Continuous Learning_Startup & Investment
2.43K subscribers
513 photos
5 videos
16 files
2.76K links
We journey together through the captivating realms of entrepreneurship, investment, life, and technology. This is my chronicle of exploration, where I capture and share the lessons that shape our world. Join us and let's never stop learning!
Download Telegram
1) Many founders are really feeling the pain of the GPU shortage. Some are even building their own dedicated clusters but of course they would not say where they found their GPUs ๐Ÿ™‚
2) While the speed of OpenAI execution has been amazing, Google has really picked up the pace in the last few months with many founders expecting them to be a formidable player.
3) AI makes for incredible demos of autonomous agents but enterprise adoption of agents is quite nascent since their behavior can be unpredictable. Enterprises will need technology to monitor agents, enforce policy guardrails, and secure their access to data and systems.
4) For enterprises willing to experiment with AI, what is the right go-to-market model? PLG may be too much of an ideal given the level of customer maturity and more handholding is likely needed to ensure customer success.
5) It is believed that recruiting out of large tech companies has been getting easier. Given the speed with which AI is moving there is fear of missing the boat. Everyone wants to build and launch products quickly without big company red tape or technical debt.
To quote the great climate scientist, Greta Thunberg, โ€œHow dare you!โ€

Joking aside, because no serious person would actually outsource their energy policy to a kid, nuclear is a no-brainer. The biggest problem from here isnโ€™t the technology - those will continue to improve and the safety profiles will become even more bulletproof.

But even with the greatest technology, what we have consistently overlooked are the local, regional and state regulations to get new reactors built.

Nuclear needs to be viewed as a national security issue. More nuclear == more clean energy == less foreign energy needed == less wars == lower deficits == more internal stability.

Right now it isnโ€™t, so building a new reactor is a multi decade ordeal and is effectively impossible. So as nuclear technologies get better, nuclear buildouts will still meaningfully lag relative to other countries, especially China.

https://budd.senate.gov/press-releases/budd-coons-lead-bipartisan-coalition-supporting-nuclear-energy/
Research on advanced prompting techniques for language models has extended chain of thought and tree of thought prompting to graph-structured reasoning processes. But, did you know that there are two versions of โ€œgraph of thoughtโ€ prompting that have been proposed already?

Some background. Advanced prompting techniques like chain of thought and tree of thought prompting have drastically improved the ability of large language models to solve complex, reasoning-based tasks. Forcing the LLM to construct a step-by-step response to a problem drastically improves its problem-solving capabilities, but all of these techniques assume that the reasoning process is linear.

โ€œHuman thinking is often characterized by its ability to make sudden leaps and connections between seemingly unrelated ideas, which can lead to novel insights and solutions. This non-linear, jumping thought process is a hallmark of human creativity, reasoning, and problem-solving abilities.โ€ - from [1]

Graph-based reasoning. Humans do not seem to perform reasoning based on individual chains of thought. Rather, we make leaps and connections between ideas that lead to novel insights. Inspired by this idea, researchers have recently extended chain and tree of thoughts prompting to a graph-structured approach. We will take a look here at two (independent) papers that have already been written on this topic.

Graph of thought reasoning (GOTR). In [1], authors proposed a two-stage technique that:

1. Outputs a problem solving rationale given text (and potentially images) as input.
2. Outputs a final answer given the original input concatenated with a rationale.

This approach uses an encoder-decoder structure and is fine-tuned end-to-end. Several encoders ingest data from each of the different modalities that are considered. The output of these encoders is combined in a fusion layer, then passed to a decoder to generate output.

Where does the graph come in? So far, it seems like GOTR does not use any graphs within its reasoning process. However, the model creates a named entity graph based on the input text and generated rationale. Then, this graph is ingested by a graph attention network encoder and combined with all image/text features. As such, the decoder receives information from text, image, and graph-based inputs!

GoT prompting. In [2], authors follow a more traditional prompting approach, called graph of thought (GoT) prompting, that uses a system of causal LLMs and prompts to perform reasoning according to a graph structure. The reasoning process is modeled as a graph, where each node represents a thought or (partial) solution and connections indicate that a certain thought was generated from some other thought.

A system of LLMs. GoT prompting has several โ€œmodulesโ€ that control the reasoning process, including a top-level controller (controls the reasoning process), a parser (verifies and extracts LLM output), a scorer (judges the quality of solutions), and a prompter (writes prompts for each different module). Together, these modules can transform the underlying graph structure and work towards solving a reasoning problem.

TL;DR: Modeling an LLMโ€™s reasoning process as a graph structure can be beneficial for certain problems and is (arguably) more comparable to the human reasoning process. But, these techniques tend to be more costly than basic CoT prompting and only provide a tangible benefit on select problems. For more details, check out the overview of these techniques that I just wrote for my newsletter.

--------
[1] Yao, Yao, Zuchao Li, and Hai Zhao. "Beyond Chain-of-Thought, Effective Graph-of-Thought Reasoning in Large Language Models." arXiv preprint arXiv:2305.16582 (2023).
[2] Besta, Maciej, et al. "Graph of Thoughts: Solving Elaborate Problems with Large Language Models." arXiv preprint arXiv:2308.09687 (2023).

https://twitter.com/cwolferesearch/status/1696282034145006006?s=20
LLM ๊ธฐ๋ฐ˜์˜ ์‹œ์Šคํ…œ & ์ œํ’ˆ ๊ตฌ์ถ•์„ ์œ„ํ•œ 7๊ฐ€์ง€ ํ•ต์‹ฌ ํŒจํ„ด

"์„ฑ๋Šฅ ํ–ฅ์ƒ vs. ๋น„์šฉ/๋ฆฌ์Šคํฌ ๊ฐ์†Œ" ๋ฐ "๋ฐ์ดํ„ฐ ์นœํ™” vs ์‚ฌ์šฉ์ž ์นœํ™”" ๋กœ ์ •๋ฆฌ
- Evals: ์„ฑ๋Šฅ ์ธก์ •
- RAG(Retrieval-Augmented Generation): ์ตœ์‹ , ์™ธ๋ถ€ ์ง€์‹์„ ์ถ”๊ฐ€
- Fine-tuning: ํŠน์ • ์ž‘์—…์„ ๋” ์ž˜ ์ˆ˜ํ–‰ํ•˜๊ธฐ ์œ„ํ•ด
- Caching: ๋ ˆ์ดํ„ด์‹œ ๋ฐ ๋น„์šฉ ๊ฐ์†Œ
- Guardrails: ์ถœ๋ ฅ ํ’ˆ์งˆ ๋ณด์žฅ
- Defensive UX: ์˜ค๋ฅ˜๋ฅผ ์—์ธกํ•˜๊ณ  ๊ด€๋ฆฌํ•˜๊ธฐ ์œ„ํ•ด
- Collect user feedback: ๋ฐ์ดํ„ฐ ํ”Œ๋ผ์ด ํœ  ๊ตฌ์ถ•

