正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

- Small Tips - you can practice hit confirmable moves by set an opponent action to "random block" Hit confirmable moves example, Easy - Leroy 1+2,1+2,1 Medium - Leo d/b2,4 Hard - Jin d/f1,4

Hi! Buff Gigas Please?

40,219 subscribers

23,228 次观看 • 1 年前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Some facts about the Oktoberfest which is starting today. Waiters don’t get an hourly wage. They buy beer cheaper from the tent (~€1 less per beer mug) and keep the difference plus tips. Official average: €5,000+ for 16 days. Realistic range: €5,000–12,000. Top boxes with wine/champagne can go higher. Catch: you front €1,500–3,500 for tokens, dirndl/lederhosen, trays. Broken mugs come out of your pocket. Hard work: 10–16 hour shifts, 20k–50k steps, 10+ full Maß/mugs at once. Good money if the tent and tips hit. Not easy money. 🍻

Some facts about the Oktoberfest which is starting today. Waiters don’t get an hourly wage. They buy beer cheaper from the tent (~€1 less per beer mug) and keep the difference plus tips. Official average: €5,000+ for 16 days. Realistic range: €5,000–12,000. Top boxes with wine/champagne can go higher. Catch: you front €1,500–3,500 for tokens, dirndl/lederhosen, trays. Broken mugs come out of your pocket. Hard work: 10–16 hour shifts, 20k–50k steps, 10+ full Maß/mugs at once. Good money if the tent and tips hit. Not easy money. 🍻

Peregrino

2,531,949 次观看 • 15 天前

Seedance 2.0 Tip: This is how I use grids to give just enough freedom for AI to work with while maintaining a controlled set of moves/techniques. Great for anything that is choreographed. Workflow: Use Grok to generate a list of techniques that you can pick from. choose 9 variations. Use banana to make it into a grid. Specify that the pose in each grid should be at the moment most representative of the tenique. SD2 Prompt: An olympic skater is performing her beautiful fast paced skate choreography. Choose at least 3 moves at random from @图片2 throughout her performance. No stopping between moves. Each move should be sharp and flow smoothly into the next move. Moves must not last longer than 4 seconds each. Emphasize her speed and accuracy of execution with dramatic cuts that shows her facial expression for emotional highlights. Must include at least 1 spin and 1 jump. Camera work should include fast 360 orbits, rapid jet cuts, ECUs to when she lands her jumps, dynamic push/pulls where her skate glides right infront of the camera, or any other that you think would be fitting. Perfect performance of a worldclass skater.

Seedance 2.0 Tip: This is how I use grids to give just enough freedom for AI to work with while maintaining a controlled set of moves/techniques. Great for anything that is choreographed. Workflow: Use Grok to generate a list of techniques that you can pick from. choose 9 variations. Use banana to make it into a grid. Specify that the pose in each grid should be at the moment most representative of the tenique. SD2 Prompt: An olympic skater is performing her beautiful fast paced skate choreography. Choose at least 3 moves at random from @图片2 throughout her performance. No stopping between moves. Each move should be sharp and flow smoothly into the next move. Moves must not last longer than 4 seconds each. Emphasize her speed and accuracy of execution with dramatic cuts that shows her facial expression for emotional highlights. Must include at least 1 spin and 1 jump. Camera work should include fast 360 orbits, rapid jet cuts, ECUs to when she lands her jumps, dynamic push/pulls where her skate glides right infront of the camera, or any other that you think would be fitting. Perfect performance of a worldclass skater.

Jason W - AI

45,912 次观看 • 5 个月前

This work makes a humanoid robot do simple parkour moves by looking with a depth camera and choosing the right move on the fly. The big deal is that it turns lots of small human moves into long, real-time robot behavior, without hand-coding every transition or retraining for each new course. A humanoid robot is usually good at steady walking, but it often fails when it has to do fast moves like jumping up, vaulting, or rolling, and then keep going to the next obstacle. The hard part is that you cannot easily collect training data for every possible obstacle shape, distance, and mistake, so robots end up learning a few moves that only work in a narrow setup. This work starts from short clips of real human parkour moves, like stepping over, vaulting, climbing, and rolling. It uses motion matching, which is basically a smart “pick the next clip that fits best right now” search, to stitch those short clips into a long, smooth plan that looks like a human doing a whole course. Then it trains a controller with reinforcement learning (RL), which means the robot learns by trial and error to copy that plan while staying balanced and not falling. After training separate expert controllers for different moves, it compresses them into 1 controller that uses only onboard depth sensing and a simple “go this fast in this direction” command. In real tests on a Unitree G1 humanoid, it can clear multiple obstacles in a row, adapt when obstacles get moved, and climb a wall up to 1.25m.

This work makes a humanoid robot do simple parkour moves by looking with a depth camera and choosing the right move on the fly. The big deal is that it turns lots of small human moves into long, real-time robot behavior, without hand-coding every transition or retraining for each new course. A humanoid robot is usually good at steady walking, but it often fails when it has to do fast moves like jumping up, vaulting, or rolling, and then keep going to the next obstacle. The hard part is that you cannot easily collect training data for every possible obstacle shape, distance, and mistake, so robots end up learning a few moves that only work in a narrow setup. This work starts from short clips of real human parkour moves, like stepping over, vaulting, climbing, and rolling. It uses motion matching, which is basically a smart “pick the next clip that fits best right now” search, to stitch those short clips into a long, smooth plan that looks like a human doing a whole course. Then it trains a controller with reinforcement learning (RL), which means the robot learns by trial and error to copy that plan while staying balanced and not falling. After training separate expert controllers for different moves, it compresses them into 1 controller that uses only onboard depth sensing and a simple “go this fast in this direction” command. In real tests on a Unitree G1 humanoid, it can clear multiple obstacles in a row, adapt when obstacles get moved, and climb a wall up to 1.25m.

Rohan Paul

37,121 次观看 • 7 个月前

i spent 4 dollars to make this post Akali 3H single routing structure with positioning optimizations these range from bnb territory to psychotic, but mainly putting together as an example for combo limits / optimization adjustments on the fly. You can always 2 hit walljump / do single hit jM to reduce some length and maintain the routing if combo time is a concern. For the third and fourth combos, there are 2 versions - one that optimizes for limit strike damage / combo time, and one that sends in extra air normals for 5s2 oki. Just posted both as an example, but routing like this also helps bridge the gap on 5s2 damage if you'd like to use that. Notes: 1) Corner to corner, keep in mind extra normals will increase the pushback and leave you closer to corner by the end, reducing frame adv from the tumble. 2) Nearing corner (starting about 75% of the stage away) adjustment on the ender for when the tumble has less frame adv, just cram in more air normals and save your double jump. 3) Here's where you'll start to see corner adjustments - round start position, this ends up being the starter for max damage midscreen, but if you were to reach the corner after the apex of your second jump loop arc, you can squeeze in extra minimum damage for a notable increase. Fortunately also maximizes corner carry while pushing damage. 4) In corner route. Stupid hard wrt combo time, just posting as an example of combo limits. you can probably just opt for above, but it does get an extra min damage source in. Alternatively, learn an easier combo lol. Have similarly structured stuff for 2H, will probably get to it later.

