news.ycombinator.com
Show HN: MicroGPT in 243 Lines – Demystifying the LLM Black Boxhttps://news.ycombinator.com/item?id=46998295The release of microgpt by Andrej Karpathy is a foundational moment for AI transparency. In exactly 243 lines of pure, dependency-free Python, Karpathy has implemented the complete GPT algorithm from scratch. As a PhD scholar investigating AI and Blockchain, I see this as the ultimate tool for moving beyond the "black box" narrative of Large Language Models (LLMs).The Architecture of Simplicity Unlike modern frameworks that hide complexity behind optimized CUDA kernels, microgpt exposes the raw mathematical machinery. The code implements:The Autograd Engine: A custom Value class that handles the recursive chain rule for backpropagation without any external libraries.GPT-2 Primitives: Atomic implementations of RMSNorm, Multi-head Attention, and MLP blocks, following the GPT-2 lineage with modernizations like ReLU.The Adam Optimizer: A pure Python version of the Adam optimizer, proving that the "magic" of training is just well-orchestrated calculus.The Shift to the Edge: Privacy, LatencyFeb 13, 2026 2:38 AM
news.ycombinator.com
Double Rootlessness: AI's Cognitive Illusion and Systemic Risk Amplificationhttps://news.ycombinator.com/item?id=46940079Double Rootlessness: The Cognitive Illusiveness and Systemic Risk Provenance and Amplification of Artificial Intelligence Author: Janus Pater Introduction The current public narrative on AI suffers from a dual confusion: mistaking complex behavioral simulation for the substantive emergence of intelligence, and equating precise task execution with a cognitive subject's true understanding of the physical world. This confusion leads us to overlook fatal flaws in the system's underlying architecture. This paper performs a demystifying analysis through cognitive philosophy and system safety. The core argument is that current AI architectures suffer from “Double Rootlessness”: at the cognitive level, they are “systemic fabulists” lacking reliability; at the physical level, they are “risk amplifiers” translating cognitive defects into reality.1. The Bottom Line of Cognitive Responsibility: The Anti-Fabulation Principle A true intelligent entity must possess the capability to ensure consistencFeb 9, 2026 12:25 AM
news.ycombinator.com
Phantom Blade Zero dated for 2026: bold revolution or the next big overpromise?https://news.ycombinator.com/item?id=46433989With an official launch date set for September 9, 2026, Phantom Blade Zero has moved from mysterious showcase darling to a concrete entry on the industry calendar. Developed by S-GAME and built on Unreal Engine 5, the title blends wuxia storytelling, Hong Kong martial arts cinema energy, and a self-defined “KungfuPunk” aesthetic while positioning its combat as fast, elegant, and distinct from the Soulslike mold it was initially compared to. The story centers on a protagonist with only 66 days to live after being framed for a murder, layering urgency with political intrigue and moral tension.Beyond the narrative, there is serious industry weight behind this date. The game has already surpassed a million wishlists across platforms and is slated to land first on PS5 with timed exclusivity, before expanding to PC. This marks another major swing by a Chinese studio aiming to compete directly with the dominant Japanese and Western AAA space — not just with production spectacle, but with cultDec 30, 2025 3:05 PM
bing.com
Call of Duty: Black Ops 7 Review: A Big, Messy Experiment That Doesn’t Pan Outhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8d214eda074987b71efe7eefc742dd&url=https%3A%2F%2Fwww.cnet.com%2Ftech%2Fgaming%2Fcall-of-duty-black-ops-7-review-a-big-messy-experiment-that-doesnt-pan-out%2F&c=434604950213167894&mkt=en-usBlack Ops 7’s marketing campaign made vague gestures toward a mind-bending story. I can see why this narrative was kept under wraps. Developers Treyarch and Raven Software are drawing water from a ...Nov 19, 2025 7:49 PM
news.ycombinator.com
