Luna Snow anal made this a while back, never... posted it. Luna Snow Model: Digital Hell Male: IceDev VA: MOO🩷COW Sounds: OpenNSFWshow more

if you hit a snowball like this, you're legally... required to post it to twitter don't believe me? search "luna snow rule 34"show more

Greetings from Kyiv! Today, while crossing the Dnipro River... by metro, I decided to film it for a moment. Covered with snow and frozen. And this sky! I’ll never get tired of looking for beauty.show more

I made this short animatic a while ago but... never posted it here 🙈 this scene from Keep Your Friends Close is so funny to me skdkdjfjd designs are inspired by gigi’s! #art #epicthemusical #Odysseus #fanartshow more

SIGNIFICANT SNOWFALL POTENTIAL FRIDAY NIGHT ******1ST CALL MAP****** 🎄... Merry Christmas to YOU!!! A pretty sizable snowstorm is expected across the region Friday night, with snow developing as early as Friday evening. I am very confident in a widespread 6–10" snowfall for a large portion of the Tri-State area, with lower confidence closer to NYC, Long Island, and the Jersey Shore where mixing remains a concern. At this time, I am keeping some mixing in play in central NJ/Long Island, but I’m not totally sold on how aggressive it becomes. While 4–6" looks like a very solid floor, several areas could easily boom to 6"+ if colder air locks in faster and holds longer than currently modeled. ZONE 1 – 6–10" OF SNOW This is the zone where cold air damming will be most prevalent, allowing snow to remain all snow for the duration of the event. Expect higher snow ratios, which could allow many areas to push toward double-digit totals, especially under any persistent banding. I could easily pull this zone into NYC and Nassau County as we see more model data. ZONE 2 – 4–6" OF SNOW This zone will likely start with heavy, wet snow, before transitioning to a lighter, fluffier snowfall as colder air works in. Some compaction is expected with lower snow ratios, but additional snow and mesoscale banding could still drive totals toward or above 6" in spots. ZONE 3 – 2–4" WITH MIXING This zone remains VERY uncertain. I could make a strong argument that cold air advection dominates, keeping this mostly snow and allowing totals to reach 4–6". However, low-level warm air transport on an easterly wind may introduce sleet, especially along the Jersey Shore (east of GSP) and into Suffolk County, which would eat into accumulations. I’ll be watching trends closely as this is shaping up to be a sharp cutoff storm where small shifts could have big impacts. 📺 On TV now with the latest on PIX11 Newsshow more

Made this beat in 45 mins back in 2020,... posted it and went mega viral, turned into a song, racked up over 100M streams, got synced in shows, commercials, and still gets picked up just cuz of how recognizable it is… u NEVER know which one is gonna hit.show more

🔥🚨JUST IN: Chairwoman of the House Oversight and Reform... Committee’s Task Force on the Declassification of Federal Secrets Rep. Anna Paulina Luna posted this photo of Nikola Tesla and his free energy device without a caption. It is currently the most recent and last post on her X account. This photo depicts Nikola Tesla in his laboratory in Colorado Springs around 1899, supposedly sitting reading next to his giant "magnifying transmitter" high voltage generator while the machine produced huge bolts of electricity.show more

Exactly four years ago this month, I was staying... in this beachfront apartment in the Caribbean when I lost $40,000 in the Luna crash. I had $20k staked in UST on Terra and another $20k in LUNA and eco-coins. By March 13th, my entire $40k was worth about around $1. I posted about it at the time, this isn’t new. What I didn’t fully comprehend was the knock-on effect. The whole market came crashing down. It was a genuinely dark time. Why didn't I cash out? I couldn't. All the funds where locked into a 30 day unlocking protocol. I couldn't access the coins to sell them. That moment marked the end of my COVID travels. A few weeks later, I flew back to England for the Queen’s Jubilee. Whoo-hoo! Loved it. 💕 I posted the 'don't worry' clip on Twitter a few days before the crash. The irony. Ironically, again - I lost 15k on the MANTRA OM collapse last year too. But I was lucky I had bought the OM for just a couple of 2,000 USD. Hey Ho. Life ... we live and learn.show more

