Vance Lowell @TheNormNetworks

AI/ML practitioner ~ My weights are bit off, but my biases are perfectly imbalanced. Joined January 2022

Tweets

261
Followers

40
Following

577
Likes

4K

Vance Lowell @TheNormNetworks

2 months ago

guys! we are soo back!

0 0 0 25 0

Download Image

Vance Lowell @TheNormNetworks

2 months ago

Prediction: will be getting blocked by Theo within 24 hours

0 0 0 24 0

Has anyone developed a method for (a) determining if the reasoning chain of thought is actually reasoning or if the model is pretending and - about half the time responses are incorrect because the model pretended to reason - feel like this would be useful for scaling RL

0 0 0 47 0

Vance Lowell @TheNormNetworks

2 months ago

openai should buy xai, they get the infrastructure and xai ain’t gonna make it…. But you couldn’t pick two ceos less likely to make a deal so xai will putter out until they go bankrupt and xai will loose billions paying providers a premium on gpus

1 0 0 60 0

Vance Lowell @TheNormNetworks

3 months ago

Prediction: Yann lecun’s days are numbered at Facebook ~ if he hasn’t been sidelined already It’s a fundamental misalignment to have a leader push the narrative that AI isn’t cutting edge & changing the world, your either on the train, or pushing in the wrong direction

0 0 0 55 0

Vance Lowell @TheNormNetworks

4 months ago

I love patterns that repeat over various periods. One I just noticed: Early Computer Era: founders working at night on mainframes that were unavailable/swamped during the day AI Era: working at night, so you don’t get quantized llms Innovators always identify ways to optimize

0 0 0 175 0

Vance Lowell @TheNormNetworks

8 months ago

This ain’t it folks, R1 is way overhyped, it can generate human like ramblings that sound like your pal thinking through a problem, but 01 does a much better job at identifying the actual problem in large code chunks and actually provides solutions, no one hyping it has used it

0 0 0 219 0

Vance Lowell @TheNormNetworks

8 months ago

I’m getting to the point of being comfortable in js, and by comfortable i mean that I’m getting used to my skin crawling and my brain slowly going into a coma atrophying while I type

0 0 0 77 0

Vance Lowell @TheNormNetworks

9 months ago

Woke up today and realized Claude has fallen off the map, I use gemini and 01 exclusively now, the only time I touch Claude is for Frontend design, @AnthropicAI what happened guys? You fell off so hard, your sonnet is conversation and over complicates everything, what happened?

0 0 0 61 0

Together AI @togethercompute

10 months ago

🚀 We have exciting news 🚀 We have acquired @codesandbox, a pioneer in code execution environments, and are joining forces to launch Together Code Interpreter—giving LLMs the ability to seamlessly execute the code they write. Here’s why this is a big deal for AI developers 👇

26 57 383 76K 143

Download Image

Vance Lowell @TheNormNetworks

10 months ago

Claude sonnet3.6 got a massive distillation update, and it’s performance has fallen off a cliff

0 0 0 50 0

Vance Lowell @TheNormNetworks

10 months ago

How does nobody see the past week of product releases by anthropic and openai as the clear desperate cries for new data that they are - give me your files (mtcs) and give me all websites (browser) wink wink

0 0 0 63 0

Vance Lowell @TheNormNetworks

a year ago

Ever feel like GPT-4o has a metaphorical gun to its head, ready to shoot if it gets too chatty? It’s like coaxing your quietest friend into conversation, while GPT-3.5 and 4-turbo play it cool. But GPT-4o? I said ‘good morning,’ and 30 minutes later, it’s still going!

0 0 2 187 0

Zack Ankner @ZackAnkner

a year ago

New paper where we explore using a small LM’s perplexity to prune the pretraining data for larger LMs. We find that small LMs can prune data for up to 30x larger LMs, data pruning works in the overtrained and data-constrained regimes, and more! arxiv.org/abs/2405.20541

11 61 329 73K 256

Chief AI Officer @chiefaioffice

a year ago

New market maps covering the latest AI startups. 1. AI agents from Insight Partners

51 192 1K 326K 2K

Download Image

Liam Bolling @liambolling

a year ago

💵 $1,000,000 and a Free Car: Join round two of the @Google Gemini Hackathon! Remember the Gen AI hackathon I judged? Well, we're doing it again, but this time we're giving away $1,000,000 and a custom made car to people who build the most creative ideas using Google Gemini.…

26 64 331 79K 366

Download Image

Swaroop Mishra @Swarooprm7

a year ago

🚨 Data is King in the LLM world! 👑 I am starting a thread of short, essential and actionable data advice. Here's Part 1: Improving Annotation 👇 Human annotators are vital for creating data & evaluating models. But annotation artifacts can lead to different types of spurious…