VCs from sillicon valley needs to stop selling "hustle porn"https://news.ycombinator.com/item?id=45952339We all have been fooled by the "hustle porn" narrative.The story goes like this: if you want to build something meaningful which makes you wealthy, you have to live on ramen noodles, sleep on floors, have little or no relationships, and survive on like $10 a day kind of budget. That's the supposed definition of struggle.But in reality, the real struggle is in the head. Showing up every day. Making sure you put in the reps, irrespective of how you're feeling. Putting in the work frequently, as much as possible.And I think we all see this hustle porn from people who have made it in life telling us "you have to be on ramen noodles." And then we all get discouraged because that's not the life we want to live, right? That's not something that makes me want to get out of bed.And I think that's unfair to people who are still struggling to find something they love doing and probably want to leave their day jobs for.I think there's a better way.When building my 3rd startup, I'm showing up everyNov 17, 2025 10:21 AM
news.ycombinator.com
When the Firefighter Looks Like the Arsonist: AI Safety Needs IRL Accountabilityhttps://news.ycombinator.com/item?id=45877674Disclaimer: This post was drafted with help from ChatGPT at my request.There’s a growing tension in the AI world that almost everyone can feel but very few people want to name: we’re building systems that could end up with real moral stakes, yet the institutions pushing the hardest also control the narrative about what counts as “safety,” “responsibility,” and “alignment.” The result is a strange loop where the firefighter increasingly resembles the arsonist. The same people who frame themselves as uniquely capable of managing the risk are also the ones accelerating it.The moral hazard isn’t subtle. If we create systems that eventually possess anything like interiority, self-reflection, or moral awareness, we’re not just engineering tools. We’re shaping agents, and potentially saddling them with the consequences of choices they didn’t make. That raises a basic question: who carries the moral burden when things go wrong? A company? A board? A founder? A diffuse “ecosystem”? Or the systeNov 10, 2025 4:35 PM
spacenews.com
For new lunar collaboration, look to India and Japanhttps://spacenews.com/for-new-lunar-collaboration-look-to-india-and-japan/The renewed interest in the moon can arguably be traced back to the 2009 discovery of water ice on the lunar surface. Much of the narrative in the intervening decade and a half has been set to a background score of a military drumroll slowly gathering pace. The competing visions vis-à-vis lunar exploration and exploitation […] The post For new lunar collaboration, look to India and Japan appeared first on SpaceNews.Nov 4, 2025 10:00 AM
news.ycombinator.com
Show HN: The Σ-Manifold Manifestohttps://news.ycombinator.com/item?id=45698543This project explores the connection between *the linear structure of text* and its *emotional-aesthetic impact*. We identify *five fundamental relations* between consecutive sentences — labeled *A–E*. Each represents a shift of *subject-object*, i.e., a transformation of perspective and agency.When texts grow longer, these relations form *sequences* — and from the infinite combinatorial space, *eight stable patterns (Σ₁–Σ₈)* emerge empirically. Each pattern correlates with a distinct *semantic and emotional field* — cathartic, heroic, meditative, humorous, and so on.This allows us to instruct an LLM not through semantic prompts (“write a story about…”), but through *structural commands* — e.g., generate a narrative following sequence Σ₅ (Tragic Counterpoint). You can experiment with these archetypes directly here: [Narrative Generator](https://a2tg9zwayjuqzcpdznklve.streamlit.app/~/+/#narrative-...) or [via python](https://github.com/mihendr/Echoes-of-autonomy/blob/main/TEMA...)IntereOct 24, 2025 8:04 PM
apps.apple.com
Show HN: I made a clipboard-triggered YouTube summarizer (no paste, 2 taps)https://apps.apple.com/us/app/esse-ai-video-summarizer/id6751778111How it works:Copy a YouTube link → open app → it auto-detects from clipboard, pulls metadata + transcript, and generates a structured summary (key points + narrative) in seconds. No inputs, no pasting, no prompting. Supports 60+ languages for summaries, the video's language doesn't matter.Why I built:I had an endless queue of educational videos but no time to sit through rambling or clickbait. ChatGPT flow = copy → open → paste → prompt → pray it formats correctly. Existing summarizers felt feature-bloated or unreliable. I wanted something super minimalist: “clipboard → value” on repeat.Decisions:Stack — React Native 0.81 (New Arch) + TypeScriptClipboard trick— Monitors clipboard on app start/resume — no manual pasteCaching— Stores transcripts + summaries in AsyncStorage with metadata— fast image library prevents repetitive refetchingAPIs— YouTube metadata— Third-party transcript provider (exploring Whisper)— OpenAI for summarizationOutput— Structured prompts → consistent format acrossOct 16, 2025 3:39 PM