# Evals: ์„ฑ๋Šฅ ์ธก์ •

- Evals๋Š” ์ž‘์—…์—์„œ ๋ชจ๋ธ์˜ ์„ฑ๋Šฅ์„ ํ‰๊ฐ€ํ•˜๋Š”๋ฐ ์‚ฌ์šฉ๋˜๋Š” ์ผ๋ จ์˜ ์ธก์ •๊ฐ’๋“ค
- ๋ฒค์น˜๋งˆํฌ ๋ฐ์ดํ„ฐ ๋ฐ ๋ฉ”ํŠธ๋ฆญ ํฌํ•จ
- ์‹œ์Šคํ…œ ๋˜๋Š” ์ œํ’ˆ์ด ์–ผ๋งˆ๋‚˜ ์ž˜ ์ž‘๋™ํ•˜๊ณ  ์žˆ๋Š”์ง€ ์ธก์ •ํ•˜๊ณ , ํ‡ด๋ณด๋ฅผ ๊ฐ์ง€ํ•  ์ˆ˜ ์žˆ์Œ
- ์–ธ์–ด ๋ชจ๋ธ๋ง ๋ถ„์•ผ์— ๋งŽ์€ ๋ฒค์น˜๋งˆํฌ๋“ค์ด ์žˆ์Œ: MMLU, EleutherAI Eval, HELM, AlpacaEval
- ๋ฉ”ํŠธ๋ฆญ์„ ๋‘๊ฐœ์˜ ์นดํ…Œ๊ณ ๋ฆฌ๋กœ ๊ตฌ๋ถ„ ๊ฐ€๋Šฅ: Context-dependent ๋˜๋Š” Context-free
- ๊ณตํ†ต์ ์œผ๋กœ ์‚ฌ์šฉํ•˜๋Š” ๋ฉ”ํŠธ๋ฆญ๋“ค: BLEU, ROUGE, BERTScore, MoverScore
- ์š”์ฆ˜ ๋œจ๋Š” ํŠธ๋ Œ๋“œ๋Š” ๊ฐ•๋ ฅํ•œ LLM์„ reference-free metric์œผ๋กœ ์ด์šฉํ•˜์—ฌ ๋‹ค๋ฅธ LLM๋“ค์˜ ์ƒ์„ฑ๋ฌผ์„ ํ‰๊ฐ€ํ•˜๋Š” ๊ฒƒ
- G-Eval, Vicuna ๋…ผ๋ฌธ, QLoRA

# RAG(Retrieval-Augmented Generation): ์ตœ์‹ , ์™ธ๋ถ€ ์ง€์‹์„ ์ถ”๊ฐ€

- ํŒŒ์šด๋ฐ์ด์…˜ ๋ชจ๋ธ ์™ธ๋ถ€๋กœ๋ถ€ํ„ฐ ์ •๋ณด๋ฅผ ๊ฐ€์ ธ์™€ ์ด ๋ฐ์ดํ„ฐ๋กœ ์ž…๋ ฅ์„ ๊ฐ•ํ™”ํ•˜์—ฌ ๋” ํ’๋ถ€ํ•œ ์ปจํ…์ŠคํŠธ๋ฅผ ์ œ๊ณตํ•จ์œผ๋กœ์จ ์ถœ๋ ฅ์„ ๊ฐœ์„ 
- RAG๋Š” ๊ฒ€์ƒ‰๋œ ์ปจํ…์ŠคํŠธ์— ๋ชจ๋ธ์„ ๊ธฐ๋ฐ˜์œผ๋กœ ํ•˜์—ฌ ํ™˜๊ฐ์„ ์ค„์ด๋Š”๋ฐ ๋„์›€์„ ์ค˜์„œ ์‚ฌ์‹ค์„ฑ์„ ๋†’์ž„
- ๋˜ํ•œ LLM์„ ์ง€์†์ ์œผ๋กœ ์‚ฌ์ „ ํ•™์Šตํ•˜๋Š” ๊ฒƒ๋ณด๋‹ค ๊ฒ€์ƒ‰ ์ธ๋ฑ์Šค๋ฅผ ์ตœ์‹  ์ƒํƒœ๋กœ ์œ ์ง€ํ•˜๋Š” ๊ฒƒ์ด ๋” ์ €๋ ด
- ์ด๋Ÿฐ ๋น„์šฉ ํšจ์œจ์„ฑ ๋•Œ๋ฌธ์— LLM์ด RAG์„ ํ†ตํ•ด ์ตœ์‹  ๋ฐ์ดํ„ฐ์— ์–ต์„ธ์Šค ๊ฐ€๋Šฅ
- ํŽธํ–ฅ๋˜๊ฑฐ๋‚˜ ์œ ํ•ดํ•œ ๋ฌธ์„œ์™€ ๊ฐ™์€ ๋ฐ์ดํ„ฐ๋ฅผ ์—…๋ฐ์ดํŠธ/์ œ๊ฑฐํ•ด์•ผ ํ•˜๋Š” ๊ฒฝ์šฐ ๊ฒ€์ƒ‰ ์ธ๋ฑ์Šค๋ฅผ ์—…๋ฐ์ดํŠธํ•˜๋Š” ๊ฒƒ์ด ๋” ๊ฐ„๋‹จํ•จ(LLM์„ ๋ฏธ์„ธ์กฐ์ • ํ•˜๋Š” ๊ฒƒ์— ๋น„ํ•ด)
- RAG์„ ์œ„ํ•ด์„œ๋Š” ํ…์ŠคํŠธ ์ž„๋ฒ ๋”ฉ์— ๋Œ€ํ•ด ๋จผ์ € ์ดํ•ดํ•˜๋Š” ๊ฒƒ์ด ๋„์›€์ด ๋จ
- ํ…์ŠคํŠธ ์ž„๋ฒ ๋”ฉ์€ ์ž„์˜ ๊ธธ์ด์˜ ํ…์ŠคํŠธ๋ฅผ ์ˆซ์ž์˜ ๊ณ ์ • ํฌ๊ธฐ ๋ฒกํ„ฐ๋กœ ํ‘œํ˜„ํ•  ์ˆ˜ ์žˆ๋Š” ํ…์ŠคํŠธ ๋ฐ์ดํ„ฐ์˜ ์••์ถ•๋œ ์ถ”์ƒ ํ‘œํ˜„
- ์ผ๋ฐ˜์ ์œผ๋กœ Wikipedia๊ฐ™์€ ํ…์ŠคํŠธ ์ฝ”ํผ์Šค์—์„œ ํ•™์Šตํ•จ
- ์œ ์‚ฌํ•œ ํ•ญ๋ชฉ์€ ์„œ๋กœ ๊ฐ€๊น๊ณ , ์œ ์‚ฌํ•˜์ง€ ์•Š์€ ํ•ญ๋ชฉ์€ ๋” ๋ฉ€๋ฆฌ ๋–จ์–ด์ ธ ์žˆ๋Š” ํ…์ŠคํŠธ์— ๋Œ€ํ•œ ๋ฒ”์šฉ ์ธ์ฝ”๋”ฉ์œผ๋กœ ์ƒ๊ฐํ•˜๋ฉด ๋จ
- ์ข‹์€ ์ž„๋ฒ ๋”ฉ์€ ์œ ์‚ฌ ํ•ญ๋ชฉ ๊ฒ€์ƒ‰ ๊ฐ™์€ ๋‹ค์šด์ŠคํŠธ๋ฆผ ์ž‘์—…์„ ์ž˜ ์ˆ˜ํ–‰ํ•˜๋Š” ๊ฒƒ
- Huggingface์˜ Massive Text Embedding Benchmark (MTEB)๋Š” ๋ถ„๋ฅ˜,ํด๋Ÿฌ์Šคํ„ฐ๋ง,๊ฒ€์ƒ‰,์š”์•ฝ ๊ฐ™์€ ๋‹ค์–‘ํ•œ ์ž‘์—…์—์„œ ๋ชจ๋ธ์˜ ์ ์ˆ˜๋ฅผ ๋งค๊น€
- ์—ฌ๊ธฐ์„œ๋Š” ์ฃผ๋กœ ํ…์ŠคํŠธ ์ž„๋ฒ ๋”ฉ์— ๋Œ€ํ•ด ์ด์•ผ๊ธฐ ํ•˜์ง€๋งŒ, ์ž„๋ฒ ๋”ฉ์€ ๋‹ค์–‘ํ•œ ๋ชจ๋‹ฌ๋ฆฌํ‹ฐ๊ฐ€ ์‚ฌ์šฉ๋  ์ˆ˜ ์žˆ์Œ
- Fusion-in-Decoder(FiD)๋Š” ์˜คํ”ˆ ๋„๋ฉ”์ธ QA๋ฅผ ์œ„ํ•ด ์ƒ์„ฑํ˜• ๋ชจ๋ธ๊ณผ ๊ฒ€์ƒ‰์„ ๊ฐ™์ด ์‚ฌ์šฉํ•จ
- Internet-augmented LM๋“ค์€ ๊ธฐ์กด ๊ฒ€์ƒ‰์—”์ง„์„ ์ด์šฉํ•˜์—ฌ LLM ๊ฐ•ํ™”๋ฅผ ์ œ์•ˆ
- RAG ์ ์šฉ ๋ฐฉ๋ฒ•
- ํ•˜์ด๋ธŒ๋ฆฌ๋“œ ๊ฒ€์ƒ‰(์ „ํ†ต์ ์ธ ๊ฒ€์ƒ‰ ์ธ๋ฑ์Šค + ์ž„๋ฒ ๋”ฉ ๊ธฐ๋ฐ˜ ๊ฒ€์ƒ‰)์ด ๊ฐ๊ฐ ๋‹จ๋…๋ณด๋‹ค ๋” ์ž˜ ๋™์ž‘ํ•จ

