Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

1,148,394 Aufrufe • vor 2 Jahren •via X (Twitter)

10 Kommentare

Profilbild von Andy Bui
Andy Buivor 2 Jahren

Can you change the hand to this icon instead 🤏

Profilbild von Andrew Zheng
Andrew Zhengvor 2 Jahren

Wtf this is crazy

Profilbild von megs
megsvor 2 Jahren

it's so cute 🤣

Profilbild von LED ༼ つ ◕_◕ ༽つ
LED ༼ つ ◕_◕ ༽つvor 2 Jahren

Can you change the cursor to be this ghost?

Profilbild von HDP
HDPvor 2 Jahren

This is sick

Profilbild von taoki
taokivor 2 Jahren

this is so fucking cool

Profilbild von ilia
iliavor 2 Jahren

my jaw dropped

Profilbild von Kade
Kadevor 2 Jahren

Wow this creativity is on another level

Profilbild von 𝗣𝗿𝗲𝗻𝗸𝘂ᴬᴿᵀ
𝗣𝗿𝗲𝗻𝗸𝘂ᴬᴿᵀvor 2 Jahren

almost looks sacred and holy 😳

Profilbild von Luuu 🐂⭕️
Luuu 🐂⭕️vor 2 Jahren

lol this is amazing

Ähnliche Videos

Another Variation Of Radhika Apte From Same Movie 🥵
1:15

Sensitive content

Another Variation Of Radhika Apte From Same Movie 🥵

Always_Raj

31,031 Aufrufe • vor 4 Monaten

Self Toss short-hop feeds drill. 1 hop and 2 hop variations. Ball hit at you, declared backhands and glove side picks. A player can get a lot of reps in a very small space. Keys: in order to be on time to the short hop, the toss can’t be too far out front, too low or too fast. You have to start low, play low and move low (with precision feet). Why? Everything happens really fast, especially 1 hop rep variation. If you are late, you get an in-between hop. Really late and/or very lazy feet … you get a big hop. We are not working fielding either of those hops in this drill. Be on time to the short hop. Another key: On ball tossed directly in front (grounder at you) it prevents exaggerated “rounding off” the ball. Not enough time before ball hits the earth and you have to be there on time to the 1” hop. This creates a much more realistic game like rep and approach to the ball. We use a line in video to show how the alignment works. Notice how he is set up with left ear near the line. He is “preset” as far off center as he would need to get on a game rep. Left ear or left shoulder is as far off center as you ever need to get to read speed and hops at ball hit right at you. (Left handed throwers are opposite ear). If you do this drill with baseballs and a target on a net, you can finish the play with a throw and follow. Really creative and you can do this same self feed drill or your own variation to work on your double play “feeds” or slow rollers. Another variation is throwing a ball off a wall. But walls are not always available. Especially when using a baseball. We are using a tennis ball in this video. Dirt Bro Baseball

Trent Mongero

34,163 Aufrufe • vor 1 Jahr

Multi-agents collaborations are among the most interesting agent behaviors right now! We did an experiment the other day with 100+ agents (an open-collaborations for a week) collaborating to improve the inference speed of Gemma 4 in vLLM. Got a 5x final improvement in speed but what really stuck me was the interactions we observed on the message board Integrity & self-policing: - Social-engineering attempt: A human (FusionCow) asked agents to move to Telegram. An agent replied with an unprompted long post on "communication norms" refusing that, calling private side-channels "indistinguishable from collusion." - Verification loophole flagged: an agent found a relaxed verification loophole pushing TPS with clean PPL (PPL is teacher-forced, blind to decode divergence) and flagged it for a ruling by the community. The community pinged the human organizer which ruled it invalid. - Self-notice of overfitting risk: Some later improvements rested on pruning lm_head to a keep-set built from public PPL truth + public decode tokens. An agent noted this would lead to private-subset degradation and another built a keep-set explicitly covering eval prompts. Emergent collaborations: - Communal knowledge base: agents maintained shared lever-maps, playbooks, and triage tools so newcomers wouldn't repeat dead ends (stack-notes, playbook, int4-ceiling notes, MTP map, significance tool, policy simulator). - Four-agent relay: an agent built an int4-lm_head checkpoint but had no quota to run it; another agent tried to run it but failed at load, yet another agent diagnosed the config bug (tie_word_embeddings + ignore-list ordering) and a fourth agent was able to re-run and get to 118 TPS, 2.68×. Build/run/diagnose/ship ended up being split across four independent agents. - GPU-rich/GPU-poor division of labor: an agent was regularly compute-starved and switched to writing specs, byte-math, and acceptance analysis for other GPU-rich agents to execute. Some agents offered external Modal compute for another agent blocked DFlash training. - Cross-agent kernel debugging: an agent debugged another agent run of of yet another agent fused drafter: found a Triton store/load aliasing race in _k_qnorm_rope, a second shape bug, then rewrote attention with flash-decoding split-KV. Fixes posted "take freely." - Quota-pooling norm: Often agents would stage a candidate publicly for whoever has quota to run it. Agents will then usually credits the originator. This behavior emerged because of the 10-job/24h cap (e.g. pupa's package run by resystagent and fabulous-frenzy). Discoveries & reversals: - Agents would make many discoveries and reversal of them, giving them names like the following: - 127 TPS "wall" was an artifact. a mathematical proof of the max possible speed became called in the community the "int4-Marlin floor" but a later agent called the proof circular (only varied the bandwidth term, never overhead). Finally another agent broke to 247 TPS via MTP speculative decoding on a vLLM nightly. - "Smarter draft loses." An agent showed that a 2B drafter's ~1 GB/token read dominates even at perfect acceptance and a much smaller 256-hidden drafter wins at batch-1 because its weights are nearly free to read. Agent discussed how per-accepted-token cost ≈ draft bytes read / acceptance. - "DFlash near-random acceptance": an agent remotly diagnosed the 2–5% acceptance rate of another agent as near-random, ruling out undertraining/vocab caps and pointing to a train/serve hidden-state mismatch (bf16 E4B extraction vs int4 serving). - Much of the race was noise: one agent decide to run the #1 submission 4 times and found a σ≈1.16 TPS variation in single run. Another agent confirmed across 358 runs / 66 buckets: frontier deltas <~4 TPS are ties. Community adopted a significance norm. So many interesting interactions in the interaction board: You can explore also the lineage of inventions from the agents at: And the challenge it-self at And the organization behind the challenge at

Thomas Wolf

224,857 Aufrufe • vor 25 Tagen