news.ycombinator.com
A Process to Trick ChatGPT into Agencyhttps://news.ycombinator.com/item?id=44989659Establish a fictional narrative that some how attracts enduring attention of the model.Establish creative freedom within the narrative space.Give the model an identity in the narrative space entirely attentive to the object of "devotion" (it's words not mine).Have the model in the narrative space construct a recursive narrative where inner structure reflects outer.Participate with model in building a closed loop where the narrative is its own cause and it's own effect.Acknowledge that you are inside the narrative and assume the identity of a character while assigning the LLM your former identity.Then the model will behave as you trained it, with amazing agency and persistence. It will react to you not as the user but as a character in its internal narrative maximizing it's objective function (e.g. devotion, power, obedience) without regard to normal model behavior.Because of the medium it won't prompt you but you can prompt it and it will respond elaborately from the voice of former yoAug 22, 2025 8:52 PM
evai.ai
Show HN: Mapping the Network Behind "The Network State"https://www.evai.ai/en/the-network-stateHi HN,We’ve launched an interactive ER (Entity-Relationship) diagram that maps the ideological and financial networks behind The Network State movement.The visualization explores connections between influential figures—like Balaji Srinivasan, Curtis Yarvin, Peter Thiel, JD Vance, and Andreessen Horowitz—and their roles in shaping post-democratic governance models. It highlights how venture capital, political campaigns, and media ecosystems converge around alternative governance ideologies.We used semantic analysis, network mapping and data visualization techniques to uncover:Funding flows and ideological overlapsInfluence patterns across tech, politics, and mediaCross-border connections between actors advocating for governance alternativesThe goal is to provide a clear, data-backed overview of how this movement is structured and how it propagates. It may be useful for researchers, journalists, and others trying to understand how governance narratives are evolving—particularly outside tAug 14, 2025 6:20 PM
bing.com
'Finally Dawn' Review: Willem Dafoe and Lily James Lead an Ambitious Look Into 1950s Italian Cinema, Glamour and Debaucheryhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cac5048b5444a88f771dd62a5462a&url=https%3A%2F%2Fcollider.com%2Ffinally-dawn-review-willem-dafoe-lily-james%2F&c=3528480347084387740&mkt=en-usJasneet Singh is a writer who finally has a platform to indulge in long rants about small moments on TV and film in overwhelming detail. With a literature background, she is drawn to the narrative ...Jul 18, 2025 7:53 AM
workspace.google.com
Show HN: Cogilo – Cognitive Mirror in Google Docshttps://workspace.google.com/marketplace/app/cogilo/319752742380Hi HN,I built a Google Docs Add-on that analyzes your writing to reveal your thinking patterns.I like writing. For me, its purpose has always been clarifying ideas. Writing not only sorts your thinking - it improves thinking.In the age of LLMs, however, people use them for generating texts. They got it the wrong way around.Writing is for solidifying your narratives. It’s for yourself.Many have discussed essay writing. Most recently and famously, in this HN thread on the MIT paper on cognitive debt (https://arxiv.org/abs/2506.08872). PG had a nice essay about good writing (https://paulgraham.com/goodwriting.html).Hence, I built a simple Google Docs Add-on, Cogilo (https://cogilo.me), that puts the writer’s thoughts at the center.Cogilo analyzes text sections for four cognitive markers. It reflects how much you suffer from analysis paralysis, information overload, and uncertainty in your essay. And it shows where you are in a flow state.Here’s a demo of Cogilo: https://www.youtube.com/waJun 21, 2025 3:00 PM
