False claims in a widely-cited paper. No corrections. No consequences
Community Discussion
The comments convey a broadly critical view of institutional practices, arguing that many professional standards and research fields—such as traffic engineering, management science, and economics—rely on outdated or biased assumptions and lack effective accountability. Peer review is described as a barrier that favors prevailing orthodoxy, while dissenting findings are often suppressed or dismissed. The discussion emphasizes concerns about self‑reinforcing biases, the erosion of evidence‑based consensus, and the tendency for influential institutions to prioritize status quo interests over rigorous truth‑seeking.
Running Tesla Model 3's computer on my desk using parts from crashed cars
Summary
- Objective: Run a Tesla Model 3 Media Control Unit (MCU) and autopilot computer on a desk to boot the vehicle’s OS for security‑research purposes.
- Parts sourced from salvaged Model 3s on eBay: MCU‑autopilot module (≈$200‑$300), touchscreen assembly ($175), and a 0‑30 V/10 A DC supply (needed up to 8 A peak).
- Wiring: Tesla’s public Electrical Reference shows the display uses a 6‑pin Rosenberger 99K10D‑1D5A5‑D connector; individual cables are not sold, only large harnesses. An attempted substitution with a BMW LVDS cable failed due to mismatched dimensions, leading to a short that burned the MAX16932 step‑down regulator, which was later replaced by a PCB shop.
- Network access: The MCU exposes a static 192.168.90.100 address with SSH (port 22, Tesla‑signed keys) and a REST‑like ODIN API on port 8080. Manual IP assignment (192.168.90.X, X > 105) is required.
- Final solution: Purchased the complete dashboard wiring loom (part 1067960‑XX‑E, ≈$80), connected power and LVDS cable, and achieved a fully operational desk‑mounted Tesla computer ready for UI interaction, CAN‑bus exploration, and firmware extraction.
Read full article →
Community Discussion
Comments express enthusiasm for Tesla’s root‑access bounty, viewing it as a well‑balanced way to enable deep security research while limiting broad vehicle exposure. Contributors discuss related automotive hacking practices, hardware setups, and the value of open documentation, noting the usefulness of wiring‑harness information and the appeal of running custom software. Some raise privacy considerations and UI aesthetics, while others share personal projects and compare tools. Overall the discussion is supportive, technical, and focused on expanding research capabilities.
ARC-AGI-3
Summary
ARC-AGI-3 is an interactive reasoning benchmark designed to evaluate AI agents’ ability to explore novel environments, acquire goals dynamically, construct adaptable world models, and engage in continuous learning. Unlike static puzzles, agents must perceive relevant information, select actions, and adjust strategies based solely on experience, without natural‑language instructions. Performance metrics include 100 % human‑solvable environments, skill‑acquisition efficiency over time, long‑horizon planning with sparse feedback, and experience‑driven adaptation across multiple steps. The benchmark quantifies the gap between AI and human learning by measuring planning horizons, memory compression, and belief updating as new evidence arises. Design principles prioritize ease of human onboarding, absence of pre‑loaded knowledge or hidden prompts, clear goals with meaningful feedback, and novelty that prevents brute‑force memorization. A perfect score indicates an AI can match human efficiency across all tested games.
Read full article →
Community Discussion
Comments show mixed reactions to ARC‑AGI‑3. Many criticize the scoring system as overly complex, biased toward the second‑best human and unsuitable for judging true AGI, noting that human baselines and action‑count limits distort results. Some appreciate the benchmark’s novelty, cost‑aware design, and its role in probing LLM reasoning, while others find the “no‑tools” constraint odd and question whether visual‑puzzle tasks reflect general intelligence. Overall, the community acknowledges the benchmark’s ingenuity but doubts its adequacy as a definitive measure of artificial general intelligence.
My astrophotography in the movie Project Hail Mary
Summary
Rod Prazeres’ astrophotography is featured in the end‑credits sequence of “Project Hail Mary.” The material consists of a compilation of his space‑focused images, presented as the visual backdrop for the closing credits. By crediting Prazeres, the production acknowledges his contribution of celestial photography that complements the narrative’s scientific theme. The segment showcases his work without additional commentary, serving primarily as a visual acknowledgment of the photographer’s role in enhancing the film’s aesthetic.
Read full article →
Community Discussion
The comments are overwhelmingly positive, celebrating the photographer’s achievement and the movie’s use of authentic astrophotography. Praise centers on the visual quality, the dedication required to capture real images, and the broader appeal of human‑generated art over synthetic alternatives. Several readers express excitement to watch the film, note the impressive equipment used, and inquire about licensing details, while a minority raise practical questions about access and cost. Overall, the community conveys admiration, enthusiasm, and support for the blend of scientific photography with mainstream cinema.
Two Studies in Compiler Optimisations
Summary
- **Case 1 – Modular increment**:
- Naïve `next_naive(cur,count){return (cur+1)%count;}` compiles to a costly 32‑bit division.
- Providing the pre‑condition `cur
Read full article →
Earthquake scientists reveal how overplowing weakens soil at experimental farm
Summary
Researchers from the University of Washington used distributed acoustic sensing (DAS) on fiber‑optic cables to monitor soil moisture dynamics across experimental plots at Harper Adams University’s farm in the UK. The site contained long‑term no‑till rows, rows tilled to 10 cm and 25 cm depth, and varying tractor‑induced compaction. Continuous ground‑motion data were recorded for 40 hours alongside weather measurements during light‑to‑moderate rain. Changes in seismic velocity—slower in wetter, less cohesive soil—were modeled to quantify moisture content. Results show that tillage and compaction fracture the soil’s capillary network, reducing its sponge‑like capacity, promoting surface pooling, crust formation, and increased erosion risk. The DAS approach proved inexpensive, high‑resolution, and capable of real‑time monitoring, offering potential applications for farm water‑management decisions, flood alerts, improved atmospheric water estimates in Earth‑system models, and refined liquefaction risk in seismic hazard maps. The study, published 19 March in *Science*, was funded by multiple foundations and research councils.