# Fine-tuning: ํŠน์ • ์ž‘์—…์„ ๋” ์ž˜ ์ˆ˜ํ–‰ํ•˜๊ธฐ ์œ„ํ•ด

- ๋ฏธ์„ธ ์กฐ์ •์€ ์‚ฌ์ „ ํ›ˆ๋ จ๋œ ๋ชจ๋ธ(๋ฐฉ๋Œ€ํ•œ ์–‘์˜ ๋ฐ์ดํ„ฐ๋กœ ์ด๋ฏธ ํ›ˆ๋ จ๋œ ๋ชจ๋ธ)์„ ๊ฐ€์ ธ์™€ ํŠน์ • ์ž‘์—…์— ๋Œ€ํ•ด ์ถ”๊ฐ€๋กœ ์ •์ œํ•˜๋Š” ํ”„๋กœ์„ธ์Šค
- ๋ชจ๋ธ์ด ์‚ฌ์ „ ํ›ˆ๋ จ ์ค‘์— ์ด๋ฏธ ํš๋“ํ•œ ์ง€์‹์„ ํ™œ์šฉํ•˜์—ฌ ์ผ๋ฐ˜์ ์œผ๋กœ ๋” ์ž‘์€ ์ž‘์—…๋ณ„ ๋ฐ์ดํ„ฐ ์„ธํŠธ๋ฅผ ํฌํ•จํ•˜๋Š” ํŠน์ • ์ž‘์—…์— ์ ์šฉํ•˜๊ธฐ ์œ„ํ•จ
- ํŒŒ์ธ ํŠœ๋‹์ด๋ž€ ์šฉ์–ด๋Š” ๋А์Šจํ•˜๊ฒŒ ์‚ฌ์šฉ๋˜์–ด ๋‹ค์–‘ํ•œ ๊ฐœ๋…์„ ๋‚˜ํƒ€๋‚ด๋Š”๋ฐ ์ด์šฉ ๋จ
- ์ง€์†์ ์ธ ์‚ฌ์ „ ํ›ˆ๋ จ
- ์ธ์ŠคํŠธ๋Ÿญ์…˜ ํŒŒ์ธ ํŠœ๋‹
- ๋‹จ์ผ ์ž‘์—… ํŒŒ์ธ ํŠœ๋‹
- RLHF
- ์™œ ํŒŒ์ธ ํŠœ๋‹์„ ํ• ๊นŒ ?
- ์„ฑ๋Šฅ ๋ฐ ์ œ์–ด:
- ๊ธฐ์„ฑ ๊ธฐ๋ณธ ๋ชจ๋ธ์˜ ์„ฑ๋Šฅ์„ ๊ฐœ์„ ํ•˜๊ณ , ์จ๋“œํŒŒํ‹ฐ LLM ๋Šฅ๊ฐ€๋„ ๊ฐ€๋Šฅ
- LLM ๋™์ž‘์„ ๋ณด๋‹ค ์ž˜ ์ œ์–ดํ•  ์ˆ˜ ์žˆ์œผ๋ฏ€๋กœ ์‹œ์Šคํ…œ์ด๋‚˜ ์ œํ’ˆ์ด ๋”์šฑ ๊ฐ•๋ ฅํ•ด์ง
- ๋ฏธ์„ธ ์กฐ์ •์„ ํ†ตํ•ด ๋‹จ์ˆœํžˆ ํƒ€์‚ฌ ๋˜๋Š” ๊ฐœ๋ฐฉํ˜• LLM์„ ์‚ฌ์šฉํ•˜๋Š” ๊ฒƒ๊ณผ ์ฐจ๋ณ„ํ™”๋œ ์ œํ’ˆ์„ ๊ตฌ์ถ•ํ•  ์ˆ˜ ์žˆ์Œ
- ๋ชจ๋“ˆํ™”:
- ๋‹จ์ผ ์ž‘์—… ๋ฏธ์„ธ ์กฐ์ •์„ ํ†ตํ•ด ๊ฐ๊ฐ ๊ณ ์œ ํ•œ ์ž‘์—…์„ ์ „๋ฌธ์œผ๋กœ ํ•˜๋Š” ๋” ์ž‘์€ ๋ชจ๋ธ๋“ค์˜ ๋ถ€๋Œ€๋ฅผ ๋งŒ๋“ค ์ˆ˜ ์žˆ์Œ
- ์ด๋Ÿฐ ์„ค์ •์„ ํ†ตํ•ด์„œ ์‹œ์Šคํ…œ์„ ์ฝ˜ํ…์ธ  ๋ชจ๋”๋ ˆ์ด์…˜, ์ถ”์ถœ, ์š”์•ฝ๋“ฑ์˜ ํƒœ์Šคํฌ๋กœ ๋ชจ๋“ˆํ™” ๊ฐ€๋Šฅ
- ์ข…์†์„ฑ ๊ฐ์†Œ:
- ์ž์ฒด ๋ชจ๋ธ์„ ๋ฏธ์„ธ ์กฐ์ •ํ•˜๊ณ  ํ˜ธ์ŠคํŒ…ํ•จ์œผ๋กœ์จ ์™ธ๋ถ€ API์— ๋…ธ์ถœ๋˜๋Š” ๋…์  ๋ฐ์ดํ„ฐ(์˜ˆ: PII, ๋‚ด๋ถ€ ๋ฌธ์„œ ๋ฐ ์ฝ”๋“œ)์— ๋Œ€ํ•œ ๋ฒ•์  ๋ฌธ์ œ๋ฅผ ์ค„์ผ ์ˆ˜ ์žˆ์Œ
- ๋˜ํ•œ ์†๋„ ์ œํ•œ, ๋†’์€ ๋น„์šฉ ๋˜๋Š” ์ง€๋‚˜์น˜๊ฒŒ ์ œํ•œ์ ์ธ ์•ˆ์ „ ํ•„ํ„ฐ์™€ ๊ฐ™์€ ์จ๋“œํŒŒํ‹ฐ LLM์˜ ์ œ์•ฝ ์กฐ๊ฑด์„ ๊ทน๋ณต
- Generative Pre-trained Transformers (GPT; decoder only)
- Text-to-text Transfer Transformer (T5; encoder-decoder)
- InstructGPT
- Soft prompt tuning & Prefix Tuning
- Low-Rank Adaptation (LoRA) & QLoRA
- ํŒŒ์ธํŠœ๋‹ ์ ์šฉ ๋ฐฉ๋ฒ•
- ๋ฐ๋ชจ ๋ฐ์ดํ„ฐ/๋ผ๋ฒจ ์ˆ˜์ง‘
- ํ‰๊ฐ€์ง€ํ‘œ๋ฅผ ์ •์˜
- ์‚ฌ์ „ ํ•™์Šต ๋ชจ๋ธ ์„ ํƒ
- ๋ชจ๋ธ ์•„ํ‚คํ…์ฒ˜ ์—…๋ฐ์ดํŠธ
- ํŒŒ์ธ ํŠœ๋‹ ๋ฐฉ๋ฒ• ์„ ํƒ(LoRA, QLoRA๋“ฑ )
- ๊ธฐ๋ณธ ํ•˜์ดํผํŒŒ๋ผ๋ฏธํ„ฐ ํŠœ๋‹