i spent 4 dollars to make this post Akali 3H single routing structure with positioning optimizations these range from bnb territory to psychotic, but mainly putting together as an example for combo limits / optimization adjustments on the fly. You can always 2 hit walljump / do single hit jM to reduce some length and maintain the routing if combo time is a concern. For the third and fourth combos, there are 2 versions - one that optimizes for limit strike damage / combo time, and one that sends in extra air normals for 5s2 oki. Just posted both as an example, but routing like this also helps bridge the gap on 5s2 damage if you'd like to use that. Notes: 1) Corner to corner, keep in mind extra normals will increase the pushback and leave you closer to corner by the end, reducing frame adv from the tumble. 2) Nearing corner (starting about 75% of the stage away) adjustment on the ender for when the tumble has less frame adv, just cram in more air normals and save your double jump. 3) Here's where you'll start to see corner adjustments - round start position, this ends up being the starter for max damage midscreen, but if you were to reach the corner after the apex of your second jump loop arc, you can squeeze in extra minimum damage for a notable increase. Fortunately also maximizes corner carry while pushing damage. 4) In corner route. Stupid hard wrt combo time, just posting as an example of combo limits. you can probably just opt for above, but it does get an extra min damage source in. Alternatively, learn an easier combo lol. Have similarly structured stuff for 2H, will probably get to it later.

SKD

21,353 次观看 • 5 个月前

My new strain ‘Drolex’💎🔥 Drolex was crafted for the for the people who love exotic genetics and premium bud. I personally designed this unique jar to match the rarity of its 1 of 1 strain. In an era where really exceptional flower paired with a fire jar is hard to come by, Drolex OG exceeds expectations. My goal was simple: deliver not just a visually appealing jar, but exotic flower that outshines it—(no shade intended). Rooted in its authentic 1980s genetics, Drolex OG is free from the diluted, over-circulated clones of today, offering that timeless old-school OG look, aroma and smoke with every hit. After years of dedication to perfecting this craft, I’m honored to share it with you and hope you value it as much as I do🙏. A limited small-batch release will hit select LA shops next week 🙏🫡 #Drolex #ZahGod #OG #WeedTwitter #CannabisCommunity

Sensitive content

My new strain ‘Drolex’💎🔥 Drolex was crafted for the for the people who love exotic genetics and premium bud. I personally designed this unique jar to match the rarity of its 1 of 1 strain. In an era where really exceptional flower paired with a fire jar is hard to come by, Drolex OG exceeds expectations. My goal was simple: deliver not just a visually appealing jar, but exotic flower that outshines it—(no shade intended). Rooted in its authentic 1980s genetics, Drolex OG is free from the diluted, over-circulated clones of today, offering that timeless old-school OG look, aroma and smoke with every hit. After years of dedication to perfecting this craft, I’m honored to share it with you and hope you value it as much as I do🙏. A limited small-batch release will hit select LA shops next week 🙏🫡 #Drolex #ZahGod #OG #WeedTwitter #CannabisCommunity

Exotic Experts™️

46,941 次观看 • 1 年前

Hood tips: Use moon knight team up 99% of the time unless you have good awareness and can be aggressive without overextending. With mk team up: start fights with the shield in front to boost demonic energy. Use it to block off heals to tanks pushing up and MELT them. Swing angles on tanks that are slow if you see the backline playing too close. If the dps aren’t right there to immediately punish, you can usually roast both healers or the very least 1 and force the 2nd to scatter and waste resources while their frontline panics. Limit testing helps you determine how much you can get away with. (Video example of swinging past a slower dps or tank when the healers are close.) Mk shield can be thrown mid fight in support ults to give your team the advantage. Same with teams backing up if they’re grouped. Pay attention to off angle dps and don’t let them go unchallenged. You can burst longer range enemies in half demon form if you alternate fire to control the bloom. He excels close to mid range and you want to maximize your uptime. Utilize the block to get max value by being aggressive and choosing when YOU want to use it. IMO his ult is best used with your team so position near them. He’s excellent when the enemy team pops a gambit ult and/or has committed to the push. You’ll be proving bonus health to nearby allies while in your ult and if positioned and timed well the enemy will have a hard time getting away even if you don’t kill all with your ult, your team is there and has an advantage. Use his void walk either defensively or offensively, you can force them to back up if you push past the frontline and drop your shield. Hes viable in almost any situation and the only thing ive had trouble dealing with is Black Cat but there’s not much you can do with that. (Team died in the clip womp womp)

Hood tips: Use moon knight team up 99% of the time unless you have good awareness and can be aggressive without overextending. With mk team up: start fights with the shield in front to boost demonic energy. Use it to block off heals to tanks pushing up and MELT them. Swing angles on tanks that are slow if you see the backline playing too close. If the dps aren’t right there to immediately punish, you can usually roast both healers or the very least 1 and force the 2nd to scatter and waste resources while their frontline panics. Limit testing helps you determine how much you can get away with. (Video example of swinging past a slower dps or tank when the healers are close.) Mk shield can be thrown mid fight in support ults to give your team the advantage. Same with teams backing up if they’re grouped. Pay attention to off angle dps and don’t let them go unchallenged. You can burst longer range enemies in half demon form if you alternate fire to control the bloom. He excels close to mid range and you want to maximize your uptime. Utilize the block to get max value by being aggressive and choosing when YOU want to use it. IMO his ult is best used with your team so position near them. He’s excellent when the enemy team pops a gambit ult and/or has committed to the push. You’ll be proving bonus health to nearby allies while in your ult and if positioned and timed well the enemy will have a hard time getting away even if you don’t kill all with your ult, your team is there and has an advantage. Use his void walk either defensively or offensively, you can force them to back up if you push past the frontline and drop your shield. Hes viable in almost any situation and the only thing ive had trouble dealing with is Black Cat but there’s not much you can do with that. (Team died in the clip womp womp)

Boxy

47,533 次观看 • 1 个月前

April 9, 2025: #ES hits 5,000, $SPY at weekly 200 EMA. News: #Trump freezes tariffs for 75+ countries for 90 days. Price spikes. October 10, 2025: $QOQ hits $600. News: Donald J. Trump slaps 100% tariffs on China. Price drops. January 11/12, 2026: #ES hits 7,000. News: #DOJ opens criminal probe into Jerome #Powell. Every example is easy to check. A coincidence happens once. I see a pattern. News often comes exactly when algos running the market need it – and I can anticipate when, not what. My content is all about algorithmic moves. ❗️Price doesn’t have to follow the news❗️ I don’t believe in global conspiracies or tin-foil hats. I just stay calm and rational. That’s probably why I see what’s hidden, even in plain sight. I could give dozens of examples, but: • If your critical thinking is sharp – that’s enough • If not – even hundreds of examples won’t convince you I’ll explain why this happens and how the mechanism works when the time comes. Not now. For now, I want to see if this content actually matters to you. I don’t want to look crazy. Hit 200 likes – and I’ll know I can share these patterns from time to time.

