Aparimeya
@aparimeya
Followers
177
Following
247
Media
41
Statuses
249
Founding eng @paraformtalent | Prev - @doordash | Duke 22 | ISEF 2017 Talk to me about AI, art, edm, history, psychology, space
San Francisco, CA
Joined September 2020
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
SEVENTEEN
• 1535141 Tweets
Hanni
• 169941 Tweets
無料クーポン
• 161105 Tweets
Jets
• 121704 Tweets
連休明け
• 93538 Tweets
Carat
• 87635 Tweets
DG X ORM KORNNAPHAT
• 77901 Tweets
अब्दुल कलाम
• 60452 Tweets
पूर्व राष्ट्रपति
• 42349 Tweets
भारत रत्न
• 32619 Tweets
FELIP TAKES ON BBPHSTAGE
• 24619 Tweets
JC BBPH MAINSTAGE
• 23908 Tweets
The Next Prince Q9
• 23618 Tweets
ミンギュ
• 17872 Tweets
金木犀の香り
• 15582 Tweets
バーニス
• 14943 Tweets
मुस्लिम एरिया
• 13980 Tweets
うさぎだらけくじ
• 12881 Tweets
ITZY WORLD IS STILL DIFFERENT
• 11960 Tweets
ジョシュア
• 11896 Tweets
スングァン
• 11446 Tweets
エスクプス
• 11017 Tweets
ドギョム
• 10178 Tweets
@ryanflorence
I told gpt-5 about my lifestyle and it predicted my future symptoms and problems… 🤯
1
0
20
@FreyaHolmer
As an artist myself I firmly believe it should be illegal to train models w images artists do not consent to use for training.
However I would still advocate for AI art because of all the new kinds of stuff I am able to make that I couldn’t before.
0
0
17
@rasbt
Amazing! ( If you happen to have a few A100s for training lying around ig lol). Btw what happened to BLOOM? Wasn’t that technically the first open source AND open access ( already trained) competitor to GPT-3?
2
1
13
@brian_armstrong
Those are but rookie numbers.
The Pentagon has failed it's fifth audit in a row after failing to account for 61% of their $3.5 Trillion in assets
But the IRS will demand the average American account for all of their $600 Venmo side hustles
Guess that $800B Pentagon budget won't fund itself🤷♀️
893
9K
37K
0
0
11
@julesgm4
@Abel_TorresM
@MichaelTrazzi
It can! Well not chatgpt but LLMs in general. Some ground breaking stuff by
@oh_that_hat
“LLMs can’t even do addition”
📄🚨We show that they CAN add! To teach algos to LLMs, the trick is to describe the algo in enough detail so that there is no room for misinterpretation
w/
@Azade_na
@hugo_larochelle
@AaronCourville
@bneyshabur
@HanieSedghi
24
154
1K
2
0
7
@jasminericegirl
@LonnieIIV
Ok wait I think I accidentally made chatGPT have a meltdown:
1) normal question, boring answer
2) prompt injecting, works but too many words
3) asked it to summarise with emojis, and … lmao
#ChatGPT
0
0
5
When I think of intelligence, I think of emergence. Here’s a preview of my personal website I’m working on that runs a simulation based on very simple rules that visualises this concept at a particle level.
Made using
@sveltejs
0
0
4
@goodside
Ever since google switched to semantic search, we have all been “prompt engineering” (kinda)
Just like how people who write prompts for Stable Diffusion add “ Art Station”, people who want real human written answers add “Reddit” to the end of their prompts.
0
0
4
@heybiji
two things can be true at once:
1) Artists must get to choose wether or not their art can be used to train models
2) AI art should be appreciated for the expression of the users ideas, because ultimately it's just a tool.
0
0
3
@SvelteSociety
@sveltejs
I spend 5 minutes making productive code and 50 minutes figuring out how to dynamically import firebase.
0
0
3
@pokecoder
I think of re making my site everyday, but then I remember this is what a $300B company's website looks like
0
0
1
@trunarla
U forgot the years of practice it takes to first google what characters to type in what order.
0
0
3
@norabelrose
Why don’t any of these decels ever mention what specifically the “great harm” these AIs are capable of actually is? Like what can gpt4 enable a bad actor to do that they couldn’t do before?
1
0
3
AI Imagines cute pets in suits and sunglasses playing poker.
.
.
.
#ai
#aiart
#StableDiffusionArt
#stablediffusion
0
0
3
Ah yes, an exclusive mandate. From the heavens perhaps? I see no way this goes wrong or is in any one particular group’s interest at the cost of everyone else.
MAGIC will:
1. Ban the development of AI systems above a certain threshold of computing power.
2. Have an exclusive mandate to conduct critical AI research.
Full paper:
52
2
26
0
0
2
@opensas
@SvelteSociety
sure hope one of them is bringing back a file naming convention that isnt 5109 files called "+page.svelte" in the same project
0
0
3
Ya boy made it
0
0
3
PWA made with
#sveltejs
(sapper) and
#firebase
for my new free platform . Best Dev experience ever
@Rich_Harris
ur a genius.
@MadeWithSvelte
0
0
1
@docmilanfar
Would be interesting to see how ResNet citations have changed over time, i.e more or less each year
1
0
1
@MattieRobinson0
@ryanflorence
I love how easy it is to make it go from "noo dont ask me medical stuff im not accurate enough!! 🙈 " to "I am right about everything always 😎 "
0
0
2
@CaseyNewton
Not really, deep mind is still a pioneer in this field and even main google research into LLMs and next gen transformer AIs seems to be pretty promising!
0
0
2
@michael_nielsen
Actually, as LLMs get better we’re going to quickly face the reality that logic is actually just an emergent property of language.