Read full article →
Community Discussion
The discussion references experimental data showing a modest overall yield advantage for no‑till fields, but with specific crops—potatoes, carrots, rutabaga, cabbage and broad beans—performing better under tillage, while others such as shallots, leeks and onions favor no‑till. Commenters note that tillage is often used for weed control, leading to herbicide reliance and resistance, and cite historical overplowing as a cause of the Dust Bowl. Additional points highlight concerns about soil health, fungal networks, permaculture advocacy, and criticism of contemporary agricultural policies, yielding a mixed view without a clear consensus.
The EU still wants to scan your private messages and photos
Summary
The document, titled “Fight Chat Control – Protect Digital Privacy in the EU,” reports that members of the European People’s Party (EPP) are attempting to initiate a new parliamentary vote on Thursday, 26 May, aimed at overturning a prior decision in which the Parliament rejected proposals for indiscriminate scanning of communications. The text characterizes the proposed vote as a challenge to democratic procedure and to the right to privacy. Two visual elements accompany the text: an image labelled “Fight Chat Control” and a logo bearing the same phrase. No further technical details, data, or policy specifications are provided.
Read full article →
Community Discussion
Comments converge on strong criticism of the EU’s “Chat Control” extension, viewing it as a step toward indiscriminate mass surveillance that undermines privacy and democratic standards. Contributors question the practicality of enforcement, warn of abuse and exemptions for officials, and highlight the absence of a counter‑legislation protecting private communications. Many cite authoritarian tendencies in member states, call for stronger encryption, and express frustration that the debate receives little public attention while urging contact with representatives. Technical inquiries about non‑EU users and GDPR compatibility appear alongside broader distrust of the legislative process.
90% of Claude-linked output going to GitHub repos w <2 stars
Summary
The update modifies the initial game configuration so that exactly one can is placed in its correct position at the start of each session. This adjustment ensures that every player begins with a single correctly positioned can, thereby standardizing the opening state and improving consistency across games. The change eliminates variability in the initial layout, providing a uniform baseline for gameplay.
Read full article →
Community Discussion
Comments converge on the view that the high proportion of Claude‑generated commits in low‑star repositories reflects a shift toward personal, utility‑focused projects rather than a lack of value, with many users noting that stars are a weak proxy for usefulness and that most code—human or AI‑assisted—is intended for private or one‑off use. Some express concern about GitHub’s future business model and potential restrictions on free or AI‑driven usage, while others downplay the metric’s significance and highlight the broader trend of reduced effort‑to‑publish costs.
My DIY FPGA board can run Quake II
Summary
The new DIY board uses an Efinix Ti60F256 FPGA (256‑pin BGA) paired with a 1 GB DDR3L IM8G16D3FFBG memory chip (96‑pin BGA) on a six‑layer PCB. The author adopted Efinix’s “DDR3 Soft Controller Core” and followed DDR3 layout guidelines, meeting tight trace‑length matching by compensating across layers. Power‑net decoupling was maximized within the 0603‑size constraint, and a stencil plus a bottom‑heater (221 °C) was used to paste‑solder the BGA components, followed by heat‑gun reflow for bottom‑side parts. Added peripherals include a TFP410 TMDS serializer, USB‑C current limiter and dual‑port power negotiation, an SD‑card voltage switch for UHS‑1 SDR104, a real‑time clock with battery backup, and an ESP32 Wi‑Fi module (later corrected pin assignment).
The SoC inside the FPGA is built from VexRiscv (generated via SpinalHDL) plus custom IP: DDR3 TileLink, UART, I2C, audio, three SPI controllers, video and DMA controllers. Utilization reaches 89 % of XLRs and 95 % of memory blocks; a single RISC‑V core runs at 207 MHz, delivering 511 DMIPS (2.46 DMIPS/MHz), 207 MFLOPS peak, CoreMark 783, 45 MB/s SD read, 348 MB/s CPU memset and 1.13 GB/s DMA memset. GPU logic was omitted due to resource limits, so the DMA controller handles limited graphics tasks.
Read full article →
Community Discussion
Comments highlight both technical and community aspects of the project. Users note a broken link in part 6 and appreciate the detailed write‑up, expressing admiration for the complexity of moving from KiCad schematics to a functional Quake‑II FPGA board. Several remarks discuss cost concerns of multilayer PCBs, inquire about the minimum hardware requirements for running Quake, and suggest broader distribution options such as crowdfunding. Additional interest appears in implementing classic game algorithms in hardware and the routing techniques used, while overall sentiment remains positive and supportive.
Supreme Court Sides with Cox in Copyright Fight over Pirated Music
Community Discussion
The comments view the Supreme Court’s unanimous reversal as a clear win for ISPs and a restraint on expanding copyright liability, emphasizing that the decision limits obligations to monitor users and reduces pressure on service providers. Observers note the relatively small number of infringing accounts compared with the ISP’s scale and draw parallels to earlier Betamax precedent. At the same time, there is caution about possible future tactics by rights holders, concerns about the durability of takedown regimes, and broader criticism of the current intellectual‑property framework.