April 9, 2025: #ES hits 5,000, $SPY at weekly 200 EMA. News: #Trump freezes tariffs for 75+ countries for 90 days. Price spikes. October 10, 2025: $QOQ hits $600. News: Donald J. Trump slaps 100% tariffs on China. Price drops. January 11/12, 2026: #ES hits 7,000. News: #DOJ opens criminal probe into Jerome #Powell. Every example is easy to check. A coincidence happens once. I see a pattern. News often comes exactly when algos running the market need it – and I can anticipate when, not what. My content is all about algorithmic moves. ❗️Price doesn’t have to follow the news❗️ I don’t believe in global conspiracies or tin-foil hats. I just stay calm and rational. That’s probably why I see what’s hidden, even in plain sight. I could give dozens of examples, but: • If your critical thinking is sharp – that’s enough • If not – even hundreds of examples won’t convince you I’ll explain why this happens and how the mechanism works when the time comes. Not now. For now, I want to see if this content actually matters to you. I don’t want to look crazy. Hit 200 likes – and I’ll know I can share these patterns from time to time.

Denistratos

46,030 次观看 • 8 个月前

Seedance 2.0 on FlovaAI =================== Prompt: [Reference Identity Lock] Image 1 is ONLY the main female protagonist. Her face, hairstyle, body type, and outfit must match Image 1 exactly and stay consistent for the entire video. Image 2 is ONLY a uniform reference. All four opponents wear the school uniform shown in Image 2. Never swap, merge, duplicate, or blend identities. The protagonist's identity comes ONLY from Image 1. The four opponents have NO reference images. They are defined by the text descriptions below. The four opponents must not resemble the protagonist, and they must not resemble each other. All five characters must remain clearly distinct and recognizable until the end. [Priority Order] 1. Preserve the protagonist's identity from Image 1. 2. Keep the four opponents visually distinct from her and from each other. 3. Maintain one continuous shot with no cuts. 4. Keep the classroom layout spatially consistent. 5. Make the action fast but readable and physically connected. 6. Keep the tone as a Korean school action drama, stylish but grounded. Korean school action drama classroom fight scene — 15 seconds, ONE CONTINUOUS SHOT, NO CUTS. A single uninterrupted handheld shot. No cuts, no scene transitions, no montage. The camera should feel handheld, with micro-jitters, slight rolling shutter, and raw unstable realism. The camera must physically travel through the same classroom space. Every transition must be motivated by camera movement, not editing. Whip pans are allowed, but they must not hide a cut. Do not teleport the camera or characters. The classroom layout and character positions must remain spatially consistent. Audio: No music. Only realistic school and classroom ambient sounds: old fluorescent light hum, distant hallway noise, ceiling fan, shoes scraping the floor, desks dragging, chair legs screeching, cloth friction, dull body impacts, and breathing that gradually becomes heavier. Breathing continues throughout the scene and keeps building. Lighting: Late afternoon in a Korean high school classroom. Mixed cool fluorescent light and warm sunlight through the windows. Dust floating in the sunlight. Soft fan shadows moving across desks and school uniforms. Main character: The Korean female high school student from Image 1, age 17–18. Cold, emotionless, calm, and intimidating. She barely speaks and does not scream during the fight. She remains composed from beginning to end. Her movements are efficient, explosive, and precise. Even if her frame is not large, she dominates through speed, timing, and accuracy. Main outfit: Exactly the outfit shown in Image 1. Do not change its colors, design, or details. Her jacket or outer layer is either removed and hanging on a chair, or worn in a slightly messy way. The action must be non-sexualized and combat-focused. Fabric movement, dust, sweat, wrinkles, and impact response should feel realistic. Opponent rules: Four Korean female high school students, all wearing the Hanlim Multi Art School uniform shown in Image 2. They have no reference images. Define them strictly by these descriptions and keep each one consistent: Opponent A: short black bob with straight bangs, medium build, round face. Opponent B: long straight hair tied in a high ponytail, tall and lean, sharp jawline. Opponent C: shoulder-length hair with side-swept bangs, slim build, narrow face. Opponent D: long wavy hair worn loose, slightly stocky and broad-shouldered. A, B, C, and D must each keep clearly different faces, hairstyles, body shapes, and silhouettes. They must not resemble the protagonist, and they must not resemble each other. No face duplication, no face merging, no identity confusion. Environment: An empty classroom at Hanlim Multi Art School, a Korean performing arts high school in Seoul. Green chalkboard, chalk tray, worn wooden desks, plastic chairs, classroom clock, class schedule poster, discipline/life-guidance posters, cleaning tools, blinds or curtains, wall study materials, and a slightly scuffed floor. Desks and chairs should react naturally to impacts, sliding, shaking, and collapsing when hit. Camera framing rules: Even during kicks, framing should stay around chest-level or eye-level. No low-angle shots under the skirt. Do not focus on legs, thighs, underwear, or fetish-like details. All action framing must prioritize faces, upper-body motion, impact, and spatial choreography. Continuous action and camera choreography: From 0 to 15 seconds, the fight continues without any cuts. The action should be stylish but readable, and every movement must be physically connected. 0–3s: The camera starts behind the protagonist at a slightly low handheld angle, drifting left through the classroom aisle. Opponent A grabs the protagonist's shoulder roughly and says in Korean: "야, 너 지금 뭐 하자는 거야?" The protagonist silently turns and lands one hard straight punch to A's face. At impact, use a very brief 15% slow motion: cheek ripple, dust particles, deep thud. A falls sideways into a desk. The camera dips slightly from the shock, then whip-pans right without cutting. 3–6s: Opponent B charges in from the right. The protagonist steps forward instead of retreating. A short body shot to the stomach. Immediate uppercut to the chin. Without pausing, she drives forward into a flying knee to B's chest. B is thrown backward across or into a desk. The camera follows the forward motion low, then rebounds upward with the impact. 6–9s: Opponent D attacks with two fast punches. The protagonist deflects both strikes with her arms, then flows into a turning backfist to D's face. As D staggers, she continues the same rotation into a spinning back elbow that lands hard on D's jaw or temple. D crashes sideways into two or three desks. The camera arcs around her shoulder and jitters slightly at each impact. No cuts. 9–12s: Opponent C rushes in from the chalkboard side. The protagonist clearly grabs C's collar with her left hand. C's face must be fully visible from the front and clearly different from the protagonist. The protagonist lands one short, hard punch to C's face, then immediately throws a powerful high kick or flying high kick into C's chest. The force sends C backward into the green chalkboard. The protagonist remains in the foreground and never touches the board. The protagonist's face should be side-profile or partially obscured. C's face should be clearly visible from the front at the moment of impact. Their faces must never overlap in frame. Use a very brief 20% slow motion at the chalkboard impact: chalk dust bursts outward, and C slides down the board. The camera pushes up with the impact, then tilts down as C slides. 12–15s: Through the chalk dust, the camera hard-pans right. D makes one final charge. The protagonist sidesteps and lands a tight uppercut to D's chin, followed immediately by a cross. D crashes into a row of desks, causing a chain reaction of collapsing desks and chairs. The camera drifts forward slowly. The protagonist adjusts her loose tie or ribbon and brushes chalk dust off her shoulder. Her expression stays cold and serious. She walks past the camera and exits the frame. Dust floats in the sunlight. Natural ending. =================== Made with Flova #FlovaAI #FlovaCPP

