1,180 results for experiment · 2.697s

News for “experiment”
14 results • 2691 ms server time
Moozonian News
bing.com• May 23, 2025• 1 min read
What Web Browsing Data Tells Us About How AI Appears OnlineAn iPhone displays Google AI Mode, an experimental search mode that uses artificial intelligence and large language models to generate interactive search results, on March 24, 2025. (Smith ...
Moozonian News
jeffwofford.com• May 7, 2025• 1 min read
Show HN: AI helped me build this financial projector. What do I do with it?I'm an experienced programmer, still figuring out the best way to use AI in my work. Does it replace me? (No.) Does it supplement me? (Yes.) Just how much of my work can it relieve me from doing, if any? Can it reliably take on large projects with essentially no coding from me, or is it better when I do the bulk of the coding and it supplements? So I'm exploring and experimenting.This weekend I decided to put the AI (Gemini 2.5 Pro, via Cursor) front and center, giving it virtually all of the programming burden. I acted as project manager and architect, directing it in application design and the broad steps of implementation but leaving it to do all of the actual programming. In practice I had to intervene a couple of times, either to quickly edit a bit of sloppiness that was easier to fix than to explain or to debug erroneous behaviors that needed browser dev tool access to figure out. The AI couldn't have done it without me, but it did 95% of it and only got stuck three or four times
Moozonian News
github.com• May 6, 2025• 1 min read
Show HN: Plexe – ML Models from a PromptHey HN! We’re Vaibhav and Marcello. We’re building Plexe (https://github.com/plexe-ai/plexe), an open-source agent that turns natural language task descriptions into trained ML models. Here’s a video walkthrough: https://www.youtube.com/watch?v=bUwCSglhcXY.There are all kinds of uses for ML models that never get realized because the process of making them is messy and convoluted. You can spend months trying to find the data, clean it, experiment with models and deploy to production, only to find out that your project has been binned for taking so long. There are many tools for “automating” ML, but it still takes teams of ML experts to actually productionize something of value. And we can’t keep throwing LLMs at every ML problem. Why use a generic 10B parameter language model, if a logistic regression trained on your data could do the job better?Our light-bulb moment was that we could use LLMs to generate task-specific ML models that would be trained on one’s own data. Thanks to the eme
Moozonian News
news.ycombinator.com• May 4, 2025• 1 min read
Show HN: Visualizing web server activity using gourceI just want to share this little weekend experiment. I used gource (usually intended to show git repo changes over time) with web server logs.Video is at: https://www.youtube.com/watch?v=NqV8wqy0v7wIt's showing the web server activity for a non profit, requesting and receiving payments, accepting new members, as well as having the two admins making changes around. This is actually surprisingly easy to do with gource by just providing a log file in a pretty simple format instead of providing the git repo. This is a low traffic server but I find it quite interesting as a way to visualize server activity. I wonder why do visualizations like these aren't seen more often.
Moozonian News
bing.com• Apr 29, 2025• 1 min read
Inteligencia artificial (IA)Desde sus primeros pasos a mediados del siglo XX hasta nuestros días, la inteligencia artificial (IA) ha experimentado una rápida evolución, transformando diversos aspectos de nuestro mundo. La IA ...
Advertisement
Moozonian News
news.ycombinator.com• Apr 21, 2025• 1 min read
OpenAI's new enterprise AI guide is a goldmine for real-world adoptionIf you’re trying to figure out how to actually deploy AI at scale, not just experiment, this guide from OpenAI is the most results-driven resource I’ve seen so far.It’s based on live enterprise deployments and focuses on what’s working, what’s not, and why.Here’s a quick breakdown of the 7 key enterprise AI adoption lessons from the report:1. Start with Evals → Begin with structured evaluations of model performance. Example: Morgan Stanley used evals to speed up advisor workflows while improving accuracy and safety.2. Embed AI in Your Products → Make your product smarter and more human. Example: Indeed uses GPT-4o mini to generate “why you’re a fit” messages, increasing job applications by 20%.3. Start Now, Invest Early → Early movers compound AI value over time. Example: Klarna’s AI assistant now handles 2/3 of support chats. 90% of staff use AI daily.4. Customize and Fine-Tune Models → Tailor models to your data to boost performance. Example: Lowe’s fine-tuned OpenAI models and saw 6
Moozonian News
github.com• Apr 13, 2025• 1 min read
Show HN: A toy MCP for AI agents to code, run, and see output of GPU code safelyEver wondered what happens when you let an AI agent loose on GPU pipeline programming, including shaders? `shaderc-vkrunner-mcp` enables you to find out, without risking any actual GPU or dealing with driver or OS stuff.The core idea is to give AI agents the ability to write, compile, and run GPU code and iterate safely or stably, then let you see the output if desired – all completely locally.It runs entirely inside Docker using Mesa's software Vulkan driver. No need to try to paravirtualize any actual physical GPU, meaning agents can experiment with fancy Vulkan features – atomics, subgroups, cooperative matrices, and more – safely on the CPU. It's surprisingly capable of testing logic, even if it's slow.I went with Vulkan since it's cross-platform (drivers provide it well these days, Mesa runs almost everywhere, and MoltenVK exists for Mac users wanting to experiment similarly), has a sufficient set of exposed primitives for hardware to accelerate the code to competitive levels, eno
