I'm working on a new project (I hope to release it soon; I hope it will be great!) and need to build it for Windows. I don't want to install Windows on my computers, and modern Win works terribly in VirtualBox. So, I bought this used Dell OptiPlex 3070. This compact 2020 computer has a Core i5-9500T/16GB of RAM and a 256GB NVMe drive with clean Windows 11 pre-installed (with a local account!); the license key is flashed to the BIOS. I really like this machine; it's compact, quiet, and works really well as a build server for my project. I was quite surprised by the nice integration of SSH and other convenient tools on modern Windows. I'm working with this machine remotely almost as if it were a Linux server.

i must be a special kind of idiot. got a uni assignment to install any type of linux distro on my laptop ‘either’ via some sort of virtualization so i spent hours trying to install ubuntu on QEMU/KVM using Omarchy Quattro, debugging why the wifi isn’t working like an idiot. if you read it and instantly found why im r3tarded then congratulations, you’re better than me. if you missed it then here: > i am already on linux. > could’ve just fastfetch, screenshot and submit. > i cant submit anymore. > im cooked. punching the air right now cuz what the fück is wrong with QEMU/KVM and their NAT network setup bs. linux networking is not fun :( virtual box wouldn’t do this to me.

Dennis Liu (@virtualbacon) says he hasn't found a single person who thinks Bitcoin hasn't bottomed. "I've talked to a bunch of people who aren't even fully sold on the cycle being two years." "But even if they say we have a six-month rally, they all agree it's bottomed." "I have not heard anyone say otherwise." "Going from 83 to 53 would look so weird."

"I sold my Robinhood coins to buy Quantus." 3am, middle of the night. Two days ago. @virtualbacon is awake, watching one thing: the moment @QuantusNetwork order books go live. When they did, he sold his Robinhood coins and rotated straight in. "I've been waiting for this thing forever." The safe trade wasn't enough for him. Real liquidity was the signal he'd been holding out for. "Now the books are actually there."

Arthur Hayes: Hyperliquid Is the Best DEX, but Not Necessarily the Most Profitable Investment Flop Labs CEO Arthur Hayes @CryptoHayes said in a September 12 interview with VirtualBacon @virtualbacon that Hyperliquid is the best-run decentralized exchange today, and he believes the project will continue to grow.However, from an investment-return perspective, Hayes said Hyperliquid is already very large, making another 5x gain much harder. In his view, Hyperliquid may be the best DEX, but not necessarily the most profitable investment. By comparison, lower-market-cap tokens that have fallen more sharply may require much less additional market value to deliver higher multiples. WuBlockchain note: Arthur Hayes has a relatively poor historical track record on market forecasts and frequently adjusts his views. His analytical framework may be more useful to follow than his specific price targets or timing calls.

The SOTF Conference is virtual this year. Worried you might miss a session or two? Have no fear - sessions will be recorded and made available to all registered attendees after the conference! #registertoday# #seeyouonline# #virtualconference# #SOTF2021#