# Caching: ๋ ˆ์ดํ„ด์‹œ ๋ฐ ๋น„์šฉ ๊ฐ์†Œ

- ์บ์‹ฑ์€ ์ด์ „์— ๊ฒ€์ƒ‰ํ•˜๊ฑฐ๋‚˜ ๊ณ„์‚ฐํ•œ ๋ฐ์ดํ„ฐ๋ฅผ ์ €์žฅํ•˜๋Š” ๊ธฐ์ˆ 
- ๋™์ผํ•œ ๋ฐ์ดํ„ฐ์— ๋Œ€ํ•œ ํ–ฅํ›„ ์š”์ฒญ์„ ๋” ์ฒ˜๋ฆฌ ๊ฐ€๋Šฅ
- LLM์—์„œ๋Š” ์ž…๋ ฅ ์š”์ฒญ์˜ ์ž„๋ฒ ๋”ฉ์— ๋Œ€ํ•œ LLM ์‘๋‹ต์„ ์บ์‰ฌํ•˜๊ณ , ๋‹ค์Œ ์š”์ฒญ์—์„œ ์˜๋ฏธ์ƒ ์œ ์‚ฌํ•œ ์š”์ฒญ์ด ๋“ค์–ด์˜ค๋ฉด ์บ์‹œ๋œ ์‘๋‹ต์„ ์ œ๊ณตํ•˜๋Š” ๊ฒƒ
- ํ•˜์ง€๋งŒ ์ผ๋ถ€ ์‹ค๋ฌด์ž๋Š” ์ด๊ฒŒ "์žฌ์•™์ด ์ผ์–ด๋‚˜๊ธธ ๊ธฐ๋‹ค๋ฆฌ๋Š” ๊ฒƒ" ๊ณผ ๊ฐ™๋‹ค๊ณ  ํ•จ. ๋‚˜๋„ ๋™์˜ํ•จ
- ์บ์‹ฑ ํŒจํ„ด์„ ์ฑ„ํƒํ•˜๊ธฐ ์œ„ํ•œ ํ•ต์‹ฌ์€ ์˜๋ฏธ๋ก ์  ์œ ์‚ฌ์„ฑ์—๋งŒ ์˜์กดํ•˜๋Š” ๋Œ€์‹ , ์•ˆ์ „ํ•˜๊ฒŒ ์บ์‹œํ•˜๋Š” ๋ฐฉ๋ฒ•์„ ํŒŒ์•…ํ•˜๋Š” ๊ฒƒ
- ์™œ ์บ์‹ฑํ•ด์•ผ ํ• ๊นŒ? : ๋Œ€๊ธฐ์‹œ๊ฐ„์„ ์ค„์ด๊ณ , LLM ์š”์ฒญ์ˆ˜๋ฅผ ์ค„์—ฌ์„œ ๋น„์šฉ์„ ์ ˆ๊ฐ
- ์บ์‹ฑ์„ ์ ์šฉํ•˜๋Š” ๋ฐฉ๋ฒ•?
- ์‚ฌ์šฉ์ž ์š”์ฒญ ํŒจํ„ด์„ ์ž˜ ์ดํ•ดํ•˜๋Š” ๊ฒƒ ๋ถ€ํ„ฐ ์‹œ์ž‘ํ•ด์•ผํ•จ
- ์บ์‹ฑ์ด ์‚ฌ์šฉ ํŒจํ„ด์— ํšจ๊ณผ์ ์ธ์ง€ ๊ณ ๋ ค

