #RLNC @get_optimum

Headline: 💻 Software Developers: Share your feedback on AI Coding Agents in a quick 13-minute study! 🚀 ​Are you a software engineer, developer, or technical architect who regularly uses AI coding assistants (like Copilot, Cursor, or ChatGPT) in your daily work? Terac is running a brief, remote study to understand developer productivity, pain points, and workflows when interacting with AI-generated code. ​Study Details: ⏱️ Format: 13-minute remote review session 🖥️ Setup: Completed via desktop computer (screen share required) 💡 What you'll do: Walk through your typical workflow, share your screen to demonstrate recent coding tasks, and talk through your thoughts on the usability of AI-generated code. ​Help shape the next generation of developer tools and AI assistants! ​👉 Check your eligibility and apply here: terac.com/refer/6de34629 ​#SoftwareEngineering #DevTools #AICoding #GitHubCopilot #SoftwareDevelopers #TechResearch

Hacking HDD Firmware: A Reverse Engineering Journey Working on an Xbox 360 race-condition exploit, I needed to modify HDD firmware to add read delays. Here's what I learned dumping & patching drives from WD, Samsung & Hitachi The Process: Dump firmware Found WD dump via HDD Guru forums; Samsung PM871a firmware updater on Lenovo's site Analyze in IDA WD used modified LZHUF compression; Samsung used simple nibble obfuscation (reversed from updater) Find the read handler Traced VSC (Vendor Specific Commands) via JTAG debugging to locate DMA READ EXT processing Patch & test Injected ~200ms delay into read loop; actual delay ~450ms (close enough for PoC) Key Discoveries: • WD drives store code in overlay modules on the platters' service area not just in SPI flash • JTAG debugging on live HDDs is possible (but fragile Windows timeouts are brutal) • VSC command tables can be mapped by poisoning memory breakpoints • Some Samsung firmware uses unknown ISAs — saved for Part 2 with AI-assisted RE The Twist: After weeks of firmware hacking… the Xbox 360 exploit started working *without* any HDD mods. The race condition resolved itself through other timing variables. Mission accomplished, but the firmware RE skills stay Tools Released: IDA loader plugins, firmware unpackers, VSC command scripts all open-sourced on GitHub. Hope this helps others dive into embedded RE #FirmwareHacking #ReverseEngineering #InfoSec #EmbeddedSecurity #Xbox360 #HardDrive #WD #Samsung #JTAG #IDAPro #CyberSecurity #EthicalHacking #ExploitDev #OpenSource #TechResearch

🧠 Remote AI Safety Fellowship 2026 Applications are now open for the Apart Research Secure Program Synthesis Fellowship, a mentor-led remote research programme exploring AI safety, formal verification, and secure software development. 🔬 Research Areas • AI-generated code verification • Formal methods and theorem proving • Specification validation and synthesis • Adversarial robustness in AI systems • Secure software engineering 🌍 Fellowship Benefits • Work with experienced mentors and researchers • Access compute resources and API credits • Collaborate on cutting-edge AI safety projects • Develop research papers and prototype tools • Present your work at a final demo day 📅 Key Dates • Application Deadline: 31 May 2026 • Fellowship Starts: 15 June 2026 • Duration: June–September 2026 • Fully Remote Participation 🔗 Learn more: globalsouthopportunities.com… #Fellowship #AISafety #ArtificialIntelligence #MachineLearning #FormalVerification #CyberSecurity #ResearchOpportunity #RemoteFellowship #TechResearch #AIResearch #GlobalSouthOpportunities

🤝 @TheFuturumGroup has entered into a definitive agreement to acquire @ETRnews. This announcement marks an important step in Futurum’s continued focus on decision-grade intelligence for technology leaders, investors, and enterprise decision-makers. By bringing together Futurum’s analyst expertise, advisory, intelligence, and strategic content capabilities with ETR’s institutional-grade technology spending data, this combination is designed to strengthen the signal organ Hear directly from @danielnewmanUV and @bmlascolea on what this means for Futurum, ETR, and the future of technology intelligence. #TheFuturumGroup #ETR #TechResearch

• Authorities: @CyberDost, @indiancert, @KolkataPolice, @WBPolice • Technical Handles: @HackerOne, @Bugcrowd #CyberSecurity #ScamAlert #CyberCrime #OnlineFraud #India #Infosec #TechResearch #Kolkata #Howrah

Launching the IIIT-B–De Gruyter Brill Series! De Gruyter Brill and IIIT-Bangalore will launch a new book series dedicated to Digital Transformation and Applied Computing For Details & guidelines: iiitb.ac.in/iiitb-press . . #IIITB #IIITBangalore #DigitalTransformation #TechResearch #DeGruyterBrill #AcademicPublishing #AI #DataScience #STEM

For any expert monitoring agentic AI, Anthropic’s new Claude Opus 4.7 reveals a fascinating structural shift #ClaudeOpus #LLM #AutonomousEngineering #Anthropic #TechResearch