ICYMI: We had a blast at @token2049 Singapore. But, it almost didn't happen… @andyyy was rerouted twice, got stuck in Dubai, and finally made it to Singapore. Landed Wednesday 2pm and walked straight into the show with Tarun. The thing is, Andy had all the production gear from The Tower in NYC. So… @OddMaxxing jumped in last minute like he'd planned it for months. He flew 30 hours to deliver the gear, telling his gf he’d “fill her in later” and proceeded to eat chicken legs in Singapore. Think he told her about that? @robbieklages carried the team all week. His rule: "If you're not having fun, you're not doing it right." The haze was thick. We all smoked a pack of cigs for every hour outside and the timeline was in shambles (typical conference dump). We did a full Quantum + Privacy Day with @NEARProtocol & @QuantusNetwork OPEN with @monad And of course, @blockworksDAS with @alpacaHQ Some of our fav quotes: @ilblackdragon on quantum: "We've designed NEAR with quantum in mind in 2018." @EliBenSasson: "If Ethereum does not move fast enough, and I'm not sure it will." @andyyy: "I know it's fear mongering, but this is a real threat." @virtualbacon: "I sold my Robinhood coins to buy Quantus." On tokenization: @chriseyin: "Today the vast majority of dollars are just sitting around doing nothing." @tomhschmidt on Pearl: "It's a block reward." @CryptoHayes: "AI agents will run all of your financial services." Plus @keoneHD, @arush, @vnovakovski, @FTDA_US and @StaniKulechov joined the show. Grayscale & SWIFT, as well. Did someone say Rollup institutional? What we took home: Institutions are here. More suits than hoodies. Tokenized equities and t-bills are the first inning of the Tokenization Supercycle. Quantum was the surprise of the week. Revenue dominates the conversations. Same week, @citrini dropped Breaking The Wall. The tokenization supercycle isn't a pitch deck anymore. It’s time to buy the f*cking tokens, Arthur says… People kept calling us the most bullish people on the timeline. @aave: "Be more bullish." @TimHaldorsson: "The best media in Token2049 by far." Then Verstappen won Singapore start to finish. Very on brand. Thank you: @NEARProtocol for co-hosting Privacy Day and tokenizing the Nasdaq. Busy week. @QuantusNetwork for making a whole conference google "post-quantum." @monad for starting our week at full speed. Literally. @alpacaHQ for tokenizing the world & allowing us to be a Patagonian farmer alongside you all. Be more bullish. See you next year (in NYC, of course). @token2049 @andyyy @robbieklages

🗺️ Even frontier GPT-5 succeeds on just 14.4% of real-world spatial tasks. A new benchmark goes beyond staring at a static image and exposes how weak AI agents still are at active spatial reasoning. Title: SpatialWorld: Benchmarking Interactive Spatial Reasoning of Multimodal Agents in Real-World Tasks URL: 📝 Overview SpatialWorld measures whether multimodal LLMs can solve tasks by actively exploring 3D environments from a vision-only, egocentric viewpoint. It unifies eight different simulators across indoor, outdoor, and digital-game settings under a shared protocol, and evaluates 15 frontier models on 760 human-annotated tasks. The agent gets no prior map and no reference solution; it has to look, move, and decide on its own. ❓ Challenges Solved Prior spatial-reasoning benchmarks relied on passive evaluation via static VQA or pre-recorded video. That can't capture the interactive spatial understanding the real world demands, where an agent must move its own viewpoint to gather visual evidence and replan on the fly under partial observability. There was a large gap between recognizing a static scene and actually moving through an unfamiliar space to get a task done. 💡 Methodology & Proposed Approach ・The task is framed as a vision-only POMDP (Partially Observable Markov Decision Process) ・The agent receives only a natural-language goal and a single native-resolution egocentric RGB image, with no depth, maps, or semantic metadata ・Actions are issued through a high-level text interface covering navigation, viewpoint control, object interaction, and task completion ・It integrates eight backends: indoor (AI2-THOR, ProcTHOR, VirtualHome), outdoor (CARLA, EmbodiedCity), and digital games (Block3D, Snake3D, Rubik's Cube) ・Success is judged by whether the final terminal state satisfies the goal, not by matching the trajectory, and is validated by human annotators ・Beyond success rate, it measures step efficiency against human reference trajectories to surface inefficient behavior 🎯 Use Cases It offers a unified, fair way to evaluate the spatial abilities of home robots and autonomous agents before real-world deployment. It can systematically diagnose where long-horizon tasks that combine navigation and manipulation break down, serving as a rigorous testbed for improving spatial-reasoning models. 📊 Experimental Results ・Across 15 frontier models, physical-task success was 14.4% for GPT-5, 12.2% for Qwen-3.5-397B, 9.2% for Gemini-3.1-Pro, and 9.2% for Kimi-K2.5 ・On digital games, Gemini-3.1-Pro led at 39.0%, followed by GPT-5 at 36.4% ・By complexity, interaction-only tasks averaged 50.2%, navigation-only dropped to 8.6%, and combined navigation-and-interaction collapsed to just 4.2% ・Models with similar success rates showed very different efficiency scores, revealing heavy reliance on trial-and-error exploration ・Model rankings shifted dramatically across environments, with no single model dominating every category #AIAgents# #SpatialReasoning#