Here’s google ALREADY using a LLM as the logic planning part of their autonomous robot:
0
0
2
@_arohan_
Imo we should be doing the exact opposite I.e use these models to make the language easier to understand for a wider audience
0
0
2
0
1
2
Here’s my shot at this, but instead I used stablediffusion2.0 on sequential frames to make an animation.
#ChatGPT
#StableDiffusion2
#AIart
0
0
2
@amasad
Was I the only one not that surprised by this? Considering how long we’ve known that in our own brains the physical locations of the language and logic centers overlap, implying that the latter is an emergent property of the former.
0
0
2
@aallan
@Raspberry_Pi
Are we so back? Btw plans to add GPU capabilities to pi? Can imagine a lot of cool stuff could be built w on edge inference
1
0
2
@OfficialLoganK
Used it to build something we were gonna use in prod, but keeps changing / breaking everytime I run it so just sorta given up for now. Maybe 1.0.0 release.
0
0
2
VALL-E for voice, SD for visuals, and GPT for scripts.
Can u imagine the flood of fully automated content that will flood social media this year?
Surprised there isn't more chatter around VALL-E
This new model by
@Microsoft
can generate speech in any voice after only hearing a 3s sample of that voice 🤯
Demo →
123
713
4K
0
0
1
So is this similar to the previous google robot demo from this year that used an LLM as the planning Center?
Introducing RT-1, a robotic model that can execute over 700 instructions in the real world at 97% success rate!
Generalizes to new tasks✅
Robust to new environments and objects✅
Fast inference for real time control✅
Can absorb multi-robot data✅
Powers SayCan✅
🧵👇
62
548
2K
0
0
1
@zack_overflow
Nice! Is the Source code available for this? Been trying to run a simple 2d particle simulation with just JS and it’s really slow lol I was thinking of looking into WASM.
1
0
1
0
0
1
🤩
StockHub by
@aparimeya
tracks social metrics for stocks. The web app
#madewithsvelte
tracks the number tweets or post on Reddit about them and does a sentiment analysis! 📈📉 -
0
1
9
1
0
1
@trunarla
JS devs and re inventing an npm package that does literally the same exact thing with just 2 lines of code less.
0
0
1
@FreyaHolmer
Imo after the initial hype and over saturation of simple images, theres gonna be a lot of new kind of art (and creative things in general) not possible before. Personally, I’ve been making AI art videos because that’s something I could never do before in PS.
0
0
1
LMMs: demonstrate logical reasoning and potential for automating just about any task once thought too complex
Web3: hey bro wanna buy this new Ding Dong coin that allows kids in Africa to trade links to JPEGs 🚀 💎 🙌
They are not the same.
The current climate in AI has so many parallels to 2021 web3 it's making me uncomfortable. Narratives based on zero data are accepted as self-evident. Everyone is expecting as a sure thing "civilization-altering" impact (& 100x returns on investment) in the next 2-3 years
218
1K
7K
0
0
1
0
0
1
@Xumiiz
I know it has millions of views and was v famous for its time but I have never seen anyone outside of YouTube talk about it or have seen it so I figured it was me and the same 10 ppl who keep rewatching it to give it views.
0
0
1
Update: site now live at
Would love to hear what u think!
#svelte
#Simulation
When I think of intelligence, I think of emergence. Here’s a preview of my personal website I’m working on that runs a simulation based on very simple rules that visualises this concept at a particle level.
Made using
@sveltejs
0
0
4
0
0
1
@quantian1
Tested copilot on some CS201 assignments and it did in-fact solve them 4/5 times so not a totally unfounded claim
1
0
1
@jchervinsky
Lol didn’t they buy a Super Bowl ad? I.e. if US residents are a core demographic for them then shouldn’t the US gov do something for consumer protection?
0
0
1
@lihautan
Yea that was my first idea but
@StephaneVanraes
suggested I make a PR and I wanted to see how it went so I did
0
0
1
My personal fav ( unicorn on fire ):
Want to try AI generated video for free?
@aparimeya
& I wanted to learn how to take a large ML model, serve it over the internet, and build an app around it: Try out !
Had a lot of fun hacking this over a weekend - we'd love to see any cool GIFs you make!
1
2
10
0
0
2
As soon as I figure out how to deploy GPU accelerated pipelines on
@modal_labs
it’s so over
Modal is generally available today, and we also raised a Series A!
38
82
692
0
0
1
Stable diffusion 2.0 is good, actually.
Asked it to make tourism ad posters for India, America and Japan.
0
0
1
AI art style transfer. Stanford dish in the style of Van Gogh’s The starry night
0
0
1
Yea man sooo sane PAUSING™ work on the one thing that can save us from demographic collapse, climate change, genetic diseases, etc.
These people will do to AI what the previous generation did to nuclear. Folding on a pair of aces, only to benefit the incumbents.
Want to hold up this sign on Saturday in San Francisco?
As part of the largest PauseAI protest in history?
Joining hundreds of sane people protesting across 8 countries?
DM me.
88
21
152
0
0
1
Started svelte when 3 just came out because
@Rich_Harris
’s demo convinced me this would set a new standard of web dev. Congrats to the amazing team behind this amazing tool!
Svelte Summit San Francisco?
0
0
1
@fireship_dev
for sure. That’s how I started!
It's Follow Friday! Please quote tweet with 1 person more people should follow in
@sveltejs
👀
let's see how long a chain we can build!
0
0
5
0
0
1
Cool if true ( it’s not )
GPT-4 is going to launch soon.
And it will make ChatGPT look like a toy...
→ GPT-3 has 175 billion parameters
→ GPT-4 has 100 trillion parameters
I think we're gonna see something absolutely mindblowing this time!
And the best part? 👇
246
1K
6K
0
0
1
The funniest tweet about this
0
0
1