Your curated collection of saved posts and media
I once was the best beta tester on the Adobe Acrobat 1.0 product. "Why?" I had 900 Aldus Pagemaker files from the magazine I worked at to feed into it. That now is Adobe Indesign. But now with @felo_ai_en you can go way further than I did. Import your PDFs into Felo and it sets up a business automation system with 24/7 on duty AI agent teams. Its new multi-personal collaborative office functions make your whole team better. Everyone working together on an intelligent canvas. It can autonomously evolve its knowledge containers, convert across formats intelligently, and take you from static files to dynamic knowledge hubs. Who does well in this new world? People who make decisions quickly. Of course it can build beautiful PDFs too. Meet Felo LiveDoc - the world's first 'Human + AI Agent' hybrid workspace! Try it at https://t.co/JGjBeygsuL
I asked @echen why Claude writes (and codes) so much better than other models. His answer: higher-quality training data. "Most people don't understand what quality even means in this space. They think you could just throw bodies at a problem and get good data, and that's completely wrong. Let me give you an example. Imagine you wanted to train a model to write an eight-line poem about the moon. What makes it a good poem? If you don't think deeply about quality, you'll be like, is this a poem? Does it contain eight lines? Does it contain the word moon? You check all these boxes? So then yeah, sure, you say it's a great poem. But that's completely different from what we want. We are looking for Nobel Prize-winning poetry. Is this poetry unique? Is it full of subtle imagery? Does it surprise you, and tug at your heart? Does it teach you something about the nature of moonlight? Does it play through emotions, and does it make you think? That's what we are thinking about when we think about a high-quality poem."
The fastest company in history to $1B did it with no VC money and fewer than 100 people. @HelloSurgeAI has become the secret weapon behind Anthropic and Google's best models. Founder Edwin Chen (@echen) built it without playing the Silicon Valley gameβno viral posts, no fundrais
Today, we're making Scouts available to everyone! Earlier this year, Scouts was born out of a simple observation β that so many of life's background (or even foreground!) tasks have a recurring flavor, e.g. house hunting, early stages of travel planning, sourcing leads, discovering rare products, job search, staying on top of niche news / research / podcasts, discovering local events, etc. It's been gratifying to see all the love and feedback from our closed beta users who've helped shape the product over the last few months β€οΈ With just a simple query, Scouts lets you deploy a team of AI agents to monitor anything for you. Running 24x7 in the background on the web. So you have the mental space to focus on what's most meaningful to you. The underlying agent architecture is incredibly powerful β subagents all the way down, powered by our own web navigation agent, and with access to way more tools / APIs than before. This what the future of interfacing with the web looks like. Where you're not sitting there manually browsing and refreshing, buried in tabs, ads, noise, distractions, context switches. Think Google Alerts on steroids. This has been, hands down, the most fun and challenging release from our team. As with all things agentic, there's a lot of noise out there, and every micro-decision matters in shipping reliable agents. To celebrate this release, we're offering all paid plans at 50% off, and here's a video we recently shot. Hope you like it.
Quick new post: Auto-grading decade-old Hacker News discussions with hindsight I took all the 930 frontpage Hacker News article+discussion of December 2015 and asked the GPT 5.1 Thinking API to do an in-hindsight analysis to identify the most/least prescient comments. This took ~3 hours to vibe code and ~1 hour and $60 to run. The idea was sparked by the HN article yesterday where Gemini 3 was asked to hallucinate the HN front page one decade forward. More generally: 1. in-hindsight analysis has always fascinated me as a way to train your forward prediction model so reading the results is really interesting and 2. it's worth contemplating what it looks like when LLM megaminds of the future can do this kind of work a lot cheaper, faster and better. Every single bit of information you contribute to the internet can (and probably will be) scrutinized in great detail if it is "free". Hence also my earlier tweet from a while back - "be good, future LLMs are watching". Congrats to the top 10 accounts pcwalton, tptacek, paulmd, cstross, greglindahl, moxie, hannob, 0xcde4c3db, Manishearth, and johncolanduoni - GPT 5.1 Thinking found your comments to be the most insightful and prescient of all comments of HN in December of 2015. Links: - A lot more detail in my blog post https://t.co/7LpJEVgbyk - GitHub repo of the project if you'd like to play https://t.co/WVQUbUzt2y - The actual results pages for your reading pleasure https://t.co/e2XIYElnc5
We have just used the @Nvidia H100 onboard Starcloud-1 to train the first LLM in space! We trained the nano-GPT model from Andrej @Karpathy on the complete works of Shakespeare and successfully ran inference on it. We have also run inference on a preloaded Gemma model, and we plan to try more exciting models in the future. Getting the first H100 to work in space required a lot of innovation and hard work from the incredible Starcloud team to make this breakthrough. This is a significant first step toward moving almost all computing off Earth to reduce the burden on our energy supplies and take advantage of abundant solar energy in space! π
π¨#BREAKING: The man stabbed in the chest by an illegal alien on the train in Charlotte NC, Kenyon Dobie, is in critical condition with a tube in his chest, pumping blood out of his lungs. According to authorities, Dobie was the only person who stood up and told the drunk illegal to stop screaming at riders, including an elderly woman on the train. For his efforts, he was knifed in the chest. "I guess itβs better off that it happened to me and not an older person. I wasnβt trying to be a macho man, but what I wonβt allow is you to attack random people for no reason, especially the elderly..."
