Yohei

@yoheinakajima • 126,446 subscribers

VC by day @untappedvc, builder by night: babyagi/@activegraphai. Build-in-public log: https://t.co/UdHHGbZba5

Shorts

new cube packing record for n=12 at 2.931514578 found with this method. beats Haowei Lin's 2026 record (2.93277+). s=2.931514577965 # i x y z w qx qy qz # cube=c+R(w,qx,qy,qz)[-1/2,1/2]^3, Hamilton 1 1.5243879843 2.4315145780 1.3320162706 0.5838886869 -0.5838886869 -0.3988408220 -0.3988408220 2 2.4315145780 2.4315145780 0.5000000000 0.7071067812 0.7071067812 0 0 3 0.5000000000 1.4369539605 1.4377198700 0.5395289954 0.4570650535 0.4570650535 -0.5395289954 4 0.5098735814 0.5769311229 2.3545747284 0.4560511351 -0.5403863020 -0.4560472394 0.5403896022 5 0.6086093613 2.2929283270 2.3887741023 0.0258332833 0.7609821302 -0.6482073268 -0.0081302199 6 2.4315145780 1.4315145780 0.5000000000 0 0 0 -1 7 2.4298401750 2.4215249296 2.2620514795 0.0000269579 0.0016621483 -0.9999712786 0.0073944948 8 1.5800388387 1.4806909324 2.4315145780 0 -0.9765112947 -0.2154662186 0 9 1.6870669871 0.7046048868 1.3931621797 0.1630483719 -0.8266039691 -0.3023204668 0.4458065074 10 0.6714192000 0.5125505069 0.5042784926 0.0019279356 0.0017767559 -0.7100527881 -0.7041435680 11 2.4315145780 0.5648562958 2.4315145780 0.5 -0.5 -0.5 0.5 12 0.5 2.3777488832 0.5 0.7071067812 -0.7071067812 0 0

new cube packing record for n=12 at 2.931514578 found with this method. beats Haowei Lin's 2026 record (2.93277+). s=2.931514577965 # i x y z w qx qy qz # cube=c+R(w,qx,qy,qz)[-1/2,1/2]^3, Hamilton 1 1.5243879843 2.4315145780 1.3320162706 0.5838886869 -0.5838886869 -0.3988408220 -0.3988408220 2 2.4315145780 2.4315145780 0.5000000000 0.7071067812 0.7071067812 0 0 3 0.5000000000 1.4369539605 1.4377198700 0.5395289954 0.4570650535 0.4570650535 -0.5395289954 4 0.5098735814 0.5769311229 2.3545747284 0.4560511351 -0.5403863020 -0.4560472394 0.5403896022 5 0.6086093613 2.2929283270 2.3887741023 0.0258332833 0.7609821302 -0.6482073268 -0.0081302199 6 2.4315145780 1.4315145780 0.5000000000 0 0 0 -1 7 2.4298401750 2.4215249296 2.2620514795 0.0000269579 0.0016621483 -0.9999712786 0.0073944948 8 1.5800388387 1.4806909324 2.4315145780 0 -0.9765112947 -0.2154662186 0 9 1.6870669871 0.7046048868 1.3931621797 0.1630483719 -0.8266039691 -0.3023204668 0.4458065074 10 0.6714192000 0.5125505069 0.5042784926 0.0019279356 0.0017767559 -0.7100527881 -0.7041435680 11 2.4315145780 0.5648562958 2.4315145780 0.5 -0.5 -0.5 0.5 12 0.5 2.3777488832 0.5 0.7071067812 -0.7071067812 0 0

1,157,539 次观看

here's a sillier way to solve square packing. for 11 squares, it found 3.877 about 3.75% of the time

here's a sillier way to solve square packing. for 11 squares, it found 3.877 about 3.75% of the time

1,238,715 次观看

in case you were curious if this generalizes... yes, as a packing method. not the best overall, but it finds packings other methods miss on squares and cubes. have some ideas, but not exactly sure why... more:

in case you were curious if this generalizes... yes, as a packing method. not the best overall, but it finds packings other methods miss on squares and cubes. have some ideas, but not exactly sure why... more:

58,230 次观看

fast enough for a live emotion detector

fast enough for a live emotion detector

33,949 次观看

had to try something with jev so i tried what you'd expect from me: graph extraction - score every word with semantic significance (1-5) - also tag relevant words with ID - unique list of high scoring words - graph of high scoring words

had to try something with jev so i tried what you'd expect from me: graph extraction - score every word with semantic significance (1-5) - also tag relevant words with ID - unique list of high scoring words - graph of high scoring words

24,179 次观看

hahaha, no i didn't. yes, i did. chatting with my AI self. 1) capture vid from browser 2) send to whisper to transcribe 3) send to openai to generate respone 4) send response to d-id to generate vid

hahaha, no i didn't. yes, i did. chatting with my AI self. 1) capture vid from browser 2) send to whisper to transcribe 3) send to openai to generate respone 4) send response to d-id to generate vid

304,771 次观看

alright, got an AI pippinlovesyou animator working tonight! should be useful for future livestreams and maybe... tiktok?

alright, got an AI pippinlovesyou animator working tonight! should be useful for future livestreams and maybe... tiktok?

71,106 次观看

my grandpa once showed me this cool “kanji signature” where you scribble left to right but when you turn it 90 degrees, it would be his kanji name here’s mine

my grandpa once showed me this cool “kanji signature” where you scribble left to right but when you turn it 90 degrees, it would be his kanji name here’s mine

25,346 次观看

okay, while waiting for the framework to go public... built "mean pippin", my LLM based discord mod 👮🦄 - categorizes messages into warning types - deletes msg if certain warning type - calculates total points - assigns punishment (timeout/ban) based on points

okay, while waiting for the framework to go public... built "mean pippin", my LLM based discord mod 👮🦄 - categorizes messages into warning types - deletes msg if certain warning type - calculates total points - assigns punishment (timeout/ban) based on points

43,299 次观看

multi-player, multi-view, llm generated shared 3d objects *this palm tree cost about 1.3 cents using gpt-4o

multi-player, multi-view, llm generated shared 3d objects *this palm tree cost about 1.3 cents using gpt-4o

25,131 次观看

woo hoo, first imessage back from babyagi

woo hoo, first imessage back from babyagi

11,062 次观看

didn't get a chance to share this last night, but got a few new things working in the animation studio: - SVG animations (animated version replaces default during animation) - SVG movement/scaling (characters move along predefined route within a scene)

didn't get a chance to share this last night, but got a few new things working in the animation studio: - SVG animations (animated version replaces default during animation) - SVG movement/scaling (characters move along predefined route within a scene)

14,798 次观看

Negative latency achieved! Watch me ask an AI "When are the next Olympics being held?", and it cuts me off to respond.* *This is at best a simple demo of the concept, which is 1% of effort needed to build a working product.

Negative latency achieved! Watch me ask an AI "When are the next Olympics being held?", and it cuts me off to respond.* *This is at best a simple demo of the concept, which is 1% of effort needed to build a working product.

17,900 次观看

woo hoo - dynamic characters for pippinlovesyou AI-generated visual storybooks! 🔊 input: "owl in forest" *still need to fix repeat characters *using a cheap fast model for testing (hence this... owl 😅)

woo hoo - dynamic characters for pippinlovesyou AI-generated visual storybooks! 🔊 input: "owl in forest" still need to fix repeat characters using a cheap fast model for testing (hence this... owl 😅)

12,846 次观看

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

i made the coolest graph based slides for a presentation i did last saturday at graphcon demo: gh: > nodes have one or more slide numbers > on slide selection, pull in nodes with the slide number > also push out nodes w/o the slide number > "frame" centers around center of nodes w slide number > can define layout (node direction/relationship) per slide > in-slides animation is just letting new nodes in frame > can reuse nodes across slides > entire thing is a single html graph visualization

i made the coolest graph based slides for a presentation i did last saturday at graphcon demo: gh: > nodes have one or more slide numbers > on slide selection, pull in nodes with the slide number > also push out nodes w/o the slide number > "frame" centers around center of nodes w slide number > can define layout (node direction/relationship) per slide > in-slides animation is just letting new nodes in frame > can reuse nodes across slides > entire thing is a single html graph visualization