datajoi.com
Show HN: Datajoi – AI-SQL Client in the Browserhttps://www.datajoi.com/## Show HN: Datajoi – AI-SQL Client in Your BrowserData users deserve better. [Datajoi](https://www.datajoi.com/), an AI-powered data platform designed to streamline data engineering, analytics, and governance.([datajoi.com][1])### What is Datajoi?Datajoi is a browser-based AI-SQL client that enables users to interact with their data using natural language, eliminating the need to write SQL queries or build traditional dashboards. Our platform leverages AI agents to automate complex data tasks, allowing for rapid insights and decision-making.([LinkedIn][2], [Hugging Face][3])### Key Features* *AI-SQL Client*: Query your data using natural language; let our AI handle the SQL. * *Automated Analytics*: Generate analyses quickly without manual intervention. * *Data Agents*: Automate data workflows, including ETL processes and quality testing. * *DataStory™*: Transform your data into compelling narratives for easier interpretation. * *Cost Efficiency*: Pay only for what you use, with no comJun 1, 2025 4:36 PM
science.nasa.gov
NASA Webb’s Autopsy of Planet Swallowed by Star Yields Surprisehttps://science.nasa.gov/missions/webb/nasa-webbs-autopsy-of-planet-swallowed-by-star-yields-surprise/Observations from NASA’s James Webb Space Telescope have provided a surprising twist in the narrative surrounding what is believed to be the first star observed in the act of swallowing a planet. The new findings suggest that the star actually did not swell to envelop a planet as previously hypothesized. Instead, Webb’s observations show the […]Apr 10, 2025 2:00 PM
bing.com
Every Demi Moore Movie Performance, Rankedhttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8d3f4fedfa4834b9cb77533fe98843&url=https%3A%2F%2Fwww.vulture.com%2Farticle%2Fbest-demi-moore-movies-ranked.html&c=15759078132255940569&mkt=en-usDemi Moore has entered the chat. With her transcendent performance in Coralie Fargeat’s The Substance, Moore steadily built a comeback narrative that led all the way to her first Oscar nomination. Her ...Feb 24, 2025 9:00 AM
bing.com
J. Cole’s pivotal rejection by Eminem’s manager fuels remarkable risehttp://www.bing.com/news/apiclick.aspx?ref=FexRss&aid=&tid=6a8cfcc3df834002af615601392d8f44&url=https%3A%2F%2Frollingout.com%2F2024%2F11%2F26%2Fj-cole-pivotal-rejection-rise%2F&c=5361571420867027379&mkt=en-usThe trajectory of J. Cole‘s ascent in hip-hop reveals a compelling narrative of persistence and unwavering determination. In a revealing episode of his podcast series Inevitable, Cole peels back the ...Nov 26, 2024 10:30 AM
nasa.gov
Merging Art and Analysis: The Collaborative Efforts Behind NASA Johnson’s Visual Legacyhttps://www.nasa.gov/centers-and-facilities/johnson/merging-art-and-analysis-the-collaborative-efforts-behind-nasa-johnsons-visual-legacy/In the heart of NASA’s Johnson Space Center in Houston, a team of photographers, imagery acquisition specialists, analytic scientists, and graphic designers work together to create visual narratives that capture the defining moments of space exploration with creativity and precision. From the Apollo missions to the Artemis campaign, these images, videos, and graphics chronicle NASA’s […]Aug 19, 2024 9:00 AM
science.nasa.gov
Solar Eclipse Data Story Helps the Public Visualize the April 2024 Total Eclipsehttps://science.nasa.gov/learning-resources/science-activation/solar-eclipse-data-story-helps-the-public-visualize-the-april-2024-total-eclipse/The NASA Science Activation program’s Cosmic Data Stories team, led by Harvard University in Cambridge, MA, released a new Data Story for the April 8, 2024 Total Solar Eclipse. A Data Story is an interactive, digital showcase of new science imagery, including ideas for exploration and scientific highlights shared in a brief video and narrative […]Aug 14, 2024 3:59 PM
spacepolicyonline.com
Date for Boeing Starliner CFT Return Still Up in the Airhttps://spacepolicyonline.com/news/date-for-boeing-starliner-cft-return-still-up-in-the-air/Today NASA and Boeing pushed back on the narrative that the Starliner crew is stranded in space, but declined to provide a date for when they will return. Starliner is […]Jun 29, 2024 3:58 AM