STARLINK NEWS: NetJets announced an agreement with SpaceX to equip Starlink on over 600 aircraft in its global fleet by the end of 2026! Rollout starts this month targeting models including Cessna Citation Latitudes, Embraer Praetor 500s, Cessna Citation Longitudes, Bombardier Challenger 350s and 650s, and the full Bombardier Global fleet in the U.S. and Europe. Per Patrick Gallagher, NetJets Aviation president, "This agreement with Starlink positions NetJets to continue delivering exceptional travel experiences for our Owners and their guests... for business or leisure, as seamless connectivity is critical to the experience our Owners need and deserve." Per Jason Fritch, SpaceX VP of Starlink Enterprise Sales, "With Starlink onboard your NetJets flight, youβll be able to boost productivity, have seamless video calls, stream on demand, and game just as you can do on the ground. Weβre excited to level up the NetJets travel experience and maintain our commitment to deliver exceptional connectivity from 40,000 feet."
EU fined π over its blue check system even though card payments already tie verification to real identity. Meanwhile they did not fine Meta, even though it also offers paid blue checks. Why penalize π but not Meta? https://t.co/OZVUT0qkYB
-Digital art is global- Your timeline is your 24/7 art fair πππ§ππ¬π’π¬-Γ 86 Γ 166 Γ 6 cm (33.9 Γ 65.4 Γ 2.4 inches) .GIF Minted on the collectorβs preferred chain. Will be minted upon purchase. Modular LED system - P1.25 Recycled HDPE plastic, aluminum profile, steel frame. 2025 Available via OTC / DMβοΈ Detailsβ¬οΈ
Made a site comparing the sizes of living things :) The great Julius Csotonyi spent 5 months painting over 60 illustrations for the site, no ai used https://t.co/8z9InrkHVQ
"ask" and you shall receive! SemTools now ships with a dedicated "ask" CLI command - performs agentic search over documents - combine with `parse` to create QA workflows over unstructured data - cache your indexes with `workspaces` Learn more: https://t.co/3f2Mfg3xCc
introducing AI Fashion Try On > upload your outfit > choose a model > generate a UGC video live today on https://t.co/u6O0YGkBCf https://t.co/wbnVvCx3Xx
What if your help docs updated themselves⦠and an AI actually used them to answer every question 24/7? Today I am shipping AI Answers for Ferndesk. The easiest way to help customers self serve from self updating docs. https://t.co/I6e6JHUZgC
Senior @SAP consultants rated results they were told came from interns as 95% accurate - then rejected the exact same work when they learned it was generated by an #AI copilot. Read the full story on VentureBeat: https://t.co/T2CdUNZ4IR https://t.co/Ii7bRObUey
Today we started rolling out SimGym β a system that creates βdigital customersβ that behave like real ones. They browse your site, complete tasks, and reveal optimization opportunities. You can even run A/B tests with *zero* live traffic! Spent a year developing it. https://t.co/W1XYgpObcs
Even if you hate AI stuff, I'd challenge you to find ONE thing in this list that you think is valuable and kinda cool. I think they all are, and each gives someone a superpower they didn't have before. https://t.co/Uh7VX0W68p
For those wondering, this was part of the 2.0.64 release: https://t.co/nm5TXZ8UxF
Today we open source Nomos 1. At just 30B parameters, it scores 87/120 on this yearβs Putnam, one of the worldβs most prestigious math competitions. This score would rank #2/3988 in 2024 and marks our first step with @hillclimbai towards creating a SOTA AI mathematician. https://t.co/YoJxrUyJWo
Holy moly, thats insane: Nomos 1 is a 30B open-source model that just scored 87/120 on this yearβs Putnam, good enough for an estimated #2/3988, showing that nearβtop human math performance is now possible with relatively small models plus good post-training and reasoning setup. This puts a state-of-the-art AI βmathematicianβ and its full reasoning capabilites into everyoneβs hands, lowering the barrier to serious math research, verification, and advanced reasoning tools.
