Latest AI/ML News
5100 matching items
Large Hadron Collider goes offline to make room for its enhanced successor
The High-Luminosity LHC will be mostly the same machine, but it'll deliver 10 times the luminosity and just as little chance of destroying the universe - sorry, conspiracy theorists
.NET's long-term support is not long-term enough, dev complains
Three-year lifecycle leaves enterprises with barely a year to adopt each LTS release
Sony’s next-gen PlayStation will go ‘beyond the living room’
Sony hinted in a recent Q&A with investors that the next generation PlayStation will offer some kind of experience that lets you play games outside of your living room. Here's the relevant portion from the transcript, emphasis mine: Q: How can you bring back to the PlayStation platform users who migrated to gaming PCs during […]
Watch out, Amazon: The Kobo eReader now has a Goodreads rival
Kobo users can now automatically sync their reading progress to StoryGraph, making it easier to track books, reading stats, and challenges without relying on Amazon’s Goodreads.
I replaced my iPhone battery at the Apple store for the first time ever - and learned a valuable lesson
A few weeks ago, I replaced my iPhone 14 Pro battery at the Apple store - and was pleasantly surprised by the results.
Gap Statistic Versus Total Sum of Squares For Cluster Centres (r language)
Hi I had a quick question about what to do when you encounter a situation where the optimal cluster center is different when you graph out the gap statistic and the total sum of squares. As I understand, you were supposed to take cluster center with the highest gap statistic and the center that is at the elbow of the total some of squares graph. Should I take both numbers suggested for the cluster centers for kmeans or am I'm doing something wrong? fviz_nbclust(iris_data, kmeans, method = "wss") # Le graph suggests that the optimal amount of cluster should be 4 where the elbow is iris_gap=clusGap(iris_data, FUN = kmeans, nstart = 25,K.max = 10, B = 50) fviz_gap_stat(iris_gap)
Swiss yodellers rehearse in public fountains as Basel swelters
Basel’s national yodelling festival unfolds during a heatwave as performers sing in fountains, compete across the city and close with street parades.
Because It Speaks In Words
There is a difference between knowing something, and truly understanding it. I think we've all had those moments, the ones where a truth you learned a long time ago really sits with you for the first time. It rests in your mind and stretches out, finally showing off the great expanse of nuance hidden deep within. Photo: mine. Most nights when you happen to spy the Moon, Jupiter, and Venus sitting in a line in the sky, you see them as the dots beside a crescent that they are. Yet sometimes, when framed amid the sunset sky above the wisps of silver cloud, you see them differently. You realize in that moment that you, your great-grandparents, Julius Caesar, and Aristotle all saw this same sky. It has always been for us. We know so much about Jupiter and Venus today, yet no human eye has ever seen more than you're seeing now. I think we are all, in this time, realizing the true power of words and what they mean, not to those who speak, but to those who listen. The Power of Stories It was Eugene Wigner who famously wrote about the unreasonable effectiveness of mathematics in the Natural Sciences, the empirical fact that advanced math is eerily accurate at predicting the behavior of nature, so much so that, for me, it can seem like mathematics is perhaps the true language that nature speaks. However, as children, we do not grow up with an innate understanding of complex mathematics, of this foundational language of nature. Instead we are all born with the innate desire to hear and…
OpenAI is teasing new hardware… for Codex
OpenAI is releasing some sort of device related to its AI-powered coding tool, Codex, on July 15th. In a video posted to X on Monday, OpenAI shows a square-shaped device with several buttons, alongside the caption, "Your favorite Codex shortcuts are getting an upgrade." This isn't the mysterious AI-powered device OpenAI is working on with […]
EU lawmakers say they were blocked from fully inspecting Italy's migrant detention centre in Albania
MEPs from the Greens/EFA group say staff refused to answer questions, blocked access to cells and provided no data during a visit to the Gjadër facility.
Generating Test Data for Pinecone
A repeatable workflow for building large, realistic vector test datasets: CC News to Parquet to local embeddings to Pinecone bulk import.
Waymo and Uber quietly part ways in Phoenix
Uber said it is readying the launch of a separate autonomous vehicle partnership in the city, but did not name the partner.
Conversation tree or branching inside a chat
Thanks for sharing this feature request. I understand how a conversation tree or branching feature could help keep long, complex workflows more organized, especially when users need to explore side topics without cluttering the main thread. I’ll pass this feedback along to the team for consideration. ~ Smith
German politicians vow to stop VW’s mass layoff plan
The giant automaker's planned cuts are among the clearest signs yet of Germany's industrial decline.