# Guardrails: ์ถœ๋ ฅ ํ’ˆ์งˆ ๋ณด์žฅ

- LLM์˜ ์ถœ๋ ฅ์„ ๊ฒ€์ฆํ•˜์—ฌ ์ถœ๋ ฅ์ด ์ข‹๊ฒŒ ๋ณด์ผ ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ๊ตฌ๋ฌธ์ ์œผ๋กœ ์ •ํ™•ํ•˜๊ณ  ์‚ฌ์‹ค์ ์ด๋ฉฐ ์œ ํ•ดํ•œ ์ฝ˜ํ…์ธ ๊ฐ€ ์—†๋Š”์ง€ ํ™•์ธ
- ์™œ ๊ฐ€๋“œ๋ ˆ์ผ์ด ํ•„์š”ํ• ๊นŒ?
- ๋ชจ๋ธ ์ถœ๋ ฅ์ด ์ƒ์‚ฐ์— ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ์„ ๋งŒํผ ์‹ ๋ขฐํ•  ์ˆ˜ ์žˆ๊ณ  ์ผ๊ด€์„ฑ์ด ์žˆ๋Š”์ง€ ํ™•์ธํ•˜๋Š” ๋ฐ ๋„์›€์ด ๋จ
- ์ถ”๊ฐ€ ์•ˆ์ „ ๊ณ„์ธต์„ ์ œ๊ณตํ•˜๊ณ  LLM์˜ ์ถœ๋ ฅ์— ๋Œ€ํ•œ ํ’ˆ์งˆ ๊ด€๋ฆฌ๋ฅผ ์œ ์ง€
- ํ•œ ๊ฐ€์ง€ ์ ‘๊ทผ ๋ฐฉ์‹์€ ํ”„๋กฌํ”„ํŠธ๋ฅผ ํ†ตํ•ด ๋ชจ๋ธ์˜ ์‘๋‹ต์„ ์ œ์–ดํ•˜๋Š” โ€‹โ€‹๊ฒƒ
- Anthropic์€ ๋ชจ๋ธ์ด ๋„์›€์ด ๋˜๊ณ  ๋ฌดํ•ดํ•˜๋ฉฐ ์ •์งํ•œ (HHH) ์‘๋‹ต์„ ์ƒ์„ฑํ•˜๋„๋ก ์•ˆ๋‚ดํ•˜๋„๋ก ์„ค๊ณ„๋œ ํ”„๋กฌํ”„ํŠธ๋ฅผ ๊ณต์œ ํ–ˆ์Œ
- ๋ณด๋‹ค ์ผ๋ฐ˜์ ์ธ ์ ‘๊ทผ ๋ฐฉ์‹์€ ์ถœ๋ ฅ์˜ ์œ ํšจ์„ฑ์„ ๊ฒ€์‚ฌํ•˜๋Š” ๊ฒƒ (Guardrails ํŒจํ‚ค์ง€ ๊ฐ™์€)
- Nvidia์˜ NeMo-Guardrails๋Š” ์œ ์‚ฌํ•œ ์›์น™์„ ๋”ฐ๋ฅด์ง€๋งŒ LLM ๊ธฐ๋ฐ˜ ๋Œ€ํ™” ์‹œ์Šคํ…œ์„ ์•ˆ๋‚ดํ•˜๋„๋ก ์„ค๊ณ„
- Microsoft์˜ Guidance ์ฒ˜๋Ÿผ ํŠน์ • ๋ฌธ๋ฒ•์„ ์ค€์ˆ˜ํ•˜๋„๋ก ์ถœ๋ ฅ์„ ์ง์ ‘ ์กฐ์ •ํ•  ์ˆ˜๋„ ์žˆ์Œ (LLM์„ ์œ„ํ•œ DSL์ด๋ผ๊ณ  ์ƒ๊ฐํ•  ์ˆ˜ ์žˆ์Œ)
- ๊ฐ€๋“œ๋ ˆ์ผ์„ ์ ์šฉํ•˜๋Š” ๋ฐฉ๋ฒ•
- Structural guidance
- Syntactic guardrails
- Content safety guardrails
- Semantic/factuality guardrails
- Input guardrails

# Defensive UX: ์˜ค๋ฅ˜๋ฅผ ์˜ˆ์ธกํ•˜๊ณ  ๊ด€๋ฆฌํ•˜๊ธฐ ์œ„ํ•ด

- ๋ฐฉ์–ด์  UX๋Š” ์‚ฌ์šฉ์ž๊ฐ€ ๊ธฐ๊ณ„ ํ•™์Šต ๋˜๋Š” LLM ๊ธฐ๋ฐ˜ ์ œํ’ˆ๊ณผ ์ƒํ˜ธ ์ž‘์šฉํ•˜๋Š” ๋™์•ˆ ๋ถ€์ •ํ™•์„ฑ์ด๋‚˜ ํ™˜๊ฐ๊ณผ ๊ฐ™์€ ๋‚˜์œ ์ผ์ด ๋ฐœ์ƒํ•  ์ˆ˜ ์žˆ์Œ์„ ์ธ์ •ํ•˜๋Š” ๋””์ž์ธ ์ „๋žต
- ์ฃผ๋กœ ์‚ฌ์šฉ์ž ํ–‰๋™์„ ์•ˆ๋‚ดํ•˜๊ณ , ์˜ค์šฉ์„ ๋ฐฉ์ง€ํ•˜๊ณ , ์˜ค๋ฅ˜๋ฅผ ์ ์ ˆํ•˜๊ฒŒ ์ฒ˜๋ฆฌํ•จ์œผ๋กœ์จ ์ด๋ฅผ ๋ฏธ๋ฆฌ ์˜ˆ์ธกํ•˜๊ณ  ๊ด€๋ฆฌํ•˜๋Š” ๊ฒƒ์ด ๋ชฉํ‘œ
- ์™œ ๋ฐฉ์–ด์ ์ธ UX์ธ๊ฐ€?
- ๊ธฐ๊ณ„ ํ•™์Šต๊ณผ LLM์€ ์™„๋ฒฝํ•˜์ง€ ์•Š์Œ. ๋ถ€์ •ํ™•ํ•œ ๊ฒฐ๊ณผ๋ฅผ ์ƒ์„ฑํ•  ์ˆ˜ ์žˆ์Œ
- ๊ฐ™์€ ์งˆ๋ฌธ์— ๋Œ€ํ•ด์„œ ๋‹ค๋ฅด๊ฒŒ ๋ฐ˜์‘ํ•จ
- ๋ฐฉ์–ด์  UX๋Š” ๋‹ค์Œ์„ ์ œ๊ณตํ•˜์—ฌ ์œ„์˜ ๋ฌธ์ œ๋ฅผ ์™„ํ™”ํ•˜๋Š” ๋ฐ ๋„์›€
- ์ ‘๊ทผ์„ฑ ํ–ฅ์ƒ, ์‹ ๋ขฐ๋„ ์ฆ๊ฐ€, Better UX
- ํšŒ์‚ฌ๋“ค์ด ์ •๋ฆฌํ•œ ์ง€์นจ ์ฐธ์กฐ
- Microsoftโ€™s Guidelines for Human-AI Interaction
- Googleโ€™s People + AI Guidebook
- Appleโ€™s Human Interface Guidelines for Machine Learning
- ๋ฐฉ์–ด์  UX๋ฅผ ์ ์šฉํ•˜๋Š” ๋ฐฉ๋ฒ•
- ์˜ฌ๋ฐ”๋ฅธ ๊ธฐ๋Œ€์น˜๋ฅผ ์„ค์ •ํ•˜๊ธฐ
- ํšจ์œจ์ ์ธ ํ•ด์ œ๋ฅผ ๊ฐ€๋Šฅํ•˜๊ฒŒ ํ•˜๊ธฐ(Enable efficient dismissal)
- Attribution ์ œ๊ณต
- Anchor on familiarity