Today we open source Nomos 1. At just 30B parameters, it scores 87/120 on this yearβs Putnam, one of the worldβs most prestigious math competitions. This score would rank #2/3988 in 2024 and marks our first step with @hillclimbai towards creating a SOTA AI mathematician. https:/
Mistral just released Devstral 2 123B Instruct 2512 https://t.co/4cPALeG6Ju
@soul_surfer78 I see. That's a bigger topic since you also want to cost controls, security, scaling etc. @LightningAI may be a good option here. E.g., see the beginner tutorials via https://t.co/RLVKveXBkk and https://t.co/GIEVvtcGX1
The European Union is batshlit crazy. They fined X (and Elon) for blue check verification. It costs $8 per month to get verified Credit/Debit cards have identifying info, including payment address. https://t.co/zsTGAlnyge
πΊπΈ ELON: WE DONβT WANT AMERICA TO TURN INTO A COMMUNIST HELLHOLE Elon spoke bluntly about the political ripple effects of concentrated voting blocs reshaping U.S. politics in ways few are willing to talk about: βWell, letβs look at Ilhan Omar, who was literally voted into Congress by a large group of people from Somalia who are in Minnesota, which is really far from Somalia. Or Mamdani who was voted to be mayor by a majority of people who were not born in America, thatβs my understanding, at least. And then California is a big-time situation. So, I donβt know. We just donβt want to turn into a communist hellhole, basically.β Source: @KatieMiller @elonmusk
πΊπΈ ELON: THE BORDER CRISIS WAS A VOTER IMPORT PROGRAM Elon called it what he sees as the biggest ongoing scam in Washington, a mass transfer of taxpayer money to fund illegal immigration and reshape the electorate: βI guess the biggest single thing is that there are massive tra
π New Nano Banana prompt idea: your profile picture as NYC subway mural! β¨ Prompt and reference images π https://t.co/Qtbe4BI4gR
Got a bit carried away! Here's more Cities using @NanoBanana and prompted by watching Matt Wolfe @mreflow https://t.co/UWj53t445q
An old line got an extension #MetroMapMaker https://t.co/DtM8AsiMRV
I donβt even know what this is called but I need like 10 of them. https://t.co/X0P5xNjjqS
Iβm speaking Saturday at @SolanaConf about how @BAXUSCo is redefining RWA as Real World Applications, starting with whisky. Weβve built IoT hardware, decentralized pricing, and blockchain rails to make real-world apps actually work. And yesβ¦ a little mobile surprise is coming π See you at Breakpoint π₯β‘οΈ
Presentation One: β Thanks to @jito_labs and the MEV Day team for having me. Now gearing up for the debate with Marco, Miller and Rebecca on Thursday for the full @SolanaConf session. https://t.co/NEcOPEOHlS
Very few are talking about proactive agents, but they are coming! Current LLM agents wait for you to ask for help. But the best assistant anticipates what you need before you ask. Existing agents follow a reactive paradigm. Users must unlock their phone, navigate to an app, and issue explicit instructions. During a conversation about travel plans, you have to manually ask for weather updates. While shopping, you have to explicitly request price comparisons. This new research introduces ProAgent, an end-to-end proactive agent system that continuously perceives your environment through wearable sensors and delivers assistance before you ask. The key idea: instead of waiting for commands, ProAgent uses egocentric video, audio, motion, and location data from AR glasses and smartphones to anticipate user needs. An on-demand tiered perception system keeps low-cost sensors always on while activating high-cost vision only when patterns suggest assistance opportunities. When you're at a bus stop, ProAgent notices the last bus just left and offers to book an Uber. During a conversation about weekend plans, it proactively checks the weather and your calendar for conflicts. While browsing headphones in a store, it finds lower prices online and gathers reviews. Results across real-world testing with 20 participants: ProAgent achieves 33.4% higher proactive prediction accuracy, 16.8% higher tool-calling F1 score, and 1.79x lower memory usage compared to baselines. User studies show 38.9% higher satisfaction across five dimensions of proactive services. The system runs on edge devices like NVIDIA Jetson Orin with 4.5-second average latency, keeping all data local for privacy. Shifting from reactive to proactive agents reduces both physical and cognitive workload. You stop missing timely information during conversations and attention-intensive tasks. Paper: https://t.co/3zFVP5igxe Learn to build effective agents in our academy: https://t.co/zQXQt0PMbG
Full spec-driven framework with templates and common mistakes: https://t.co/HgvX7tb9UN Includes: β Spec template you can copy β 4-phase workflow β Mistakes that burn time
The founders who master this will ship their first paying feature while others are still debugging tutorial projects. Check out https://t.co/RspY4vohjW