TSUBAKI

19,709 次观看 • 2 个月前

Vector Database by Hand ✍️ Vector databases are revolutionizing how we search and analyze complex data. They have become the backbone of Retrieval Augmented Generation (#RAG). How do vector databases work? [1] Given ↳ A dataset of three sentences, each has 3 words (or tokens) ↳ In practice, a dataset may contain millions or billions of sentences. The max number of tokens may be tens of thousands (e.g., 32,768 mistral-7b). Process "how are you" [2] 🟨 Word Embeddings ↳ For each word, look up corresponding word embedding vector from a table of 22 vectors, where 22 is the vocabulary size. ↳ In practice, the vocabulary size can be tens of thousands. The word embedding dimensions are in the thousands (e.g., 1024, 4096) [3] 🟩 Encoding ↳ Feed the sequence of word embeddings to an encoder to obtain a sequence of feature vectors, one per word. ↳ Here, the encoder is a simple one layer perceptron (linear layer + ReLU) ↳ In practice, the encoder is a transformer or one of its many variants. [4] 🟩 Mean Pooling ↳ Merge the sequence of feature vectors into a single vector using "mean pooling" which is to average across the columns. ↳ The result is a single vector. We often call it "text embeddings" or "sentence embeddings." ↳ Other pooling techniques are possible, such as CLS. But mean pooling is the most common. [5] 🟦 Indexing ↳ Reduce the dimensions of the text embedding vector by a projection matrix. The reduction rate is 50% (4->2). ↳ In practice, the values in this projection matrix is much more random. ↳ The purpose is similar to that of hashing, which is to obtain a short representation to allow faster comparison and retrieval. ↳ The resulting dimension-reduced index vector is saved in the vector storage. [6] Process "who are you" ↳ Repeat [2]-[5] [7] Process "who am I" ↳ Repeat [2]-[5] Now we have indexed our dataset in the vector database. [8] 🟥 Query: "am I you" ↳ Repeat [2]-[5] ↳ The result is a 2-d query vector. [9] 🟥 Dot Products ↳ Take dot product between the query vector and database vectors. They are all 2-d. ↳ The purpose is to use dot product to estimate similarity. ↳ By transposing the query vector, this step becomes a matrix multiplication. [10] 🟥 Nearest Neighbor ↳ Find the largest dot product by linear scan. ↳ The sentence with the highest dot product is "who am I" ↳ In practice, because scanning billions of vectors is slow, we use an Approximate Nearest Neighbor (ANN) algorithm like the Hierarchical Navigable Small Worlds (HNSW).

Vector Database by Hand ✍️ Vector databases are revolutionizing how we search and analyze complex data. They have become the backbone of Retrieval Augmented Generation (#RAG). How do vector databases work? [1] Given ↳ A dataset of three sentences, each has 3 words (or tokens) ↳ In practice, a dataset may contain millions or billions of sentences. The max number of tokens may be tens of thousands (e.g., 32,768 mistral-7b). Process "how are you" [2] 🟨 Word Embeddings ↳ For each word, look up corresponding word embedding vector from a table of 22 vectors, where 22 is the vocabulary size. ↳ In practice, the vocabulary size can be tens of thousands. The word embedding dimensions are in the thousands (e.g., 1024, 4096) [3] 🟩 Encoding ↳ Feed the sequence of word embeddings to an encoder to obtain a sequence of feature vectors, one per word. ↳ Here, the encoder is a simple one layer perceptron (linear layer + ReLU) ↳ In practice, the encoder is a transformer or one of its many variants. [4] 🟩 Mean Pooling ↳ Merge the sequence of feature vectors into a single vector using "mean pooling" which is to average across the columns. ↳ The result is a single vector. We often call it "text embeddings" or "sentence embeddings." ↳ Other pooling techniques are possible, such as CLS. But mean pooling is the most common. [5] 🟦 Indexing ↳ Reduce the dimensions of the text embedding vector by a projection matrix. The reduction rate is 50% (4->2). ↳ In practice, the values in this projection matrix is much more random. ↳ The purpose is similar to that of hashing, which is to obtain a short representation to allow faster comparison and retrieval. ↳ The resulting dimension-reduced index vector is saved in the vector storage. [6] Process "who are you" ↳ Repeat [2]-[5] [7] Process "who am I" ↳ Repeat [2]-[5] Now we have indexed our dataset in the vector database. [8] 🟥 Query: "am I you" ↳ Repeat [2]-[5] ↳ The result is a 2-d query vector. [9] 🟥 Dot Products ↳ Take dot product between the query vector and database vectors. They are all 2-d. ↳ The purpose is to use dot product to estimate similarity. ↳ By transposing the query vector, this step becomes a matrix multiplication. [10] 🟥 Nearest Neighbor ↳ Find the largest dot product by linear scan. ↳ The sentence with the highest dot product is "who am I" ↳ In practice, because scanning billions of vectors is slow, we use an Approximate Nearest Neighbor (ANN) algorithm like the Hierarchical Navigable Small Worlds (HNSW).

Tom Yeh

192,022 次观看 • 2 年前

Who remembers being surprised to see "Directory Art" after you typed: LOAD "$", 8 LIST Instead of a boring list of filenames and block counts, the entire directory screen transformed into a beautiful piece of directory art (dir art / d/art). The surprise hit hard the first time you saw it on a cracked game disk; I was like, "What just happened?!" The drive's "directory" isn't a normal file — when you ask for "$" (the special filename meaning "show directory"), the drive generates a fake BASIC program on the fly. The first two bytes are the load address (like any PRG file), then it spits out tokenized BASIC lines: line numbers (those 18, 40, etc.), the block counts as "PRINT" statements, filenames in quotes, file types like PRG or DEL, and finally "107 BLOCKS FREE." and "READY." as the last "lines." Normally this just looks like a messy program listing when you LIST it. But clever sceners realized: "Hey, we can overwrite those 'lines' with PETSCII block characters (like the diamond borders, checker patterns, hearts, etc.) while keeping the structure intact so the drive still thinks it's a valid directory." They used deleted files (type DEL) or carefully placed dummy entries to draw in the 16-column filename area, turning the directory into ASCII/PETSCII art. When this popped up on your C64, it was amazing. Who remembers seeing directory art on their Commodore 64?

Who remembers being surprised to see "Directory Art" after you typed: LOAD "$", 8 LIST Instead of a boring list of filenames and block counts, the entire directory screen transformed into a beautiful piece of directory art (dir art / d/art). The surprise hit hard the first time you saw it on a cracked game disk; I was like, "What just happened?!" The drive's "directory" isn't a normal file — when you ask for "$" (the special filename meaning "show directory"), the drive generates a fake BASIC program on the fly. The first two bytes are the load address (like any PRG file), then it spits out tokenized BASIC lines: line numbers (those 18, 40, etc.), the block counts as "PRINT" statements, filenames in quotes, file types like PRG or DEL, and finally "107 BLOCKS FREE." and "READY." as the last "lines." Normally this just looks like a messy program listing when you LIST it. But clever sceners realized: "Hey, we can overwrite those 'lines' with PETSCII block characters (like the diamond borders, checker patterns, hearts, etc.) while keeping the structure intact so the drive still thinks it's a valid directory." They used deleted files (type DEL) or carefully placed dummy entries to draw in the 16-column filename area, turning the directory into ASCII/PETSCII art. When this popped up on your C64, it was amazing. Who remembers seeing directory art on their Commodore 64?