"Trust the AI, Doubt Yourself." 🤖📉 A fascinating new paper from McMaster University reveals a striking cognitive asymmetry in Human-AI interaction. Are we sacrificing our own confidence for machine efficiency? Let’s dive in. 👇 #AI #HumanComputerInteraction #TechResearch

One of the more interesting structural decisions in Perceptron is how the network observes bandwidth contribution through ongoing interaction rather than static registration. Nodes participate continuously in network activity, which allows the system to evaluate real connectivity conditions such as responsiveness and throughput. That observation layer becomes essential when the system needs to coordinate resources across many independent environments. The node layer is intentionally lightweight, which reduces operational friction for participants. Instead of requiring specialized infrastructure, the client focuses on exposing available connectivity and maintaining stable network interaction. When participation is simple to maintain, distributed systems tend to develop a wider and more diverse node topology. Another design choice is the clear separation between infrastructure participation and coordination logic. Nodes provide connectivity and bandwidth, while a higher-level orchestration layer determines how workloads are distributed and verified. By keeping orchestration external to individual nodes, the system prevents local complexity from spreading across the entire network. This structure also creates a feedback loop where infrastructure growth improves routing flexibility. As more nodes become active, the network gains additional paths for workload distribution. Increased path diversity typically improves resilience because the system can redirect activity away from congested or unstable segments. Taken together, the architecture suggests a network that evolves through measurable contribution, modular coordination, and distributed participation. The long-term stability of the system will likely depend on how effectively these mechanisms continue adapting to changing network conditions as the infrastructure expands. @PerceptronNTWK #Perceptron #DistributedInfrastructure #NetworkArchitecture #EdgeNetworks #InternetInfrastructure #SystemsEngineering #TechResearch

One detail that continues to stand out to me while following Fluton is how the project treats information exposure as a structural problem, not just a user-level issue. In most open computational systems, the moment an action is submitted, its details become visible to the entire environment. That visibility creates a predictable pattern where strategies, balances, and execution paths can be analyzed in real time. Fluton attempts to address that problem at the earliest stage of interaction through its Encrypted Intent model. Instead of broadcasting raw instructions, users submit encrypted intentions describing what they want executed. This prevents sensitive details from becoming visible before computation even begins. What makes this approach interesting is how it shifts the architecture of coordination. Many systems rely on full transparency for verification, but that transparency also exposes strategic information. Fluton’s model suggests that coordination can still happen even when the underlying data remains confidential. Another component that reinforces this direction is the use of Fully Homomorphic Encryption (FHE). With FHE, computation can be performed directly on encrypted data. In practice, that means instructions can be processed while the system never sees the underlying information in plaintext form. From an infrastructure perspective, that’s a meaningful step toward enabling confidential computation at scale. Of course, encrypted computation introduces another challenge: verification. Systems still need to prove that results are correct. Fluton addresses this through Zero-Knowledge verification, which allows the network to confirm the validity of outcomes without revealing the private data used to generate them. When I look at these components together—Encrypted Intents, encrypted computation through FHE, and verification through Zero-Knowledge proofs—I see a consistent design philosophy emerging. Each piece solves a different stage of the same problem: how to preserve confidentiality while maintaining verifiable execution. Another element that I find significant is the project’s orientation toward interoperability. The architecture appears to position Fluton as a privacy layer that other computational environments can integrate with, rather than building a closed system around itself. That kind of design signals a long-term focus on infrastructure rather than isolated applications. After spending time observing the project’s direction, what keeps my attention is how deliberate the design choices appear to be. Instead of patching privacy problems after they appear, the architecture attempts to reshape how information enters and moves through computational systems from the start. If that model proves scalable, it could influence how future digital infrastructure balances openness with confidentiality. @FlutonIO #Fluton #EncryptedIntents #FHE #ZeroKnowledge #ConfidentialComputation #DistributedSystems #TechResearch

Perceptron’s coordination layer is structured to treat network capacity as something that can be dynamically discovered rather than statically assigned. Nodes expose their available bandwidth to the system, and the orchestration layer schedules workloads based on real-time responsiveness and availability. This matters because network conditions constantly change, and static allocation would leave large portions of capacity unused. Another important aspect is how the node environment prioritizes lightweight deployment. The infrastructure does not require specialized hardware or complex configuration, which lowers the barrier to participation. When infrastructure systems reduce operational friction, they tend to grow faster and achieve broader geographic distribution. The verification mechanism is also notable. Instead of relying solely on declared node capacity, the system evaluates observable network behavior to confirm whether nodes are actually delivering connectivity. That distinction is important in distributed infrastructure, where reliability depends on measurable contribution rather than theoretical capacity. There is also a clear separation between resource exposure and task execution. Nodes primarily provide bandwidth and connectivity, while the coordination layer focuses on distributing workloads and validating outcomes. Structurally separating these responsibilities prevents the network from becoming tightly coupled around a single operational function. Over time, this type of architecture tends to produce infrastructure that becomes stronger as participation increases. Each additional node adds both capacity and redundancy, which improves resilience against congestion or localized failures. @PerceptronNTWK #Perceptron #NetworkArchitecture #DistributedSystems #EdgeInfrastructure #InternetInfrastructure #SystemsDesign #TechResearch