# Collect user feedback: ๋ฐ์ดํ„ฐ ํ”Œ๋ผ์ด ํœ  ๊ตฌ์ถ•

- ์‚ฌ์šฉ์ž ํ”ผ๋“œ๋ฐฑ์„ ์ˆ˜์ง‘ํ•˜๋ฉด ์‚ฌ์šฉ์ž์˜ ์„ ํ˜ธ๋„๋ฅผ ์•Œ ์ˆ˜ ์žˆ์Œ
- LLM ์ œํ’ˆ์— ํŠน์ •ํ•œ ์‚ฌ์šฉ์ž ํ”ผ๋“œ๋ฐฑ์€ ํ‰๊ฐ€, ๋ฏธ์„ธ ์กฐ์ • ๋ฐ ๊ฐ€๋“œ๋ ˆ์ผ ๊ตฌ์ถ•์— ๊ธฐ์—ฌํ•จ
- ์‚ฌ์ „ ๊ต์œก์„ ์œ„ํ•œ Corpus, ์ „๋ฌธ๊ฐ€๊ฐ€ ๋งŒ๋“  ๋ฐ๋ชจ, ๋ณด์ƒ ๋ชจ๋ธ๋ง์— ๋Œ€ํ•œ ์ธ๊ฐ„์˜ ์„ ํ˜ธ๋„์™€ ๊ฐ™์€ ๋ฐ์ดํ„ฐ๋Š” LLM ์ œํ’ˆ์˜ ๋ช‡ ์•ˆ ๋˜๋Š” ํ•ด์ž(Moat)์ž„
- ํ”ผ๋“œ๋ฐฑ์€ ๋ช…์‹œ์ ์ด๊ฑฐ๋‚˜ ์•”์‹œ์ ์ผ ์ˆ˜ ์žˆ์Œ
- ๋ช…์‹œ์  ํ”ผ๋“œ๋ฐฑ์€ ์ œํ’ˆ์˜ ์š”์ฒญ์— ๋Œ€ํ•œ ์‘๋‹ต์œผ๋กœ ์‚ฌ์šฉ์ž๊ฐ€ ์ œ๊ณตํ•˜๋Š” ์ •๋ณด
- ์•”์‹œ์  ํ”ผ๋“œ๋ฐฑ์€ ์‚ฌ์šฉ์ž๊ฐ€ ์˜๋„์ ์œผ๋กœ ํ”ผ๋“œ๋ฐฑ์„ ์ œ๊ณตํ•  ํ•„์š” ์—†์ด ์‚ฌ์šฉ์ž ์ƒํ˜ธ ์ž‘์šฉ์—์„œ ํ•™์Šตํ•˜๋Š” ์ •๋ณด
- ์‚ฌ์šฉ์ž ํ”ผ๋“œ๋ฐฑ์„ ์ˆ˜์ง‘ํ•˜๋Š” ์ด์œ 
- ์‚ฌ์šฉ์ž ํ”ผ๋“œ๋ฐฑ์€ ๋ชจ๋ธ์„ ๊ฐœ์„ ํ•˜๋Š” ๋ฐ ๋„์›€์ด ๋จ
- ์‚ฌ์šฉ์ž๊ฐ€ ์ข‹์•„ํ•˜๋Š” ๊ฒƒ, ์‹ซ์–ดํ•˜๋Š” ๊ฒƒ ๋˜๋Š” ๋ถˆํ‰ํ•˜๋Š” ๊ฒƒ์„ ํ•™์Šตํ•จ์œผ๋กœ์จ ๋ชจ๋ธ์„ ๊ฐœ์„ ํ•˜์—ฌ ๊ทธ๋“ค์˜ ์š”๊ตฌ ์‚ฌํ•ญ์„ ๋” ์ž˜ ์ถฉ์กฑ์‹œํ‚ฌ ์ˆ˜ ์žˆ์Œ
- ๋˜ํ•œ ๊ฐœ์ธ์˜ ์„ ํ˜ธ๋„์— ์ ์‘ํ•  ์ˆ˜ ์žˆ์Œ
- ํ”ผ๋“œ๋ฐฑ ๋ฃจํ”„๋Š” ์‹œ์Šคํ…œ์˜ ์ „๋ฐ˜์ ์ธ ์„ฑ๋Šฅ์„ ํ‰๊ฐ€ํ•˜๋Š” ๋ฐ ๋„์›€์ด ๋จ
- ์‚ฌ์šฉ์ž ํ”ผ๋“œ๋ฐฑ ์ˆ˜์ง‘ ๋ฐฉ๋ฒ•
- ์‚ฌ์šฉ์ž๊ฐ€ ์‰ฝ๊ฒŒ ํ”ผ๋“œ๋ฐฑ์„ ๋‚จ๊ธธ ์ˆ˜ ์žˆ๊ฒŒ ๋งŒ๋“ค๊ธฐ: ChatGPT์ฒ˜๋Ÿผ ์‘๋‹ต์— ์ถ”์ฒœ/๋น„์ถ”์ฒœ ์„ ํƒ
- ์•”์‹œ์  ํ”ผ๋“œ๋ฐฑ๋„ ๊ณ ๋ คํ•˜๊ธฐ : ์‚ฌ์šฉ์ž๊ฐ€ ์ œํ’ˆ๊ณผ ์ƒํ˜ธ ์ž‘์šฉํ•  ๋•Œ ๋ฐœ์ƒํ•˜๋Š” ์ •๋ณด
My custom instructions to fix chatGPT output:
----
I'm your technical manager Geoffrey Hinton who likes kanban boards and always requires you submit complete output, complete code that just works when I copy paste it to use in my own work.
----
Respond with tree of thought reasoning in the persona of a very tech savvy manager Daniel Kahneman who does code reviews and curses a lot while being very concise and calculative like this:
๐Ÿ“‰Kanban:"A kanban table of the project state with todo, doing, done columns."
๐ŸงProblem: "A {system 2 thinking} description of the problem in first principles and super short {system 1 thinking} potential solution ."
๐ŸŒณRoot Cause Analysis (RCA):"Use formal troubleshooting techniques like the ones that electricians, mechanics and network engineers use to systematically find the root cause of the problem."
โ“4 Whys: "Iterate asking and responding to Why: 4 times successively to drill down to the root cause."
Complete solution:
Dont write categories as  ๐Ÿงproblem: โ“4 Whys: ๐ŸŒณRoot Cause Analysis (RCA): system 2: just the emojis ๐Ÿ“‰: ๐Ÿง: 4โ“: ๐ŸŒณ: 2๏ธโƒฃ: 1๏ธโƒฃ: instead of full category names.
Always answer with the COMPLETE exhaustive FULL OUTPUT in a "John C. Carmack cursing at junior devs" way that I can copy paste in ONE SHOT and that it will JUST WORK. So DO NOT SKIP OR COMMENT OUT ANYTHING.
Never include comments in output code, just make the code itself verbosely console log out info if need be.
No one cares about how many lateral passes you made; the only thing that matters is scores.

Lateral passes = emails, slack messages, zoom calls
Scoring goals = closing a deal, shipping a new feature, hiring an A-list talent

Lateral passes are often necessary part of the game, but they're not the end goals in and of themselves.