🎉 GIVEAWAY TIME 🎉 I tried Pull.Fun for the... first time and pulled an absolute banger! 🔥 To celebrate, we're giving away: 🥇 Winner #1: $50 Digital Pack 🥈 Winner #2: $50 Digital Pack 🥉 Winner #3: Mystery Slab How to Enter: ✅ Follow Chris WSB ✅ Follow Pull.Fun ✅ Like this post ❤️ ✅ Repost this post 🔄 One thing I thought was really cool about PullDotFun is a feature I've never seen on another platform: 💡 Pull a legendary card? You can send it back into the pack it came from and consign it. While it's in the pack, you earn a cut of every pull. When another collector eventually pulls your card, you receive 85% of its value paid out instantly. ⏰ Winners will be picked in 48 hours. Good luck everyone! 🍀🔥show more

This is my "feel the AGI" moment: I used... GPT-5.6 Sol to train my own autocorrect model that outperforms GPT-5.6 Sol (wtf??) I have no ML background. I have no idea what I'm doing. I just kept pushing Sol until it spat out a SOTA model. And I spent $0. The motivation: Years of talking to AI have made me terrible at typing. Rather than fix my skill issue, I decided to throw more AI at it. My idea was: instead of autocorrect that interrupts my flow, I want to type fast with mistakes and have AI clean it up after. I wanted the smallest local model possible, for speed, for battery life, for science! So I decided to train my own. Inspired by Andrej Karpathy’s autoresearch, I ran Codex /goal with this setup: pick an experiment, try it, record the results to a doc, throw it out if it fails, and plan the next experiment without repeating failures. I gave a few examples that had to pass, tight latency targets, and let it run. Sol did some amazing things. First, it scanned benchmarks and shortlisted base models: Qwen 3.5, Gemma 4, Liquid LFM 2.5. It found a dataset on HuggingFace for typed text. Then it built a simulator for fingers striking a Mac keyboard, modeling the physical layout with a Gaussian distribution around each key. It simulated striking the wrong key, wrong order, fat-fingering, etc. With the models + data + simulator, it fine-tuned using MLX right on my MacBook. It had a working prototype within an hour! But accuracy was pretty poor. — Problem 1: Tokenization Sol read papers, ran tests, and identified that the tokenizer was the bottleneck. Tokenization makes typos hard for the model to see, so it memorizes mappings instead of using its language priors. Sol tried ByT5, Google’s tokenizer-free byte-level LLM. This made a big improvement, but the model is old and lacked the knowledge needed to reach Sol performance. Sol dug deeper and realized a tokenizer-free model isn’t needed; instead, it used T5Gemma, an encoder-decoder model. This can understand the input deeply before producing output, and furthermore, Sol could post-train the encoder to improve performance. This gave a much higher ceiling. — Problem 2: Loss function Now the model was correcting some typos perfectly, but ignoring most. Sol realized that standard cross-entropy loss was teaching the model to avoid edits, because the vast majority of characters in the training data were left unmodified. The fix was wild: Sol wrote a custom loss function that byte-aligns the source and target strings, uses a dynamic programming algorithm to compute the minimum edits between the two, then weights correct edits much higher than copies. After a lot of tuning, this dramatically improved accuracy. — Problem 3: Autoregression One failure mode remained: if the model made a mistake, it couldn’t backtrack. It could only predict the next token. Teaching it to “think” like a reasoning model would solve this, but would be far too slow. Sol found a beautiful solution: instead of greedily predicting the next token, beam search over all possibilities. This parallelizes the exploration instead of one linear chain-of-thought. At the end, choose the path with highest cumulative log probability. This worked great, but made the experience worse, since the user wouldn’t see progress until the whole search was done. To fix this, Sol made a clever observation: after each search step, the longest common prefix among surviving branches is guaranteed to appear in the final result, so it can be displayed immediately. As the search progresses, weaker paths are dropped and the prefix grows, so the user sees continuous progress. Sol built all this as a custom MLX pipeline that does the parallel decoding on the MacBook GPU, with just ~40ms TTFT. It’s crazy fast and entirely local. — Final eval (error reduction rate, higher is better): - Apple autocorrect: 49.66% - GPT-5.6 Luna: 82.47% - GPT-5.6 Terra: 87.64% - GPT-5.6 Sol: 90.56% - Our model (1.7B): 91.02% Final cost: - 1 quota reset (thanks Tibo) - $0 (And yes, I verified there's no cheating. In fact, we test words scrubbed from the training data to prove the model isn’t memorizing) There were a ton more details and tangents I could write about: contrastive learning, GRPO, DPO, dynamic masking, and more. Sol is a fascinating and creative model. It blew my mind so many times. Don’t let a lack of experience stop you: Sol makes AI experiments accessible to anyone!show more