Art, a medium for expression and channeling emotions, has also found its way into solving real-world problems. A challenge as common and complex as packing irregular objects efficiently may have found 3D Anamorphic Art as the missing piece of its puzzle. When light shines on seemingly random objects placed on a table, their combined shadows can form a clear image, such as a face or a design. From one viewpoint, the image makes perfect sense, but from another, it appears completely different. In this study, researchers from IITGN and Portland State University draw on this principle to develop a differentiable rendering-based framework called RASP - Revisiting 3D Anamorphic Art for Shadow-guided Packing of Irregular Objects. The algorithm aligns computer-generated object silhouettes with the projected shape of the container they need to fit into. By repeatedly comparing with the target projection, object positions are gradually adjusted until the shadows match the targets and fit inside the container without overlapping. This Signed Distance Field (SDF) approach, which serves as a 3D map, prevents objects from protruding beyond the container and avoids invalid placements and collisions. While traditional packing strategies typically account only for basic shapes such as cubes, spheres, and cylinders, the team developed a method that considers the reality of irregular objects. Beyond improving packing efficiency, this system could help in robotics, automated assembly, logistics optimisation, and even the digital reconstruction of broken artefacts. It also opens new possibilities for computational art, where object arrangements are designed to create meaningful shadows. The team includes Soumyaratna Debnath, Ashish Tiwari, Kaustubh Sadekar, Prof Shanmuganathan Raman. To read more: openaccess.thecvf.com/conten… #RnD #IITGN #IITGNResearchCapsules #ResearchCapsule #IITGNResearchers #ComputerVision #CVPR2025 #3DGeometry #ArtificialIntelligence #ComputationalDesign #AlgorithmicArt #3DModeling #RoboticsResearch #ScientificInnovation #TechResearch #EngineeringSolutions #LogisticsOptimization #DigitalReconstruction @EduMinOfIndia | @PIB_India | @Google | @ANRFIndia | @IndiaDST | @qualcomm_in | @Meta | @Apple | @MSFTResearch | @TCS | @amazonIN | @OpenAI

Feels a bit like a Google Maps moment for scientific research. Instead of endlessly searching, you start navigating. @Oall_global #ODataMap #Oall #EmbodiedAI #ArtificialIntelligence #TechResearch #Innovation

Attend the RISE 2026 Open House and get a chance to apply for a Summer Internship at IIIT-Bangalore. Through RISE, you will get to know the research happening at IIIT-Bangalore and also get a chance to apply for the Summer Internship. Why wait? Limited seats are available Register soon: hi.3it.in/RISE26 For details, visit: rise.iiitb.ac.in Questions? Write to: rise@iiitb.ac.in . . #IIITB #IIITBangalore #RISE2026 #RISEOpenHouse #SummerInternship #ResearchOpportunities #Innovation #TechResearch #FutureResearchers #StudentOpportunities #CampusEvents #ExploreResearch

📢 Call for Submissions: Advances in Gel Materials: Synthesis, Characterization and Applications ✏️ Guest edited by Dr. Nina Danchova and Prof. Dr. Stoyan Gutzov 🔗 Issue: brnw.ch/21x0vZU 📌 #SciComm #STEM #Innovation #TechResearch #SustainableMaterials #ResearchImpact

⚡️ NEW STUDY: BITCOIN POLICY INSTITUTE TESTED 36 AI MODELS ACROSS 9,000 CONVERSATIONS. 🔍📊 #AI #Bitcoin #TechResearch #AIStudy

AI is increasingly shaping crypto security. 👾🔐 OpenAI and Paradigm introduced EVMbench - a framework designed to evaluate how AI systems detect and analyze Ethereum smart contract vulnerabilities. 📊 Recent benchmarks show detection rates improving significantly over time. As AI capabilities advance, its role in DeFi security may expand from audits to deeper risk analysis. The key question: How should the ecosystem balance innovation with proper safeguards? 🏗️ #AI #Ethereum #CyberSecurity #DeFi #TechResearch

Imagine building a university timetable — assigning classes to rooms and time slots while respecting countless constraints. Now imagine needing to present two valid timetables to a decision-maker. Ideally, they'd look as different as possible, since near-identical options don't offer a real choice. This is the diverse solutions problem: finding a curated set of sufficiently dissimilar solutions. In this work, Misra, Mittal, and Rai study this for Boolean satisfiability — where even finding one solution is typically hard, and listing all is harder still. They map out when finding two maximally different solutions is tractable across several structured formula classes. 𝗙𝗼𝗿 𝗺𝗼𝗿𝗲 𝗶𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 : doi.org/10.1016/j.tcs.2025.1… #scicomm #PhD #IITGNResearchSnips #Research #IITGNResearchers #IIT #ComplexityTheory #SATProblems #ComputerScience #Algorithms #NPComplete #InnovationInMath #TechResearch @EduMinOfIndia | @PIB_India | @ANRFIndia | @IndiaDST