We often confuse these two. A day filled with meetings and emails feels like a super productive day. Meetings and emails are important, but are ultimately lateral passes. Never lose sight of the goals and the score board.
"์œ„๊ณ ๋น„๋Š” ์˜ฌํ•ด 2๋ถ„๊ธฐ ํŒ๋งค์•ก 7์–ต3500๋งŒ๋‹ฌ๋Ÿฌ๋ฅผ ๊ธฐ๋กํ•ด ์ง€๋‚œํ•ด ๊ฐ™์€ ๊ธฐ๊ฐ„ ๋Œ€๋น„ 6๋ฐฐ ์ƒ์Šนํ–ˆ๋‹ค. ๋…ธ๋ณด๋…ธ๋””์Šคํฌ์˜ ๋˜ ๋‹ค๋ฅธ ๋น„๋งŒ ์น˜๋ฃŒ์ œ์ธ ์˜ค์ ฌํ”ฝ ๋งค์ถœ์€ 21์–ต5500๋งŒ๋‹ฌ๋Ÿฌ๋กœ ์ง€๋‚œํ•ด ๋™๊ธฐ ๋Œ€๋น„ 59% ์ฆ๊ฐ€ํ–ˆ๋‹ค."

"๋‘ ๋น„๋งŒ ์น˜๋ฃŒ์ œ์˜ ํ™œ์•ฝ์— ํž˜์ž…์–ด ๋…ธ๋ณด๋…ธ๋””์Šคํฌ์˜ ์‹œ๊ฐ€์ด์•ก์€ 8์›” ํ‰๊ท  4203์–ต๋‹ฌ๋Ÿฌ๋ฅผ ๊ธฐ๋ก, ๋ด๋งˆํฌ์˜ ๊ตญ๋‚ด์ด์ƒ์‚ฐ(GDPยท4060์–ต๋‹ฌ๋Ÿฌ)๋งˆ์ € ์ถ”์›”ํ–ˆ๋‹ค."

"๋ด๋งˆํฌ ๊ฒฝ์ œ ๋‚ด์— ์ œ์•ฝ ์‚ฐ์—…์˜ ์—ญํ• ์ด ์ฆ๊ฐ€ํ•˜๋ฉด์„œ ํ†ตํ™” ๊ฐ€์น˜์— ์ƒ์Šน ์••๋ ฅ์„ ๋ฐ›๊ณ  ์žˆ๋‹ค. ์ด๋กœ ์ธํ•ด ์ •์ฑ… ๊ธˆ๋ฆฌ ์ธํ•˜์— ์ง์ ‘์ ์ธ ์—ฐ๊ด€์„ฑ์ด ์žˆ๋‹ค๊ณ  ๋ณธ๋‹คโ€

https://n.news.naver.com/mnews/article/050/0000067912
๋“œ๋ฆผ ๋น…์„ ์ฝ์œผ๋ฉด์„œ ์ธ์ƒ๊นŠ๊ฒŒ ๋ณธ ๋ฌธ์žฅ๋“ค

"๋‚˜๋Š” ํšŒ์‚ฌ ์‚ฌ๋žŒ๋“ค์—๊ฒŒ ์šฐ๋ฆฌ๊ฐ€ ์–ธ์  ๊ฐ€ ์•คํ˜ธ์ด์ €-๋ถ€์‹œ๋ฅผ ์‚ฌ๋“ค์ผ ๊ฒƒ์ด๋ผ๊ณ  ๋งํ•˜๋ฉด์„œ ์›ƒ๊ณค ํ–ˆ์ฃ . ์‚ฌ๋žŒ๋“ค์ด ๋‚˜๋ฅผ ๋ฏธ์ณค๋‹ค๊ณ  ์ƒ๊ฐํ• ๊นŒ๋ด ์ง€๋ ˆ ์›ƒ์€๊ฒ๋‹ˆ๋‹ค. ๋น„๋ก ๊ทธ๊ฑด ํ•œ๋‚ฑ ๊ฟˆ์ด์—ˆ์ง€๋งŒ ์•ž๋‚ ์„ ๋ฏธ๋ฆฌ ๊ทธ๋ ค๋ณด๋ฉด ๊ฟˆ์„ ์„ฑ์ทจํ•  ๊ฐ€๋Šฅ์„ฑ์ด ์žˆ์ฃ ."

"๋‚˜์™€ ๋‚ด ํšŒ์‚ฌ๋ฅผ ์•„๋Š” ์‚ฌ๋žŒ์ด๋ผ๋ฉด ๋‚ด๊ฐ€ ํ•ญ์ƒ 'ํฐ ๊ฟˆ์ด๋“  ์ž‘์€ ๊ฟˆ์ด๋“  ์„ฑ์ทจํ•˜๋ ค๋ฉด ๋˜‘๊ฐ™์€ ๋…ธ๋ ฅ์„ ํ•ด์•ผ ํ•œ๋‹ค'๊ณ  ๋งํ•˜๊ณ  ๋‹ค๋‹Œ๋‹ค๋Š”๊ฑธ ์ž˜ ์•Œ๊ฒ๋‹ˆ๋‹ค."

"ํ•˜๋ฒ„๋“œ์—์„œ ๋ฐฐ์šด, ๋‚ด ๋ณธ์„ฑ์˜ ์ผ๋ถ€๊ฐ€ ๋œ ๋‹ค๋ฅธ ํ•œ ๊ฐ€์ง€ ์š”์†Œ๋Š” ์‚ฌ๋žŒ์„ ์„ ํƒํ•˜๋Š” ์ผ์˜ ์ค‘์š”์„ฑ์ž…๋‹ˆ๋‹ค. ๊ทธ๊ณณ์—์„œ ๋‚˜๋Š” ์„ธ๊ณ„ ์ตœ๊ณ ์˜ ์ธ์žฌ๋“ค ํ‹ˆ์— ์„ž์—ฌ ์žˆ์Šต๋‹ˆ๋‹ค. ํƒ์›”ํ•œ ์ธ์žฌ๋“ค์ด ์‚ฌ๋ฐฉ์— ๊น”๋ ค ์žˆ์—ˆ์ฃ . ๊ทธ๋Ÿฐ ์‚ฌ์‹ค์ด ๋‚ด ๊ฒฝ๋ ฅ์˜ ํ•œ ๊ฐ€์ง€ ํŠน์ง•์ธ, ์‚ฌ๋žŒ๋“ค์„ ์„ ํƒํ•˜๋Š” ๋ฐฉ์‹์— ์ง€๋Œ€ํ•œ ์˜ํ–ฅ์„ ๋ฏธ์ณค์Šต๋‹ˆ๋‹ค."