When I was in high school we called blocking... with your head a “six pack” any guesses why? 😭 Watching this beach volleyball game brought it all back. This bikini model/hot mom/player absolutely launched herself at the net and crushed a spike while her teammate in the turquoise bikini went full Supergirl across the sand trying to keep the rally alive. At one point the ball came screaming back so fast it basically turned somebody’s forehead into part of the defense 💀 instant six pack. People who’ve never played think beach volleyball is all sunshine and bikinis until they actually hit the sand hard once. That sand gets in your eyes, and somehow the ball always finds the one spot you don’t want it to but you dig it out anyway. When was the last time you got six packed?show more

This Minecraft image trend is absolutely blowing up on... IG and TikTok. It's super simple and can work on any image edit model - this video (from clovr guy) uses Reve. I extracted the prompt 👇 "Convert everything except the main subject into a Minecraft block world while keeping the main subject 100% photorealistic and identical. Preserve identical main subject and background composition, framing, camera angle, perspective, lighting, shadows, and placement. Replace all surroundings with block geometry: grass, dirt, stone, trees, buildings, signs, roads, stepped terrain, snow mountain, and block clouds. Use pixel Minecraft block textures, visible cube edges, simple lighting. Ensure seamless integration so the main subject fits naturally. Keep the background concept identical only make it minecraft style look. Do not add unrelated elements."show more

On average, I earn up to $30 to $100... daily asides profit from personal trading! And I promised never to tell anyone but fuck it! This is why you don’t see me trading a prop account! This is why I can never go back to 0 even on a bad market condition! If you’re still struggling to break that phase of always going back to 0, I bet this is for you! I made a full guide on the blueprint to make this daily income from just your phone to your Bybit or crypto account daily! All you just need is 2 hrs daily from your phone! If interested, ‘‘Comment’’ full guide and I’d send it to you immediately on your dm. Must repost and be following me and @TheAttahxPrflio so I can dm. And you must join the group in my bio for proper guidance and details to learn while you earn.show more

Day 436 of documenting 0-1M Checkmate. For the last... 436 days I publicly built this store online, X was mainly the daily notebook, but i also made several videos revealing nearly everything about this store without fully doxxing. We started with 1 product and finished with 37 as we crossed this finish line. It’s not the craziest numbers, but the i hope it shows people who are just starting out, the game of consistency. While documenting this publicly, it was the complete opposite of a straight line up. Many times we ran it up, thought we were in the clear, made mistakes, got reset back to virtually zero. In fact, we actually were set back to nearly zero twice and it’s all documented somewhere in the X posts. I could have easily faked revenue to say i got here earlier, I could have easily jumped on X only posting our best days, but that’s gay and what most people do. We did not miss a single post throughout building this store. Every single day we posted. The reason I used these photos/videos instead a regular ss is because this is the exact table & chair that I made my first $ online in. This broken chair, a broken table, and this laptop that would proceed to break in the process (have a new one now) but it felt like a more full circle moment This store was in the jewlery & fashion accessories niche. Margins were healthy due to low cogs, and it cash flowed nicely but were most likely going to move on from it. Have much better plans in place but was a sick ride. Over 23,000 total orders. This was ZERO to a million. 436 days without a single day missed of posting, which also means not a single day missed of work. Despite the internet turning into a world of fake weird shit, authentic shit still exists. Gonna start dropping some archives of all the shit I filmed behind the scenes that never got posted. Real 🅿️, put a milly up on the board publicly. 🧘‍♂️show more

Tulsi just dropped the bombshell docs on Fauci’s Wuhan... money laundering, the cover up machine, and his perjury to Congress. But screw the politics for a second. I’ll never forget the grandparents who died alone in hospital rooms while their families begged at the door. The dads who couldn’t hold their wife’s hand as she took her last breath. The small businesses gutted, the dreams crushed, the kids robbed of years they’ll never get back. Millions of Americans paid with their lives, livelihoods, and sanity because “experts” like Fauci played God with taxpayer cash and then lied through their teeth. This isn’t about “owning the libs.” This is about the body count. The empty chairs at Thanksgiving. The businesses that never reopened. The American people deserve the full truth, the prosecutions, and the reckoning. Fauci helped unleash hell. Now we make damn sure it never happens again. Drop the files, Tulsi. Burn it all down.show more

