Machine Learning CHOPS

Budding ground of future CTO and machine learning architects 🚀🚀🚀🛰️🛰️🛰️🛸🛸

[Andrej Karpathy @karpathy]

We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks:

Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for aerospace maintenance documentation. LLMs well-versed in this language and it comes with heavy constraints on clean writing style that I often find a lot more readable. Sometimes I've tried to soften it a bit e.g. ask for "80% of the way to ASD-STE100" because the spec is quite stringent. But even better:

Diagrams / images. Instead of writing, ask your LLM to create a diagram. These can be a lot easier to process, parse, and understand. But even better:

Web pages. Ask for output "in HTML" to get a beautiful, interactive webpage. LLMs are getting really good at frontend and can create beautiful experiences, animations, etc. But even better:

Explainer videos. The output format I am most bullish on is fully custom / bespoke explainer videos generated on any arbitrary topic. Experiment with things like "Create a 3b1b style video explainer on X. Use my ElevenLabs API key for audio narration". (you'd need an API key for the latter or you can ask your LLM to find you decent free alternatives that use your local compute). This is actually starting to work!

In summary:
- As LLMs get better, they will do more and more of the legwork autonomously, and a lot more of our work will rise up the abstractions into oversight and understanding.
- Luckily, LLMs can help here too because as intelligence and code are increasingly abundant, you can ask for large, custom, discardable software artifacts (e.g. web apps, video explainers) that would have never made sense to create before. Push the boundaries here and you'll be surprised.

[Noisy @noisyb0y1]

She is 18 built a game with 4 agents and sold it to Microsoft for $2.4M - and at Stanford shared how to repeat it from scratch:

00:09 - how Opus 5 in Claude Code built a game making $100k a month
23:34 - 4 agents replaced a developer team worth $200k a year
59:47 - how to build your first game from scratch in 60 minutes

after watching I gave Claude Code the game idea I had been holding for 2 years - in 60 minutes it already existed in the App Store and was making money.

save & watch - the article below shows how 4 agents and Claude Code replace a developer team and build a game Microsoft will buy for $2.4M.

[CopyRebeldia @CopyRebeldia]

El curso completo de Stanford, 19 clases, gratis:
https://t.co/psWQXxhndF

[CopyRebeldia @CopyRebeldia]

Así se ve un sueldo de 750.000 dólares al año: un tipo en camiseta blanca, un pizarrón y 2 horas y media.

Stanford, CS336. Percy Liang construye un LLM desde cero. Lo que hay debajo de Claude y ChatGPT, y arranca por la parte que todos saltan: el modelo no lee tu texto, lee números.

Anthropic paga ese sueldo a los ingenieros que entienden esa capa.

Lo único que cobra Stanford son 2 horas y media de tu atención.

[Stanford NLP Group @stanfordnlp]

Where Did All the Computer Science Professors Go?

With comments from @Stanford’s Chris Gregg.

The AI industry is exciting.

The need is to keep academia exciting: “the chasm has widened as the Trump administration has cut back on scientific funding”

https://t.co/On6M974MnR

[Machine Learning CHOPS @MLCHOPSNYC]

@BecomingCritter At least they placed the chairs

[Alex Lieberman @businessbarista]

Literally the only class in the world that would make me go back to school

[AA @measure_plan]

i made tetris but the board and pieces are attached to your body and it's quite tiring to play

[Machine Learning CHOPS @MLCHOPSNYC]

GEN AI COURSE

https://t.co/BK5tNLfOHj

[Machine Learning CHOPS @MLCHOPSNYC]

Welcome back to east coast

[Jim Fan @DrJimFan]

Kaiming He, inventor of ResNet, is leaving industry to join MIT faculty in 2024!! He’s one of the most impactful figures in deep learning.

- Residual layer is a fundamental building block of LLMs.
- Faster/Mask R-CNN are industrial standards for image segmentation and robot perception stack.
- Panoptic segmentation redefined a research sub-field in vision.
- Mask AutoEncoder (MAE) is among the best general-purpose self-supervised algorithms for computer vision and beyond.
- Before MAE, Momentum Contrast (MoCo) was a SOTA contrastive learning technique.
- SlowFast network was among the default backbones for video learning until ViTs took over.
- Too many other groundbreaking works to enumerate …

I recently observe an exodus of researchers from big techs to academia. It’s an interesting movement given the current LLM gold rush 🤔

Normally we congratulate someone who joins MIT, but this time I congratulate @MITEECS to have Kaiming! 🎉

[A. Patrick Behrer @abehrer]

PhD students! You need a website! Your department website isn’t good enough. And LinkedIn is a nightmare.

Even if you’re not on the market yet, create a simple website so I can find you when I Google your name trying to offer you a job.

[ICML Conference @icmlconf]

Attend the EXPO talks on Sunday for a chance to win prizes announced at the Monday Welcome Reception.

Current list of Exhibitors:
https://t.co/8ai6hm1qB1

[ICML Conference @icmlconf]

#ICML2023 starts one month from today! Know where the awesome retail food locations are!

https://t.co/WkEXWh0q2V

[Gautam Kamath @thegautamkamath]

Proceedings PDFs for @icmlconf accepted papers are now available on OpenReview: https://t.co/vGlDgJhfLY #ICML2023

[ICML Conference @icmlconf]

Join us at the #ICML2023 Welcome Reception on Monday, July 24th at 6:15pm - 8pm in Exhibit Hall III on level one and 3rd floor foyer at the Hawaii Convention Center.

[WiML @WiMLworkshop]

✨ We are thrilled to announce our inspiring line-up of invited speakers for the #WiML un-workshop at @icmlconf! ✨

These are @RihabGrs (@instadeepai ),
@doudna_lab ( @UCBerkeley @igisci ) and @jpineau1 ( @MetaAI, @mcgillu )

@WiMLunworkshop #ML #WiML #WomenInSTEM #AI

[ICML Conference @icmlconf]

We are now on LinkedIn.

Follow us here:
https://t.co/3DuQjyBA1q

[SpaceX Super Intelligence @SpaceXSI]

https://t.co/1KhhMLLkRw

[Machine Learning CHOPS @MLCHOPSNYC]

hello @XAI

[Elon Musk @elonmusk]

Announcing formation of @xAI to understand reality