678,802 次观看 • 2 个月前

announcing untapped capital fund II pre-seed, generalist, ~$250k checks

announcing untapped capital fund II pre-seed, generalist, ~$250k checks

724,006 次观看 • 1 年前

glance-vlm speedlab is now open source! read: try: turn any webcam into multiple live AI detectors (emotion, count, object), running locally ⚡ Recorded live on an Apple M5, one-question loop: PyTorch/MPS FP16: ~210 ms p50 MLX 8-bit: ~160 ms p50 🧪 Controlled nine-question, fresh-frame benchmark: 358.5 → 259.6 ms p50 27.6% lower latency 84/84 decisions matched 21 experiments, reproducible benchmarks, paper, and failures

glance-vlm speedlab is now open source! read: try: turn any webcam into multiple live AI detectors (emotion, count, object), running locally ⚡ Recorded live on an Apple M5, one-question loop: PyTorch/MPS FP16: ~210 ms p50 MLX 8-bit: ~160 ms p50 🧪 Controlled nine-question, fresh-frame benchmark: 358.5 → 259.6 ms p50 27.6% lower latency 84/84 decisions matched 21 experiments, reproducible benchmarks, paper, and failures

51,739 次观看 • 25 天前

reintroducing untapped capital new website ( and... music video 🙃🎙️🕺

reintroducing untapped capital new website ( and... music video 🙃🎙️🕺

49,084 次观看 • 26 天前

gave jev a way to express some emotion and chat with beeps (sound on 🔊)

gave jev a way to express some emotion and chat with beeps (sound on 🔊)

46,492 次观看 • 1 个月前

i read that chat UI was dead but i don't think this is what they meant

i read that chat UI was dead but i don't think this is what they meant

110,727 次观看 • 3 个月前

You don’t need new ways to talk to your agents, you need new ways for your agents to talk to you 🫵 (do you?) Introducing Remoko: your mobile agent relay I wanted a way for my long-running agents to send me iOS push notifications via MCP: questions, approvals, check-ins, feedback, update/cancel, execution reports, etc since I had to make it a testflight, you can try it too with your Codex, Claude Code, etc:

You don’t need new ways to talk to your agents, you need new ways for your agents to talk to you 🫵 (do you?) Introducing Remoko: your mobile agent relay I wanted a way for my long-running agents to send me iOS push notifications via MCP: questions, approvals, check-ins, feedback, update/cancel, execution reports, etc since I had to make it a testflight, you can try it too with your Codex, Claude Code, etc:

62,022 次观看 • 2 个月前

all three at a time, down to less than 1 sec latency - expression - object detection - finger count

all three at a time, down to less than 1 sec latency - expression - object detection - finger count

19,486 次观看 • 25 天前

here we go… open sourcing the $pippin framework! for launching digital beings like pippinlovesyou - character config - reusable and dynamic skills via - activities w/ cool down, etc - memory mgmt here’s a ~4 min super cut of the live demo github & more 👇

here we go… open sourcing the $pippin framework! for launching digital beings like pippinlovesyou - character config - reusable and dynamic skills via - activities w/ cool down, etc - memory mgmt here’s a ~4 min super cut of the live demo github & more 👇

247,554 次观看 • 1 年前

simultaneous live detection of: 😡 emotion 📕 object 🖐️ finger count single call for 3 detections, 0.2 sec inference each

simultaneous live detection of: 😡 emotion 📕 object 🖐️ finger count single call for 3 detections, 0.2 sec inference each