Commodore Computer Museum 🕹

12,362 次观看 • 6 个月前

What if you could choreograph an entire action scene before generating a single frame? 🥋🎬 That’s what I wanted to test with Pippitofficial’s 3D Director Studio + Seedance 2.5. I created a scene with two girls demonstrating a series of karate moves and used the 3D workspace to figure out the sequence first. The interesting part is that I wasn’t relying on a giant prompt to explain every movement. I could place both characters in the scene, set up the environment, choose props from 150+ available options, and then map out how the performance should unfold. One move at a time. Step in → stance → punch → block → kick → reset → next move. The timeline makes it possible to arrange actions and control their timing, while movement trajectories help define where the characters should go. And then there’s the camera. Instead of hoping the generated video finds the right angle, I could plan the camera movement around the choreography and decide how the sequence should be captured. That’s the part I found most useful. You can think through the scene visually before asking AI to generate the final result. Characters. Movement. Timing. Camera. All planned first. Then Seedance 2.5 turns that setup into the actual video. For me, this feels like a shift from simply prompting AI to actually directing the shot. Try Pippit 3D Director Studio: #3DDirectorStudio #PippitAI #Seedance25 #PippitPartner

What if you could choreograph an entire action scene before generating a single frame? 🥋🎬 That’s what I wanted to test with Pippitofficial’s 3D Director Studio + Seedance 2.5. I created a scene with two girls demonstrating a series of karate moves and used the 3D workspace to figure out the sequence first. The interesting part is that I wasn’t relying on a giant prompt to explain every movement. I could place both characters in the scene, set up the environment, choose props from 150+ available options, and then map out how the performance should unfold. One move at a time. Step in → stance → punch → block → kick → reset → next move. The timeline makes it possible to arrange actions and control their timing, while movement trajectories help define where the characters should go. And then there’s the camera. Instead of hoping the generated video finds the right angle, I could plan the camera movement around the choreography and decide how the sequence should be captured. That’s the part I found most useful. You can think through the scene visually before asking AI to generate the final result. Characters. Movement. Timing. Camera. All planned first. Then Seedance 2.5 turns that setup into the actual video. For me, this feels like a shift from simply prompting AI to actually directing the shot. Try Pippit 3D Director Studio: #3DDirectorStudio #PippitAI #Seedance25 #PippitPartner

Shabnam Parveen

51,749 次观看 • 27 天前

A friend of mine just recently got his first gun, a striker fired Glock 19. He is terrified to leave a round chambered. To anyone reading this who has this same fear, I want to alleviate your concerns. I'm going to explain why there is nothing to be afraid of. First and foremost, I fully understand your concern. It's not irrational. And you're not stupid for being nervous about a bullet being aimed at your leg with the striker cocked back under spring tension about an 1/8 of an inch from the primer. I get it. I shared this concern when I was new to carrying. But your fear is due to a lack of understanding of the internal safety features of a modern striker fired handgun. On a properly maintained modern striker fired pistol the striker CANNOT hit the primer on the chambered round without the trigger being pulled. Can't. Not shouldn't. Can't. (If you have any good Sig P320 jokes, this is the proper place to share them) There is a block that the striker will hit and stop if it is released without the trigger being depressed. It works just like a lock and key. There is a channel in the blocker that the striker can pass through if it is in the fire position. It only goes into the fire position if the trigger is pulled which pushes the blocker into the fire position. A spring keeps the blocker firmly in the block position until then. The blocker is a solid piece of steel that won't break. And if your gun is securely in a hard sided Kydex holster nothing can move the trigger unless it is unholstered. Many guns like Glocks have a trigger safety that prevents the trigger from moving due to force of dropping. Your finger has to be on the trigger safety to move it. So the gun cannot fire unless you pull the trigger. And it's easy for you to test this safety feature to see how it works! See the video below. You can try to push the striker forward by moving that piece at the back of the slide forward. That piece is the back of the striker. Push it forward and it will stop. The tip won't come out through the breach. You see it trying to come through but it gets stuck. Now push that little button in. That is the actual blocker. When you push it in you will now be able to move the striker forward through the breach and see how it will strike the primer. You can test this easily every time you disassemble your pistol to verify it is still working. You can break this safety feature by removing the blocker or losing the spring. But if it's there it will work. (You would be the first person in firearms history who has ever lost a spring. Literally the first one. It has never happened before. 😐) But it doesn't come out as part of regular maintenance. You have to disassemble the firing assembly to get it out. That won't happen by accident. And if you do that on purpose you should do this blocker test when you reassemble it to make sure the safety feature is working. My video shows a Glock slide. All Glocks will look like this. Other brands may put the blocker in a different spot or it may be a slightly different shape. But they all (mostly) have the same design. If you test yours as demonstrated and it works like mine does, your gun WILL NOT fire without pulling the trigger. (If you just thought of another P320 joke, this is another good place to toss that out there)

A friend of mine just recently got his first gun, a striker fired Glock 19. He is terrified to leave a round chambered. To anyone reading this who has this same fear, I want to alleviate your concerns. I'm going to explain why there is nothing to be afraid of. First and foremost, I fully understand your concern. It's not irrational. And you're not stupid for being nervous about a bullet being aimed at your leg with the striker cocked back under spring tension about an 1/8 of an inch from the primer. I get it. I shared this concern when I was new to carrying. But your fear is due to a lack of understanding of the internal safety features of a modern striker fired handgun. On a properly maintained modern striker fired pistol the striker CANNOT hit the primer on the chambered round without the trigger being pulled. Can't. Not shouldn't. Can't. (If you have any good Sig P320 jokes, this is the proper place to share them) There is a block that the striker will hit and stop if it is released without the trigger being depressed. It works just like a lock and key. There is a channel in the blocker that the striker can pass through if it is in the fire position. It only goes into the fire position if the trigger is pulled which pushes the blocker into the fire position. A spring keeps the blocker firmly in the block position until then. The blocker is a solid piece of steel that won't break. And if your gun is securely in a hard sided Kydex holster nothing can move the trigger unless it is unholstered. Many guns like Glocks have a trigger safety that prevents the trigger from moving due to force of dropping. Your finger has to be on the trigger safety to move it. So the gun cannot fire unless you pull the trigger. And it's easy for you to test this safety feature to see how it works! See the video below. You can try to push the striker forward by moving that piece at the back of the slide forward. That piece is the back of the striker. Push it forward and it will stop. The tip won't come out through the breach. You see it trying to come through but it gets stuck. Now push that little button in. That is the actual blocker. When you push it in you will now be able to move the striker forward through the breach and see how it will strike the primer. You can test this easily every time you disassemble your pistol to verify it is still working. You can break this safety feature by removing the blocker or losing the spring. But if it's there it will work. (You would be the first person in firearms history who has ever lost a spring. Literally the first one. It has never happened before. 😐) But it doesn't come out as part of regular maintenance. You have to disassemble the firing assembly to get it out. That won't happen by accident. And if you do that on purpose you should do this blocker test when you reassemble it to make sure the safety feature is working. My video shows a Glock slide. All Glocks will look like this. Other brands may put the blocker in a different spot or it may be a slightly different shape. But they all (mostly) have the same design. If you test yours as demonstrated and it works like mine does, your gun WILL NOT fire without pulling the trigger. (If you just thought of another P320 joke, this is another good place to toss that out there)

