jessysu

Make data fun

[HPC-AI Tech @HPCAITech]

🚀 HPC-AI is heading to #CVPR2026! Find us at Booth 237 from June 3–7! Whether you're building AI Agents or hunting for top-tier compute, we’ve got you covered.

🎁 The Ultimate CVPR Swag & Perks:
1️⃣ Limited Swag: Grab custom caps, canvas bags 🧢 🛍️
2️⃣ Top-up Reward: 💰$15 = GPU Hour + Model APIs + Summer mini-fans / Bluetooth speaker (Only 20 set)🔥
3️⃣ Accepted Author Special: Top up $40 Get $20 voucher 💰
4️⃣ Survey Gift: Just do a quick survey, choose your refrigerator magnets 🧲

Stop by, talk tech, and claim your rewards! 🤝
📍 Booth 237 | June 3–7 | Colorado Convention Center
#CVPR2026 #HPCAI #AI #MachineLearning #GPU #ModelAPIs

[jessysu @jsu80]

AttnMod empowers you to create new art styles that are beyond description. https://t.co/dLSTiHnWlo

[Greg Kamradt @GregKamradt]

Pressure Testing GPT-4-128K With Long Context Recall

128K tokens of context is awesome - but what's performance like?

I wanted to find out so I did a “needle in a haystack” analysis

Some expected (and unexpected) results

Here's what I found:

Findings:
* GPT-4’s recall performance started to degrade above 73K tokens
* Low recall performance was correlated when the fact to be recalled was placed between at 7%-50% document depth
* If the fact was at the beginning of the document, it was recalled regardless of context length

So what:
* No Guarantees - Your facts are not guaranteed to be retrieved. Don’t bake the assumption they will into your applications
* Less context = more accuracy - This is well know, but when possible reduce the amount of context you send to GPT-4 to increase its ability to recall
* Position matters - Also well know, but facts placed at the very beginning and 2nd half of the document seem to be recalled better

Overview of the process:
* Use Paul Graham essays as ‘background’ tokens. With 218 essays it’s easy to get up to 128K tokens
* Place a random statement within the document at various depths. Fact used: “The best thing to do in San Francisco is eat a sandwich and sit in Dolores Park on a sunny day.”
* Ask GPT-4 to answer this question only using the context provided
* Evaluate GPT-4s answer with another model (gpt-4 again) using @langchain evals
* Rinse and repeat for 15x document depths between 0% (top of document) and 100% (bottom of document) and 15x context lengths (1K Tokens > 128K Tokens)

Next Steps To Take This Further:
* Iterations of this analysis were evenly distributed, it’s been suggested that doing a sigmoid distribution would be better (it would tease out more nuanced at the start and end of the document)
* For rigor, one should do a key:value retrieval step. However for relatability I did a San Francisco line within PGs essays.

Notes:
* While I think this will be directionally correct, more testing is needed to get a firmer grip on GPT4s abilities
* Switching up prompt with vary results
* 2x tests were run at large context lengths to tease out more performance
* This test cost ~$200 for API calls (a single call at 128K input tokens costs $1.28)
* Thank you to @charles_irl for being a sounding board and providing great next steps

[jessysu @jsu80]

Latent Painter turns stable diffusion output into a painting animation https://t.co/B4UTY5hE2c as seen in #NeurIPS2023 #stablediffusion

[Geoffrey Hinton @geoffreyhinton]

Dishonest CBC headline:
"Canada's AI pioneer Geoffrey Hinton says AI could wipe out humans. In the meantime, there's money to be made".
The second sentence was said by a journalist, not me, but you wouldn't know that.

[The Olympic Games™ @Olympics]

Even if you're in last place. 🏃
Even if the weather is terrible. 🌧️
Even if it feels like you can't do it. 🚫

𝙉𝙚𝙫𝙚𝙧 𝙜𝙞𝙫𝙚 𝙪𝙥 💪

Nothing was going to stop Cambodia's Bou Samnang 🇰🇭 from finishing the women's 5,000 metre race at the #SEAGames.

[Aran Komatsuzaki @arankomatsuzaki]

Align your Latents: High-Resolution Video Synthesis with Latent Diffusion Models

Turns LDM Stable Diffusion into an efficient and expressive text-to-video model with resolution up to 1280 x 2048.

proj: https://t.co/BS3fFbLCSs
abs: https://t.co/oICRrLdSjE

[pharmapsychotic @pharmapsychotic]

you can never get behind this cat
#stablediffusion #dreamstudio #aicat

[Pete @nonmayorpete]

JUST IN: NVIDIA dropped new text-to-video research.

While still far from Hollywood quality, it's pretty damn impressive how fast this is moving.

"a storm trooper vacuuming a beach"

[wallsti @wallsti2]

Top Reddit #WSB meme stocks on high quarterly earnings growth: https://t.co/6kx79GsWuP
1. $CRSR (37%)
2. $TSLA (26%)
3. $PYPL (12%)
4. $CELZ (6%)
5. $PUBM (4.4%)
6. $SHEN (2.9%)
7. $AMD (2.4%)
8. $JD (2.4%)
9. $AMZN (2.2%)
10. $UWMC (1.4%)

[Google DeepMind @GoogleDeepMind]

Introducing UNREAL, our UNsupervised REinforcement & Auxiliary Learning agent! Results on Labyrinth & Atari https://t.co/GczQ8UlQko

[morethancure @morethancure]

What to Do If Your Parents Are Causing You Career Angst RT @HarvardBiz https://t.co/xJJlOtnG70 #Anxiety

[morethancure @morethancure]

A Puzzling Projection Simulates What It’s Like to Live With Dementia https://t.co/ybdEk5WULv @MoreThanCure #Dementia https://t.co/7tYXha68TH

[jessysu @jsu80]

Face2Face: Real-time Face Capture and Reenactment of RGB Videos (CVPR 20... https://t.co/p94yMfO6bA via @YouTube

[Elon Musk @elonmusk]

Model 3 will be RWD, with dual motor AWD optional. Even RWD will have great traction on ice due to fast torque response of Tesla drivetrain.

[jessysu @jsu80]

New York firm hires those with autism to help squash software bugs https://t.co/g4HNERHdkL via @InaFried

[morethancure @morethancure]

Severe chronic sinus issues not responding to any treatment. https://t.co/5BTEWOV8f7 #Edema

[Elon Musk @elonmusk]

Model 3 design sketches

[jessysu @jsu80]

Facebook sets its sights on identity theft https://t.co/BFWJw6Qxrj via @alphr

[Kansas City Royals @Royals]

Chien-Ming Wang throws zeros on the scoreboard against the A’s: https://t.co/qqP540Jwno #RoyalsST