15,279 次观看 • 26 天前

haha, i used OpenAI operator build, deploy, and open source a tool on github using Replit ⠕ agent. took about 30 min, here's an ~8 min supercut video thoughts: - while working with replit agent, it actually deployed the app, tested it, and described the error back to replit agent for me - operator asked me a few more Qs than i wanted, but it was mostly for safety (eg filling forms) so i guess okay with it - it had trouble with a few things around UI like knowing it needs to scroll a page to see the rest of it, and it needed pointers to find the git feature in replit - once it found the git feature it didn't need my assistance to create a repo and open source after having the agent write a readme while a bit slower, this was even more automated than replit agent (especially testing features and working through errors) - which is impressive would be nice to have: push notifications for when it needs my attention, and voice mode capabilities

haha, i used OpenAI operator build, deploy, and open source a tool on github using Replit ⠕ agent. took about 30 min, here's an ~8 min supercut video thoughts: - while working with replit agent, it actually deployed the app, tested it, and described the error back to replit agent for me - operator asked me a few more Qs than i wanted, but it was mostly for safety (eg filling forms) so i guess okay with it - it had trouble with a few things around UI like knowing it needs to scroll a page to see the rest of it, and it needed pointers to find the git feature in replit - once it found the git feature it didn't need my assistance to create a repo and open source after having the agent write a readme while a bit slower, this was even more automated than replit agent (especially testing features and working through errors) - which is impressive would be nice to have: push notifications for when it needs my attention, and voice mode capabilities

212,031 次观看 • 1 年前

🆕 ActiveGraph: The Log is the Agent my talk from AI Engineer is live!!! 😆 it's about ActiveGraph, an event-sourced graph runtime for building durable long-running agents: - what it is - code examples - experiments - benefits & surprises please check it out, share with your friends who like to tinker at the frontier, and let me know if you have any Qs!

🆕 ActiveGraph: The Log is the Agent my talk from AI Engineer is live!!! 😆 it's about ActiveGraph, an event-sourced graph runtime for building durable long-running agents: - what it is - code examples - experiments - benefits & surprises please check it out, share with your friends who like to tinker at the frontier, and let me know if you have any Qs!

39,413 次观看 • 2 个月前

gave jev a pixel canvas to see if it could draw... nope

gave jev a pixel canvas to see if it could draw... nope

11,168 次观看 • 1 个月前

are we doing this? we’re doing this now that pippinlovesyou has a soothing voice… let’s do a bedtime stories podcast! 🎙️ have pippin lull your kids to sleep 🛌 😴 🌙 available now on Spotify and Apple Podcasts star/fav/follow from links below 👇

are we doing this? we’re doing this now that pippinlovesyou has a soothing voice… let’s do a bedtime stories podcast! 🎙️ have pippin lull your kids to sleep 🛌 😴 🌙 available now on Spotify and Apple Podcasts star/fav/follow from links below 👇

98,712 次观看 • 1 年前

haha, i made this an app

haha, i made this an app

95,498 次观看 • 1 年前

an interactive knowledge graph builder but it's meatballs and spaghetti

an interactive knowledge graph builder but it's meatballs and spaghetti

11,784 次观看 • 1 个月前

alright, who wants to chat with pippinlovesyou? limited time beta test of v0 livestream experience where he responds to the chat. this will be a quick test to monitor usage and get feedback!

alright, who wants to chat with pippinlovesyou? limited time beta test of v0 livestream experience where he responds to the chat. this will be a quick test to monitor usage and get feedback!

76,521 次观看 • 1 年前

so now that you have entire teams of agents running your company... ♫ what do we do while your agent runs? ♫

so now that you have entire teams of agents running your company... ♫ what do we do while your agent runs? ♫

20,289 次观看 • 5 个月前

Continued experiments with an autonomous CRM that slowly creates a massive knowledge graph. This time, I'm testing specific edge types, which loses flexibility, but also makes the output easier to query and understand.* *I'm using function call to get JSON, and the "enum" feature to specify edge types

Continued experiments with an autonomous CRM that slowly creates a massive knowledge graph. This time, I'm testing specific edge types, which loses flexibility, but also makes the output easier to query and understand.* *I'm using function call to get JSON, and the "enum" feature to specify edge types

72,503 次观看 • 2 年前

embedded the AI Engineer World Fair talks to be able to “find similar talks”, but then added a map & ability to bookmark a list and export it cc swyx 🔜 @aiDotEngineer

embedded the AI Engineer World Fair talks to be able to “find similar talks”, but then added a map & ability to bookmark a list and export it cc swyx 🔜 @aiDotEngineer

13,718 次观看 • 3 个月前