Count the number of Safari tabs
Tiniest TIL, using AppleScript to count the number of open browser tabs in Safari: osascript -e 'tell application "Safari" to count tabs of every window' Tags: safari , til , applescript
No GPT Name in warning email
How about as room for improvement: when NEVER flagging an account for existing assets such as GPTs that have been operating for two years, GPTs that may then be completely unmanageable by the user not having a Plus subscription but yet still using ChatGPT for casual and their past chats, you instead: NEVER threaten the status of the account NEVER make automated decisions against the account DO demote items from the store sharing if they match denial policy, as determined by multiple types of graders each returning a positive AND the weight of no positive flags for an ongoing two years taking precedence against one non-deterministic AI (the type that can decide “I should delete all these files on a whim”. And then: Do not take any automated action against accounts without personnel to handle appeals thoughtfully and in a timely manner. Do not employ AI to judge or classify people nor make business-critical decisions (see terms and conditions)
Is AIC-based model reduction before adding latent class membership a defensible modelling strategy?
I am an MSc Biostatistics student working on a study of willingness to quit smoking. I derived a 3-class latent variable using Latent Class Analysis (LCA) from FTND and two motivational indicators. Since these variables were used to construct the latent class, I excluded them from my logistic regression to avoid modelling both the component indicators and their derived latent construct simultaneously. I first fitted a full logistic regression using the remaining conventional predictors, performed AIC-based backward model selection to obtain a parsimonious model, and then added latent class membership to evaluate its incremental explanatory value using changes in AIC and BIC. Would you consider this a statistically sound and defensible modelling strategy? Are there any methodological references or alternative approaches that you would recommend?
Implement a backup strategy for Amazon Quick Sight BI assets
In this post, we cover best practices for implementing an effective backup strategy for BI assets in Quick Sight. We start by covering the options for selecting the assets to include in your backup, then explain the high-level APIs available for that purpose, and finalize with sample code to help you get started quickly.
Anthropic and Gov. Newsom forge deal allowing California government to use Claude at half price
As Anthropic forges a closer relationship with the state of California, the federal government has made an enemy out of the OpenAI rival.
South Korean tech giants commit over $550B to ease ‘RAMageddon’
The world's two largest memory chip companies vow to build more memory lab fabs as South Korea positions itself as an AI tech powerhouse country.
Newsom and Anthropic reach deal to give local governments discounted access to Claude
The agreement would make Claude the first artificial intelligence tool available to all state agencies and local governments.
Amazon engineers are reportedly distilling Anthropic models to cut costs before new token-based pricing kicks in
Amazon engineers are already distilling Anthropic models into smaller, cheaper versions for internal use. Starting next year, Amazon will pay by tokens processed rather than compute hours, which could push costs up sharply. The company is also exploring alternatives like OpenAI. The article Amazon engineers are reportedly distilling Anthropic models to cut costs before new token-based pricing kicks in appeared first on The Decoder .
MEPs urge FIFA to investigate chief Infantino over Trump peace prize
Gianni Infantino's public statements in favor of the U.S. president are in breach of FIFA statute, according to an ethics complaint.
Thousands take part in New York Pride parade
Thousands join New York Pride parade in Manhattan with floats, music and political leaders, marking celebration and remembrance of LGBTQ+ rights.
US government wants to have a useful quantum computer by 2028
The US government is trying to speed up the development of quantum computers so it can have one sooner
AI agents are not your “coworkers”
This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. Imagine coming in to work to learn that a new underling will report to you. The worker is not a person but an AI tool—one that your company nonetheless calls Alex, an…
Trump's student-loan overhaul takes effect July 1 — but it's not the deadline for borrowers to act on repayment changes
Trump's student-loan changes take effect on July 1, but the millions of borrowers enrolled in SAVE will have time to switch to a new repayment plan.
EU seeks AI independence as Austria proposes luring Anthropic to Europe
Austria's State Secretary for Digitalization, Alexander Pröll, is calling on the European Commission to explore bringing Anthropic to Europe. He's responding to the U.S. ban on advanced AI models from OpenAI and Anthropic for foreign users. The initiative is likely unrealistic, though. And the alternative floating in the background, Chinese AI models, would just trade one dependency for another. The article EU seeks AI independence as Austria proposes luring Anthropic to Europe appeared first on The Decoder .
Automated Tablet Defect Inspection
Train a pill detection model, crop each tablet, classify defects with a VLM, and sort every inspection into pass or fail.
Huge improvement in long-form technical workflows
Hey @ CandyButcher , welcome to the community! Really appreciate you taking the time to share this. It’s really nice to hear that ChatGPT has felt more useful for long, iterative software projects, not just one-off prompts. The example around building a document extraction pipeline module by module is especially helpful context. This kind of real-world feedback is helpful, and I’m glad it’s been making a noticeable difference in your day-to-day work. - Sunny
Comment on Here’s To Drinking Safely With AI! by David Zinczenko
Interesting tech, but from my own experience, no gadget can stop someone who is not ready to stop. I needed full immersion away from my daily triggers, not predictions or alerts. That is why Inpatient rehab https://canadiancentreforaddictions.org/understanding-addiction/cocaine-addiction/ was the only thing that actually worked for me. These tools might help some, but real change comes from within.