Spaceballs The X Account

197,088 次观看 • 1 个月前

Polymarket. Calculator. Public data. That’s enough to beat 95% of traders How? By trading mispriced markets Liquidity is thin, so NO ONE talks about it But I'll reveal it step by step so you can win too: There're hundreds of markets where real outcome is already decided, but odds are still far from reality. Why? BECAUSE PEOPLE ARE LAZY. If you’re willing to spend 5 minutes checking numbers, you’re already ahead of ~95% of traders, trust me. Most users come to Polymarket chasing a miracle like sudden 10x, one lucky click, instant retirement. That’s not how consistent money is made. The boring stuff works. Repetition works. Simple logic works. Let’s walk through a real example. “Will MrBeast reach ___ million subscribers by March 31?” Market link: [ Now check public data: > Current number: 460M followers > Average growth = +1M subs every 6 days > Time left until March 31 ≈ 77 days 77/6 = ~13 So around 13M subscribers are left on the clock. Needed growth to hit the target: 10-12M And that’s without pricing in: > Viral uploads > New Beast Games season > Algorithmic boosts Result? A YES outcome is almost locked. At ~72% odds, you’re buying something that behaves like a near-certainty for a ~28% return. (And if you don't want to risk, you can bet on the first option to turn every $0.94 into $1) 475M line is riskier, but still logical. Even at conservative growth, the projection lands around ~473M, meaning a small push or event clears it. This isn’t gambling. This is probability abuse. And the wild part? You can find dozens of setups like this every week. All public data. All obvious. All mispriced. No edge required. Just math and patience most people don’t have. Always DYOR before trading. NFA.

Polymarket. Calculator. Public data. That’s enough to beat 95% of traders How? By trading mispriced markets Liquidity is thin, so NO ONE talks about it But I'll reveal it step by step so you can win too: There're hundreds of markets where real outcome is already decided, but odds are still far from reality. Why? BECAUSE PEOPLE ARE LAZY. If you’re willing to spend 5 minutes checking numbers, you’re already ahead of ~95% of traders, trust me. Most users come to Polymarket chasing a miracle like sudden 10x, one lucky click, instant retirement. That’s not how consistent money is made. The boring stuff works. Repetition works. Simple logic works. Let’s walk through a real example. “Will MrBeast reach ___ million subscribers by March 31?” Market link: [ Now check public data: > Current number: 460M followers > Average growth = +1M subs every 6 days > Time left until March 31 ≈ 77 days 77/6 = ~13 So around 13M subscribers are left on the clock. Needed growth to hit the target: 10-12M And that’s without pricing in: > Viral uploads > New Beast Games season > Algorithmic boosts Result? A YES outcome is almost locked. At ~72% odds, you’re buying something that behaves like a near-certainty for a ~28% return. (And if you don't want to risk, you can bet on the first option to turn every $0.94 into $1) 475M line is riskier, but still logical. Even at conservative growth, the projection lands around ~473M, meaning a small push or event clears it. This isn’t gambling. This is probability abuse. And the wild part? You can find dozens of setups like this every week. All public data. All obvious. All mispriced. No edge required. Just math and patience most people don’t have. Always DYOR before trading. NFA.

Dexter's Lab

42,865 次观看 • 8 个月前

Do mindshare, sentiment, and price really work in sync? We're looking at the case of recent $FARTCOIN moves as an example of how to read and analyze data. For most of the last 2 weeks, $FARTCOIN followed the script: uptrend/downtrend in mindshare and sentiment → price followed, as shown on the multi-metric chart at But over the last 3 days: → Sentiment climbed → Price lagged and stayed coherent with mindshare. Could this mean that if the pattern holds, price is primed to catch up? Is sentiment the first thing that climbs, followed by mindshare? If so, this could mean that projects need to focus on generating positive sentiment from current users and holders first, as they are the primary drivers of mindshare growth. Let’s break down what’s really happening with $FARTCOIN and why context is everything. Numbers and price action are only part of the story. They tell you what happened, but not why it happened or what might come next. With you can analyze the context behind the moves: By tracking how people talk, feel, and engage with a project over time along with how much influence they carry. Key Drivers for $FARTCOIN sentiment growth: #1 Buzz Highlight (Cookie Deep Research): “Fartcoin reaches milestone of 155,000 holders showing strong community growth” On the surface, this sounds bullish. But a quick scroll through the Smart Feed shows that CT’s reaction was muted. That nuance matters because not every headline creates FOMO. #2 Smart Feed Signals: Sentiment was up, but not euphoric. While the charts show a clear climb in sentiment, the Smart Feed reveals the tone was measured, not manic. Again, that subtlety matters it’s not just about sentiment rising, but how it rises. #3 Top Voices: Dip Wheeler and CRG have dominated the conversation across most timeframes. Both are $FARTCOIN maxis, this was easily discoverable by stalking their smart feeds, hence if their tone shifts, that’s a signal worth tracking. Knowing who is influencing the narrative is just as important as understanding the narrative. So where does this leave us? If the sentiment → mindshare → price pattern continues to hold, we might see price play catch-up soon. But right now, the Smart Feed doesn’t show full conviction. That’s why looking at price alone isn’t enough. To understand what’s really going on, you need to go deeper into the sentiment, mindshare, and influence data that shapes market moves. Everyone deserves access to market context. That’s what decentralized InfoFi delivers. Disclaimer: This post is for informational purposes only. Always DYOR.

Do mindshare, sentiment, and price really work in sync? We're looking at the case of recent $FARTCOIN moves as an example of how to read and analyze data. For most of the last 2 weeks, $FARTCOIN followed the script: uptrend/downtrend in mindshare and sentiment → price followed, as shown on the multi-metric chart at But over the last 3 days: → Sentiment climbed → Price lagged and stayed coherent with mindshare. Could this mean that if the pattern holds, price is primed to catch up? Is sentiment the first thing that climbs, followed by mindshare? If so, this could mean that projects need to focus on generating positive sentiment from current users and holders first, as they are the primary drivers of mindshare growth. Let’s break down what’s really happening with $FARTCOIN and why context is everything. Numbers and price action are only part of the story. They tell you what happened, but not why it happened or what might come next. With you can analyze the context behind the moves: By tracking how people talk, feel, and engage with a project over time along with how much influence they carry. Key Drivers for $FARTCOIN sentiment growth: #1 Buzz Highlight (Cookie Deep Research): “Fartcoin reaches milestone of 155,000 holders showing strong community growth” On the surface, this sounds bullish. But a quick scroll through the Smart Feed shows that CT’s reaction was muted. That nuance matters because not every headline creates FOMO. #2 Smart Feed Signals: Sentiment was up, but not euphoric. While the charts show a clear climb in sentiment, the Smart Feed reveals the tone was measured, not manic. Again, that subtlety matters it’s not just about sentiment rising, but how it rises. #3 Top Voices: Dip Wheeler and CRG have dominated the conversation across most timeframes. Both are $FARTCOIN maxis, this was easily discoverable by stalking their smart feeds, hence if their tone shifts, that’s a signal worth tracking. Knowing who is influencing the narrative is just as important as understanding the narrative. So where does this leave us? If the sentiment → mindshare → price pattern continues to hold, we might see price play catch-up soon. But right now, the Smart Feed doesn’t show full conviction. That’s why looking at price alone isn’t enough. To understand what’s really going on, you need to go deeper into the sentiment, mindshare, and influence data that shapes market moves. Everyone deserves access to market context. That’s what decentralized InfoFi delivers. Disclaimer: This post is for informational purposes only. Always DYOR.