Moozonian News
reporoast.fun• Apr 9, 2025• 1 min read
Show HN: Repo Roast – a real-time feed of the funniest comments on GitHubHey HN - I'm Mason, and I'm excited to share a fun side project I've been working on recently: Repo Roast.For me, Github often feels like a serious (and sometimes soulless) place. I built this website to showcase the human, silly side of developers. Repo Roast is built on top of LLM-powered GitHub analytics tooling we’re developing internally at Merit Systems [0].One of the historically cited weak points for LLMs has been their sense of humor. I found that with the right prompt, at scale, they mostly can _identify_ humor pretty well. We experimented with a couple of different models to tune the cost/effectiveness and eventually landed on a strategy where we have gpt-4o-mini do a very light filter pass to reduce the number of calls to GPT-4o to do the actual humor scoring. Interestingly, our first prompt attempts ended up surfacing wayyy more aggressively NSFW content, which is funny that this happens despite all the censoring that OpenAI is doing.So far we have processed the last 90 da
Moozonian News
news.ycombinator.com• Apr 2, 2025• 1 min read
Show HN: Vibe Styler – Transform Any Website's Style With a Prompt + Gemini 2.5I vibe coded a Chrome extension that lets you redesign any website using a simple prompt, powered by Gemini 2.5 Pro's million-token context window. It analyzes the full DOM and existing CSS, then generates contextually-aware styles based on your requests – from specific tweaks ("make the header sticky") to complete themes ("apply cyberpunk aesthetics").The extension maintains style persistence across visits, handles CSP gracefully, and lets you manage styles per website. All processing happens through the Gemini API (you'll need your own key), with no intermediate servers. The API is currently free to use.Note: Since the extension sends the entire context of the website to Gemini, be careful not to send any sensitive data.Try asking it to style as "Star Wars" or "Simpsons", or "add subtle animations to all buttons" – it's pretty fun to experiment with!GitHub: https://github.com/majidmanzarpour/vibe-stylerDemo: https://x.com/majidmanzarpour/status/1907275311798206561
Advertisement
Moozonian News
proxylity.com• Mar 7, 2025• 1 min read
Show HN: UDP and ServerlessHi HN!I'm Lee, the founder and developer of Proxylity.I became interested in UDP as a part of exploring game development, experimenting with IoT/embedded, and running global RADIUS authentication systems. I even wrote a specialized load balancer for it[1].One of my takeaways is how little innovation has happened in UDP-based backend software. It's been the same pattern for decades: old code for stateful processes running on clusters, VMs or containers. Familiar, but expensive (time and/or money) and restrictive.Meanwhile, while the HTTP world has benefited from eliminating most barriers to experimentation: setting up a new HTTP API or site has little to no initial cost, and usage-based cost once running. As a result we have "the web" as we know it today -- full of experimentation and creativity. I'd like to see the same happen for network services in general, and ride the wave created by giants[2][3].UDP is stateless and serverless is stateless, which seems like a match that works. So
Moozonian News
news.ycombinator.com• Feb 26, 2025• 1 min read
Ask HN: How do I finish this lottery scheme?This is a cryptography problem I haven't solved yet.[1] Maybe there is some background research in distributed computing that I don't know about yet.The State of Utopia is a distributed, AI-run, Utopian state, and has certain scarce resources such as citizenships and money that in the interest of fairness must sometimes be granted by lottery, since there is no such thing as half a citizenship for example, or it might have a Universal Basic Income level that it sets in an experiment and only has money to give to a certain number of people, and it doesn't scale to trying to run the experiment on everyone at once.However, absent trust, there is a problem with attackers who want to compromise the entire system to choose their own lottery winner - for example so they can set up a corrupt black market where they choose the lottery winners and require payment for their choice. Or they can just choose themselves as winners.How could the population of Utopia trust the random number which repres
Moozonian News
oss-llmops-stack.com• Feb 26, 2025• 1 min read
Open Source LLMOps StackSome background: I work on Langfuse and we've been collaborating with LiteLLM.(LiteLLM is a Python library and proxy/gateway that handles cost management, virtual keys, caching, and rate-limiting for OpenAI or other LLM APIs. Langfuse manages LLM tracing, evaluation, prompt management, and experiments.)We’ve each been building our open-source projects since early 2023 and learned that many devs and especially platform teams use the two together, so we created an integrated “OSS LLMOps stack.”This is a fully self-hostable, technology-agnostic setup that lets you (1) Use LLMs via a standardized interface without adding complexity to the application; (2) Keep LLM Tracing, Evaluation, Prompt Management in-house for compliance; (3) Track cost and usage via a single interface, create virtual API keys for attribution of costsIt also enables direct transfer of LLM traces from the LiteLLM proxy to Langfuse. This simplifies the rollout of LLMOps practices (observability and evaluations) across m