Dillon Erb @dlnrb

Building AI @ DigitalOcean. Formerly CEO / co-founder @HelloPaperspace paperspace.com New York, USA Joined August 2010

Tweets

296
Followers

479
Following

947
Likes

1K

Jack Morris @jxmnop

8 months ago

posted the other day about model distillation. pretty much everyone responded with their theories professors, leading lab researchers, students, pseudoanonymous anime-profile posters seems there's no clear consensus why it works, but here are the theories 🧵

Jack Morris @jxmnop

8 months ago

251 176 3K 372K 2K

14 44 570 71K 493

MOrdO Kai @kai_mordo

9 months ago

@deepfates simple, low gpu use, built in api server, and RAG retrieval - GPT4ALL

3 1 53 4K 51

Download Image

Yining Shi @yining_shi

10 months ago

We're looking for a hands-on engineering manager to lead our growing front-end team at @runwayml. If you're passionate about building AI-driven products and growing high-performing and fast-paced engineering teams, we'd love to hear from you 🙏. job-boards.greenhouse.io/runwayml/jobs/…

4 10 35 5K 5

Etched @Etched

10 months ago

Introducing Oasis: the first playable AI-generated game. We partnered with @DecartAI to build a real-time, interactive world model that runs >10x faster on Sohu. We're open-sourcing the model architecture, weights, and research. Here's how it works (and a demo you can play!):

1K 1K 13K 24.6M 10K

Download Video

Etched @Etched

a year ago

Meet Sohu, the fastest AI chip of all time. With over 500,000 tokens per second running Llama 70B, Sohu lets you build products that are impossible on GPUs. One 8xSohu server replaces 160 H100s. Sohu is the first specialized chip (ASIC) for transformer models. By specializing,…

342 1K 6K 2.5M 2K

Download Image

Dillon Erb @dlnrb

a year ago

Incredible update to the best open source embedding model. Congrats @nomic_ai 💪

CalCo @calco_io

a year ago

Incredible update to the best open source embedding model. Congrats @nomic_ai 💪

22 125 631 103K 483

Download Video

1 0 11 599 0

Cameron R. Wolfe, Ph.D. @cwolferesearch

a year ago

@rasbt Super interesting, data curation is becoming the cornerstone of LLM research

0 1 8 2K 0

Dillon Erb @dlnrb

a year ago

If you are in NYC next week come see some amazing speakers at our first ever AI Dev Day @ DO's SoHo office starring @nomic_ai @NormalComputing @TrainyAI and @LeptonAI . RSVP link below 👇👇

DigitalOcean @digitalocean

a year ago

If you are in NYC next week come see some amazing speakers at our first ever AI Dev Day @ DO's SoHo office starring @nomic_ai @NormalComputing @TrainyAI and @LeptonAI . RSVP link below 👇👇

1 3 11 9K 4

Download Image

0 2 11 1K 1

Dillon Erb @dlnrb

a year ago

I'm hosting the first of many Dev Day's at DigitalOcean. If you are in the NYC are would love to have you join!

DigitalOcean @digitalocean

a year ago

I'm hosting the first of many Dev Day's at DigitalOcean. If you are in the NYC are would love to have you join!

1 3 11 9K 4

Download Image

1 1 9 2K 0

Nikita Bier @nikitabier

2 years ago

Every investor who has stood in the way of a company being acquired has regretted it. Once an acquisition is in motion, the most important thing to do as an investor is to fall in line. The process is so incredibly stressful that any effort to derail it will also derail the…

35 45 1K 480K 243

Harj Taggar @harjtaggar

2 years ago

One thing that’s made office hours with the AI startups in the current YC batch so fun is imagining what their software might be capable of doing in the future. We can throw out crazy ideas knowing it might only be a few months before they don’t seem crazy anymore.

11 16 134 54K 20

Mark Tenenholtz @marktenenholtz

2 years ago

The best research is done with fast iterations on small models. This paper is one of the first that I’ve seen where the authors explained how they did that small scale testing.

3 9 133 17K 61

Download Image

Siqi Chen @blader

2 years ago

after two decades of trying to get better at “leadership”, this is what i’ve got: leadership is about having extremely high expectations, deeply believing in people’s ability to grow to meet those expectations, and having the responsibility to support people in that growth.

23 36 435 42K 250

Ivan Fioravanti ᯅ @ivanfioravanti

2 years ago

An example of Apple MLX power and "hackability": visualizing model confidence. NeuralBeagle14 here. I'm "playing" with code to add --colorize param to mlx_lm to be able to show confidence of the model on the next token using standard top_p algorithm. I started from @antirez…

3 13 129 21K 92

Download Video

Supasorn Suwajanakorn @supasornaek

2 years ago

Introducing DiffusionLight---a simple yet effective technique to estimate lighting from any in-the-wild input image. How? ... by inpainting a chrome ball into the image with diffusion models! (1/3) paper: arxiv.org/abs/2312.09168 diffusionlight.github.io huggingface.co/DiffusionLight…

22 158 843 104K 429

Download Video

Simon Suo @disiok

2 years ago

Do we really want AGI or just really cheap and fast open source thought calculators that runs in our laptop

2 3 16 2K 1

Dillon Erb @dlnrb

2 years ago

Fascinating. I expect we start seeing a lot more of these LLM middlewares and routers soon.

Tomas Hernando Kofman @tomas_hk

2 years ago

Fascinating. I expect we start seeing a lot more of these LLM middlewares and routers soon.

13 64 481 96K 412

1 0 1 213 1

Quantіan @quantian1

2 years ago

There’s no good theoretical reason I can think of for why the space of human legible images should be both so sparse *and* so compact in the space of all possible JPEGs that this works. Most images you generate, even for tiny perturbations, “should” look like random TV snow.

16 15 507 46K 31

Patrick Coles @ColesThermoAI

2 years ago

We have exciting news to share 🔥 The team at @NormalComputing has performed the first thermodynamic linear algebra experiment. This is an experimental follow-up to our theory paper that came out in August. The details are in this blog, see thread below blog.normalcomputing.ai/posts/2023-11-…