Cookie DAO 🍪

54,184 次观看 • 1 年前

$Softmax vs Sigmoid ✍️ Interact 👉 = Softmax = Softmax is how deep networks turn raw scores into a probability distribution — the final layer of every classifier, and the core of every attention head in a transformer. To see what it does, picture five boba tea shops on the same block, all competing for your dollar. Five candidates: a, b, c, d, e — different chains, different brewing styles, different pearls. A boba reviewer hands you a 𝘤𝘩𝘦𝘸𝘪𝘯𝘦𝘴𝘴 𝘴𝘤𝘰𝘳𝘦 for each — higher means perfectly chewy "QQ" pearls with the right bite (ask a Taiwanese friend to find out what QQ means). Negative scores are real: mushy bobas, overcooked pearls, a batch left sitting too long. How do you turn five chewiness scores into an allocation that adds to a whole dollar? You could spend everything at the chewiest shop, but that ignores how good the runners-up are. Softmax is the smooth alternative. Read the diagram left to right. First, raise each score to e^{x} — this does two things: it turns negative chewiness into small positives, and it stretches the gaps between scores exponentially. Then sum all five into a single total Z. Finally, divide each e^{x} by Z to get a probability. The five probabilities add up to one, so you can read them as percentages of your dollar. The chewiest shop gets the biggest slice — but never the whole dollar. That's the point of softmax: it ranks confidently while still leaving room for the others. = Sigmoid = Sigmoid squashes any real number into a probability between 0 and 1 — the classic activation for binary classification, and still the gating function inside LSTMs and GRUs. Same boba block as the previous Softmax example, narrowed to just two contenders — a hot new shop `a` with chewiness score x, and your usual go-to `b` whose score is pinned at zero (the neutral baseline you've come to expect). Sigmoid is just softmax with two players, one of them pinned to zero. Read the diagram left to right. First, raise each score to e^{x} — for the usual shop `b` whose score is zero, this is just e^0 = 1 (the constant baseline). Then sum the two into a total Z. Finally, divide each e^{x} by Z to get a probability. The two probabilities add up to one — the new shop wins more of your dollar when its pearls get chewier, and your usual keeps the rest. That's the point of sigmoid: it turns a single chewiness score into a clean 0-to-1 chance you'll try the new place over your usual. --- AI Math, Algorithms, Architectures by hand ✍️ Subscribe to my 60K+ reader newsletter 👉$

Softmax vs Sigmoid ✍️ Interact 👉 = Softmax = Softmax is how deep networks turn raw scores into a probability distribution — the final layer of every classifier, and the core of every attention head in a transformer. To see what it does, picture five boba tea shops on the same block, all competing for your dollar. Five candidates: a, b, c, d, e — different chains, different brewing styles, different pearls. A boba reviewer hands you a 𝘤𝘩𝘦𝘸𝘪𝘯𝘦𝘴𝘴 𝘴𝘤𝘰𝘳𝘦 for each — higher means perfectly chewy "QQ" pearls with the right bite (ask a Taiwanese friend to find out what QQ means). Negative scores are real: mushy bobas, overcooked pearls, a batch left sitting too long. How do you turn five chewiness scores into an allocation that adds to a whole dollar? You could spend everything at the chewiest shop, but that ignores how good the runners-up are. Softmax is the smooth alternative. Read the diagram left to right. First, raise each score to e^{x} — this does two things: it turns negative chewiness into small positives, and it stretches the gaps between scores exponentially. Then sum all five into a single total Z. Finally, divide each e^{x} by Z to get a probability. The five probabilities add up to one, so you can read them as percentages of your dollar. The chewiest shop gets the biggest slice — but never the whole dollar. That's the point of softmax: it ranks confidently while still leaving room for the others. = Sigmoid = Sigmoid squashes any real number into a probability between 0 and 1 — the classic activation for binary classification, and still the gating function inside LSTMs and GRUs. Same boba block as the previous Softmax example, narrowed to just two contenders — a hot new shop `a` with chewiness score x, and your usual go-to `b` whose score is pinned at zero (the neutral baseline you've come to expect). Sigmoid is just softmax with two players, one of them pinned to zero. Read the diagram left to right. First, raise each score to e^{x} — for the usual shop `b` whose score is zero, this is just e^0 = 1 (the constant baseline). Then sum the two into a total Z. Finally, divide each e^{x} by Z to get a probability. The two probabilities add up to one — the new shop wins more of your dollar when its pearls get chewier, and your usual keeps the rest. That's the point of sigmoid: it turns a single chewiness score into a clean 0-to-1 chance you'll try the new place over your usual. --- AI Math, Algorithms, Architectures by hand ✍️ Subscribe to my 60K+ reader newsletter 👉

Tom Yeh

74,503 次观看 • 5 个月前

RHP/1B Trace Phillips (Middle Tennessee Baseball) is an interesting draft-eligible sophomore on which to keep close tabs. Last spring, he hit .304/.360/.547 with 8 2B, 13 HR and 47 RBIs, but it’s his ability on the mound that's the root of his buzz. Phillips' back of the baseball card numbers last season are modest, but there is more to it than meets the eye. At 6'3" and 185-lbs, Phillips has an athletic, high-waisted frame. In the box, he has a simple setup with an ear-high handset and the bat horizontal above his back shoulder. Phillips has a slight load in which he has a noticeable barrel tip and a small stride. He moves well in the box and has some scissor action with his back leg. Phillips’ pure hit tool is fringey, but he has plus raw power and has shown the ability to drive the baseball to all fields. Max EV last spring of 109.4, if he's able to lift the ball more it will lead to a more impressive power output. Phillips’ current prospect status is largely driven by his ability on the mound. His numbers were unimpressive last season, but he took a step forward this fall. There’s some low hanging fruit to clean up in his delivery, but he has a short, whippy arm stroke and attacks from a low-three quarters slot with plenty of arm speed. Phillips’ FB has been up to 96 with ride through the zone (particular life at the top), though his best pitch is a high-70s-to-low-80s CH. He does a nice job of killing spin on it, and It averaged almost 15 MPH off his fastball this fall. It’s a plus pitch that flashes both fade to the arm side as well late tumble. Phillips completes his arsenal with a mid-80s slider that flashed above-average this fall with sharp, two-plane tilt. Both his changeup and slider profile as legitimate swing-and-miss offerings. Another positive development with Phillips this fall is that his strike-throwing took a step in the right direction. Mid-4th-6th round type this July. (📽️: Middle Tennessee Baseball)