How the AI bubble could pop and take down the global economy, according to the BIS
Central bank for central banks sees shades of dotcom mania in hyperscaler capex binge
Delhi EV policy to accelerate electric two wheeler adoption
The Delhi government on Monday approved a new Electric Vehicle (EV) Policy that is expected to accelerate electric two wheeler adoption through purchase incentives, mandatory EV only registrations from 2028 and a major expansion of charging infrastructure. The policy offers incentives of Rs 30,000 for electric two wheelers in the first year, Rs 20,000 in the second year and Rs 10,000 in the third year. From April 1, 2028, only electric two wheelers will be eligible for fresh registration in Delhi. Existing petrol and diesel vehicles can continue to be used as per the current rules. The move is expected to benefit electric two wheeler manufacturers such as Ather Energy, Ola Electric, TVS Motor Company, Bajaj Auto and Ultraviolette Automotive. Delhi is one of the country's largest two wheeler markets and the policy provides long term visibility for EV adoption. Reacting to the announcement, Tarun Mehta, cofounder and CEO of Ather Energy, said that Delhi has approved one of the most significant city level EV policies in India. "The combination of incentives, phased electrification mandates and charging infrastructure creates a very strong foundation. If Delhi can become a majority EV market, it has the opportunity to become a benchmark for the rest of the country," Mehta said in a post on X . He added that long term policies give the EV ecosystem the confidence to continue investments and product development. Narayan Subramaniam, CEO and Head of Design at Ultraviolette Automoti…
Pair Nova 2 Lite with Claude for cost-optimized document processing
In this post, we show how pairing Amazon Nova 2 Lite with Anthropic’s Claude Sonnet 4.6 delivers an efficient solution for digitizing scanned documents at scale. We built a two-model pipeline on Amazon Bedrock for digitizing scanned yearbook pages. Amazon Nova 2 Lite handles native multimodal extraction in a single call: detecting photos, extracting visible names with coordinates, and returning page-level metadata. Claude Sonnet 4.6 then performs spatial reasoning to match names to faces based on page layout.
After enabling advanced account security on the Codex app, phone number verification is mandatory (Case 10556951)
@openai_support Any progress today? Add verification methods or support for modifying phone numbers
I tried a Windows handheld PC, and its docking system made it my ideal travel companion
MSI's Claw 8 EX AI+ is a worthy sequel, with stronger performance, better ergonomics, and highly effective cooling.
Andy Burnham turns down invite to America’s 250th birthday bash in London
Britain’s likely next PM won’t be among the guests of the U.S. ambassador’s legendary summer party.
At $499, Apple’s M3-powered iPad Air is a good deal
Most of Apple’s price increases have gone into effect, resulting in iPads and other products costing hundreds more than they did a few days ago. If last week’s Prime Day sale wasn’t a good time to consider buying an iPad, we found a deal worth considering that’s just under $500 right now. The 128GB iPad […]
![[Linkpost]Frame Error](https://res.cloudinary.com/lesswrong-2-0/image/upload/v1654295382/new_mississippi_river_fjdmww.jpg)
[Linkpost]Frame Error
Today I want to explain what I call “ frame errors ”: a third kind of mistake, distinct from logical fallacies and empirical errors. You can get every fact right and every inference valid, and still be wrong, because your framework for reasoning about the problem is structurally inadequate. In some sense, the idea is not new. Many other thinkers have pointed out specific instantiations of these mistakes before. However, I coin the “frame error” term to draw attention to the common category across these mistakes, provide tools and worked examples to help readers spot these errors when they occur, and offer advice to help readers understand and identify the meta-errors that might lead them or others to commit frame errors. This post will go through five common classes of frame errors, with worked examples. I will start with examples that you likely have heard of before, and then end with more novel examples where I was the first person to explicitly point out such errors. Along the way, I provide conceptual tools to model good thinking, so you can hopefully learn useful rationality and reasoning tips even if you do not buy the frame error construct. [...] See more at: https://linch.substack.com/p/frame-error Discuss
After testing Thread, Zigbee, and Matter, here's how I'm building my smart home differently
All three are related to smart home connectivity, and it's critical that you understand their differences.
Multi-tenant LLM analytics with row-level security: How we built a secure agent on AWS
In this post, we show you how PAR built a production-ready multi-tenant LLM analytics system that enforces row-level security through a three-layer architecture: cryptographic request signing with AWS SigV4, semantic validation on Amazon Bedrock, and programmatic data isolation via Split-Plane SQL. We demonstrate how each layer operates independently to reduce the risk of cross-tenant data exposure, even when the LLM itself is compromised or manipulated.
Arena, the AI leaderboard everyone uses, is now a $100M business
The startup, which runs a popular free AI leaderboard, launched its commercial service just last September.
Injection Molding Defect Detection for Medical Components
Use RF-DETR and Gemini 2.5 Pro to detect defects on molded components and generate automated inspection observations.