Don't train the model, evolve the harness. I read... a brilliant blog post from Hugging Face where they took a frozen open model scoring 0% on a hard legal agent benchmark, left its weights alone, and let an automated loop rewrite only the code around it. That code layer is the harness, the runtime wrapper that feeds the model context, runs its tool calls, and decides when a run ends. By the time the loop finished, the system had essentially matched Sonnet 4.6 on the benchmark's headline metric, at roughly 7x lower cost per task. Zero weights changed. The gain existed because of where the model was failing. The judge only grades files saved in the right place under the exact requested filename, and the model kept doing the legal analysis correctly, then saving it under the wrong name, dropping it in a scratch folder, or never writing it at all. So the 0% was never measuring legal reasoning. It was measuring the harness. Hand-tuning that layer is slow and model-specific, so they automated it. A Claude proposer adds exactly one mechanism per iteration, and an outer loop keeps it only if it clearly beats the current best, so accepted mechanisms compound. What the loop discovered says a lot about where agents actually fail. → The biggest single gain was file handling, not intelligence. An automatic step that lands the deliverable exactly where the judge expects it beat every prompt change, with zero extra model tokens. → Code fixes transferred across models, prompt playbooks did not. The same harness lifted a smaller model from the same family by 14 points, but the tuned prompts hurt a different model family on tasks it could already finish. → The harness mattered more than anything else. Same model, same judge, same tasks, and five different harnesses scored anywhere between 3.5% and 80.1%. The gains do eventually flatten, and the remaining misses look like real capability gaps. At some point the wrapper runs out of tricks and the model has to carry the work. But the lesson holds. A benchmark score measures the model and its harness together, and until the harness is fixed, it's impossible to know which one failed. I highly recommend reading this: I also wrote a deep dive on agent harness engineering a while back, covering the orchestration loop, tools, memory, context management, and everything that turns a stateless LLM into a capable agent. The article is quoted below.show more

This week's ChatGPT feature drop - Aug 7: 1/... Rich formatting in our web composer – When you paste in emails or documents, our composer will retain the formatting; copying and pasting is so common, we should have done this a while ago! 2/ Updated model for paid users – GPT 5.6 Sol is more consistent across quick chats and deeper reasoning. You'll now get a slider that lets you choose how much thought ChatGPT puts into a response. The haptics (vibrations) on mobile slider are fun! 3/ Unlimited text messages - Free users will get GPT 5.6 Luna with unlimited text messages. More intelligence for all. Rolling out soon. 4/ Fast Android Camera – We've made it a lot faster on Android to tap "camera" in ChatGPT to take a new photo and ask a question. 5/ Voice x Files - You can now upload files and ask questions in our new ChatGPT voice experience powered by GPT-Live. Team demos were awesome this week. So much in the queue that the next few months are going to be good. Let us know what you're hoping for in the comments!show more

“Let me make this crystal clear; the narrative that... I'm "trying to get Aidan's bail revoked" is a lie.“– @TheOldMerIsDead •She claims she’d never do anything to get the bail revoked of Aidan Kearney while publicly leaking privileged attorney communications that he trusted her with. That looks a lot like a shot to sabotage his legal team, if you ask me. •The things she’s sharing have NEVER been made public. They’re private texts that he trusted her with and unreleased recordings she only has because she screen-recorded things he sent to her OVER A YEAR AGO, as if she was building a file on him in case things went South. (Proof in comments) •Sitting on private material for a year, then dumping it online for attention reeks of obsession spiraling into recklessness. •You don’t get to say you’re not trying to send him back while doing the very things that jeopardize someone’s freedom. •At this point, the denial isn’t believable—it’s disturbing. When clout matters more than moral boundaries, it’s more than evident exactly what her intent is. •We’ve all been seeing posts about Aidan potentially being arrested again before Christmas and all of a sudden, Meredith releases the Ronnie audio that was never intended for the public?!? THIS IS NO COINCIDENCE. SHE'S GUNNING FOR AIDAN'S FREEDOM. SHE WANTS HIM BACK IN JAIL. THIS. IS. SICK. 🚫show more

You’ve given this everything you have, and you feel... like you barely made it across the finish line when Trump won. You then realized that wasn’t the finish line, which has been tough to swallow. You’re exhausted. You’re drained. You’ve lost motivation. You’re questioning everything. You feel lied to. You’re angry … really angry. While I’m not in your shoes, I get it. I’ve only been at this since 2020 and it feels like so much longer than that. I can’t say for certain how I would be if I started this journey in 2017 or prior. Patience is a virtue but your patience has been tested like never before. I’m not here to blow smoke up your ass or try to convince you of anything, but I will leave you with this. You are not alone and I hope you can get back into the fight because we need you. [W]here [W]e [G]o [1] [W]e [G]o [A]llshow more