Peter Flaherty III

10,984 次观看 • 1 年前

Let me ask you an honest question: Who’s your daddy? If you’re a high school student athlete in Illinois, the answer is easy when it comes to one’s “Right to Play.” It’s not your Mom or Dad. It’s a private third-party vendor called the Illinois High School Association (IHSA). If you play a high school sport, the IHSA controls you even when your school day is over and even more when school isn’t even in session. It’s long been an overreach of control from a small group of narrow-minded ego-driven adults literally dictating what student athletes and families can do in their free time. But much-needed and welcome change is getting closer for all student athletes and their families, as more and more student athletes and parents come forward from present and past to speak up and talk about the disadvantages the IHSA - and those who benefit from it - have and are causing. The first victory earlier this week was by slaughter rule, a 12-0 thumping over the IHSA in committee vote. Now it moves to the big leagues. You can read more here: I encourage you to engage and educate yourself in a topic that 75 percent of high school student athletes across the U.S. can do, but it’s something high school student athletes in Illinois can’t do. Over the next 30 days, you are going to see more stories and hear directly from student athletes, families, leading athletic trainers and doctors, school leaders from Illinois and other states on local and national media outlets talking about this overreach and how the IHSA continues to put student athletes at a disadvantage. So I ask you again: Since when does a private third-party vendor get to control what student athletes and their families do, especially during their free time? Think about that. As you’ll see once again, it’s another classic case of cherry-picking to benefit one’s self by an organization that has now lost three consecutive court cases due to its unlawful dictating overreach. As always, I’ll layout both sides and show you how this has and continues to work easily and well for high school student athletes and schools across the U.S. - except in Illinois. They’ll say “health & safety.” They’ll say “scheduling & other logistics.” I’ll show you how those stocked talking points aren’t even tired, “they’re outdated” as one athletic trainer and other athletic director said to me today. Change for adults who control is always hard. It’s time to do what is right and long overdue for high school student athletes and families in Illinois. You’ll see why loud and clear beginning this week. **This is the first in a daily series.**

Let me ask you an honest question: Who’s your daddy? If you’re a high school student athlete in Illinois, the answer is easy when it comes to one’s “Right to Play.” It’s not your Mom or Dad. It’s a private third-party vendor called the Illinois High School Association (IHSA). If you play a high school sport, the IHSA controls you even when your school day is over and even more when school isn’t even in session. It’s long been an overreach of control from a small group of narrow-minded ego-driven adults literally dictating what student athletes and families can do in their free time. But much-needed and welcome change is getting closer for all student athletes and their families, as more and more student athletes and parents come forward from present and past to speak up and talk about the disadvantages the IHSA - and those who benefit from it - have and are causing. The first victory earlier this week was by slaughter rule, a 12-0 thumping over the IHSA in committee vote. Now it moves to the big leagues. You can read more here: I encourage you to engage and educate yourself in a topic that 75 percent of high school student athletes across the U.S. can do, but it’s something high school student athletes in Illinois can’t do. Over the next 30 days, you are going to see more stories and hear directly from student athletes, families, leading athletic trainers and doctors, school leaders from Illinois and other states on local and national media outlets talking about this overreach and how the IHSA continues to put student athletes at a disadvantage. So I ask you again: Since when does a private third-party vendor get to control what student athletes and their families do, especially during their free time? Think about that. As you’ll see once again, it’s another classic case of cherry-picking to benefit one’s self by an organization that has now lost three consecutive court cases due to its unlawful dictating overreach. As always, I’ll layout both sides and show you how this has and continues to work easily and well for high school student athletes and schools across the U.S. - except in Illinois. They’ll say “health & safety.” They’ll say “scheduling & other logistics.” I’ll show you how those stocked talking points aren’t even tired, “they’re outdated” as one athletic trainer and other athletic director said to me today. Change for adults who control is always hard. It’s time to do what is right and long overdue for high school student athletes and families in Illinois. You’ll see why loud and clear beginning this week. This is the first in a daily series.

Trost

11,116 次观看 • 1 年前

$I genuinely think the Terafab is going to end up being one of the biggest moves ever made in human history to secure the future of AI... and I think most people still don’t fully see what Elon is trying to do here. The signs are clear to me. This is Tesla, xAI, and SpaceX essentially hinting to us that they are not going to wait on the world to give them the compute the team needs. They are going to build it themselves at a scale no one has ever attempted. When you really break it down, it gets a bit nutty. This is going to be a fully vertically integrated chip factory that will be producing over 1 terawatt of AI compute per year. This is NEXT LEVEL BIG. Today, AI is limited by chips. You can have the best models, the best engineers, the best everything... but if you don’t have enough compute, you will eventually hit a wall. Elon told us, the world can only supply a tiny fraction of the chips his companies will need. So this is the solution. Terafab puts everything under one roof like design, manufacturing, memory, packaging, testing, which means that they can build chips very fast.. like really fast. I'm talking about 100-200 billion custom AI chips per year at full capacity. Chips designed specifically for: • Tesla cars and Optimus robots • xAI models • Space-based compute You see, while other companies and CEOs are thinking Earth, Elon is planning for AI in space. Around ~80% of the compute is expected to go orbital, powered by solar energy bc Earth simply doesn’t have enough electricity. The U.S. grid is only about ~0.5 terawatts, while space has basically UNLIMITED energy if you can capture it. And this is the steps to get it: Starship launches → space compute → solar-powered AI → feeds back into everything to Earth. Bro... Elon and his companies are playing at a whole different level... And this is why I keep telling people that the Terafab is going to be the secret ingredient that will be the real unlock for everything: • Robotaxis at scale • Billions of Optimus robots • Massive AI models running 24/7 • Future off-world, other planet infrastructure Without these chips, none of this can happen... but with the Terafab, all of this becomes possible. That’s why Elon is calling it “the final missing piece.” I agree.$

I genuinely think the Terafab is going to end up being one of the biggest moves ever made in human history to secure the future of AI... and I think most people still don’t fully see what Elon is trying to do here. The signs are clear to me. This is Tesla, xAI, and SpaceX essentially hinting to us that they are not going to wait on the world to give them the compute the team needs. They are going to build it themselves at a scale no one has ever attempted. When you really break it down, it gets a bit nutty. This is going to be a fully vertically integrated chip factory that will be producing over 1 terawatt of AI compute per year. This is NEXT LEVEL BIG. Today, AI is limited by chips. You can have the best models, the best engineers, the best everything... but if you don’t have enough compute, you will eventually hit a wall. Elon told us, the world can only supply a tiny fraction of the chips his companies will need. So this is the solution. Terafab puts everything under one roof like design, manufacturing, memory, packaging, testing, which means that they can build chips very fast.. like really fast. I'm talking about 100-200 billion custom AI chips per year at full capacity. Chips designed specifically for: • Tesla cars and Optimus robots • xAI models • Space-based compute You see, while other companies and CEOs are thinking Earth, Elon is planning for AI in space. Around ~80% of the compute is expected to go orbital, powered by solar energy bc Earth simply doesn’t have enough electricity. The U.S. grid is only about ~0.5 terawatts, while space has basically UNLIMITED energy if you can capture it. And this is the steps to get it: Starship launches → space compute → solar-powered AI → feeds back into everything to Earth. Bro... Elon and his companies are playing at a whole different level... And this is why I keep telling people that the Terafab is going to be the secret ingredient that will be the real unlock for everything: • Robotaxis at scale • Billions of Optimus robots • Massive AI models running 24/7 • Future off-world, other planet infrastructure Without these chips, none of this can happen... but with the Terafab, all of this becomes possible. That’s why Elon is calling it “the final missing piece.” I agree.

Teslaconomics

25,494 次观看 • 6 个月前