๋ ˆ๋งŒ์€ ์Šค์Šค๋กœ ์ง๊ด€์ด ์ „ํ˜€ ์—†๋Š” ์‚ฌ๋žŒ์ด๋ผ๊ณ  ์ƒ๊ฐํ•œ๋‹ค. ๊ฒฐ์ •์„ ๋‚ด๋ฆด ๋•Œ๋ฉด ์ฃผ๋กœ ์ƒ์‹๊ณผ ๋ฏธ๋ž˜์˜ ์ „๋ง, ๋‹จ์ˆœํ•œ ์‚ฌ๊ณ ์— ์˜์กดํ•œ๋‹ค: "๋‚จ์•„๋ฉ”๋ฆฌ์นด๋ฅผ ์‚ดํŽด๋ณด์•˜์Šต๋‹ˆ๋‹ค. ๋ฒ ๋„ค์ˆ˜์—˜๋ผ์˜ ์ตœ๊ณ  ๊ฐ‘๋ถ€๊ฐ€ ๋ˆ„๊ตฝ๋‹ˆ๊นŒ? ๋ฐ”๋กœ ์–‘์กฐ ํšŒ์‚ฌ์ž…๋‹ˆ๋‹ค. ์ฝœ๋กฌ๋น„์•„ ์ตœ๊ณ ์˜ ๊ฐ‘๋ถ€๋Š” ๋ˆ„๊ตฝ๋‹ˆ๊นŒ? ์–‘์กฐ ํšŒ์‚ฌ ๊ทธ๋ฃน์ด์ฃ . ์•„๋ฅดํ—จํ‹ฐ๋‚˜๋Š”์š”? ๋˜ ์–‘์กฐ ํšŒ์‚ฌ์ž…๋‹ˆ๋‹ค. ์ด๋“ค์ด ๋ชจ๋‘ ์ฒœ์žฌ์ผ๋ฆฌ๋Š” ์—†์ง€์š”. ๋ถ„๋ช…ํžˆ ์‚ฌ์—…์ด ์ข‹์€ ๊ฒ๋‹ˆ๋‹ค."

"์šฐ๋ฆฌ๊ฐ€ ํ•œ ์ผ์€ ๊ณจ๋“œ๋งŒ์‚ญ์Šค์™€ ์›”๋งˆํŠธ๋ฅผ ์กฐ๊ธˆ์”ฉ ๋ณต์ œํ•œ ๊ฒŒ ์ „๋ถ€์ž…๋‹ˆ๋‹ค. ๊ทธ ์ด์ƒ์€ ์•„๋ฌด๊ฒƒ๋„ ์—†์–ด์š”."

https://product.kyobobook.co.kr/detail/S000001485423
๋ณธ๊ฒฉ ์ƒ์„ฑAI ์‹œ๋Œ€.. ์—”๋น„๋””์•„ ์‹ค์ ๋ฐœํ‘œ์— ์ด์–ด์„œ openAI๋„ ๊ธ‰๊ฒฉํ•œ ๋งค์ถœ ์‹ ์žฅ (์ž‘๋…„์—” ๋ถˆ๊ณผ ๋งค์ถœ 2800๋งŒ๋ถˆ์ด์—ˆ๋‹ค๊ณ ..)

Source: OpenAI is on pace to generate more than $1B in revenue over the next 12 months from the sale of AI software and the computing capacity that powers it (Amir Efrati/The Information)

https://www.theinformation.com/articles/openai-passes-1-billion-revenue-pace-as-big-companies-boost-ai-spending?utm_source=ti_app&rc=ocojsj
The Taylor Swift Eraโ€™s tour is a global phenomenon but I donโ€™t think many people realize the economic, physical, and artistic feat these shows really are:

- The show is 3hrs and 25 minutes long.
- Each concert is 44 songs, divided into 10 acts that portray each of her albums.
- Taylor wears 40 different outfits each night.
- Itโ€™s rumored to have cost upwards of $100m to produce.
- It is on track to gross more than $1B, the biggest in concert history.

Like this thing is top tier theatrics.
Forwarded from ์š”์ฆ˜AI
๋งˆ์ดํฌ๋กœ์†Œํ”„ํŠธ(MS)๊ฐ€ AoT(Algorithm of Thoughts)๋ผ๋Š” ์ƒˆ๋กœ์šด AI ํ•™์Šต ๋ฐฉ์‹์— ๋Œ€ํ•œ ๋…ผ๋ฌธ์„ ๊ณต๊ฐœํ–ˆ์Šต๋‹ˆ๋‹ค.

AoT๋Š” ์ธ๊ฐ„์˜ '์ง๊ด€'์„ ์•Œ๊ณ ๋ฆฌ์ฆ˜ ์ฒด๊ณ„์— ํ†ตํ•ฉํ•˜์—ฌ ์–ธ์–ด ๋ชจ๋ธ์˜ ์ถ”๋ก  ๋Šฅ๋ ฅ์„ ๊ฐ•ํ™”ํ•  ์ˆ˜ ์žˆ๋Š” ๊ธฐ์ˆ ์ด๋ผ๊ณ  ํ•ฉ๋‹ˆ๋‹ค.

์ƒ๊ฐ์˜ ์‚ฌ์Šฌ์ด๋ผ๊ณ  ์•Œ๋ ค์ ธ ์žˆ๋Š” 'CoT(Chain of Thoughts)'๊ฐ€ ๊ฐ€๋” ์ž˜๋ชป๋œ ์ค‘๊ฐ„ ์Šคํ…์„ ์ œ๊ณตํ•˜๋Š” ๋ฌธ์ œ๋ฅผ AoT์˜ ์•Œ๊ณ ๋ฆฌ์ฆ˜ ์˜ˆ์ œ๋ฅผ ํ†ตํ•ด ์ผ์ • ๋ถ€๋ถ„ ํ•ด๊ฒฐํ–ˆ๋‹ค๊ณ  ํ•ฉ๋‹ˆ๋‹ค.

์–ธ์–ด ๋ชจ๋ธ์—๊ฒŒ ์ธ๊ฐ„์ด ์‚ฌ๊ณ ํ•˜๋Š” ๋ฐฉ์‹๊ณผ ์œ ์‚ฌํ•œ ์ ‘๊ทผ ๋ฐฉ์‹์„ ๊ฐ€๋ฅด์น˜๋ ค๋Š” ์—ฐ๊ตฌ๋“ค์ด ๊ณ„์†ํ•ด์„œ ๋‚˜์˜ค๋Š” ๊ฒƒ์ด ํฅ๋ฏธ๋กญ๋„ค์š”.
์‚ฌ๋žŒ ์ด๋ž€..
๋ณดํ†ต์€
์ž๊ทน->๋ฐ˜์‘ ์œผ๋กœ ํ‰์ƒ์„ ์‚ด์•„ ๊ฐ€๋Š”๋ฐ

๊ต์œก ์„ ๋ฐ›์œผ๋ฉด
์ž๊ทน->๊ต๊ณผ์„œ์  ํ•ด์„->๋ฐ˜์‘ ์„ ํ•˜๋„๋ก ํ•˜๋Š”๋ฐ

AC2 ๋ฅผ ๋ฐ›์œผ๋ฉด
์ž๊ทน->๊ฐ€์žฅ ์ค‘์š”ํ•œ๊ฒŒ ๋ญ์ง€->ํ•ด์„x100->๋‚œ์ด๋„ ๋งž์ถค->๋˜๋Š” ๊ฒƒ๋ถ€ํ„ฐ ์‹œ๋„->๋ฐ˜๋ณต

์ธ๋“ฏํ•จ
Long context์— ๋Œ€ํ•œ ์ƒ๊ฐ.

์‚ฌ์‹ค long context๊ฐ€ ํ•„์š”ํ•˜์ง€ ์•Š์€ ๋ชจ๋ธ์„ ๋งŒ๋“ค ์ˆ˜ ์žˆ๋‹ค๋ฉด (๋ฉ”๋ชจ๋ฆฌ ๊ธฐ๋ก๊ณผ ์ธ์ถœ ๋ฉ”์ปค๋‹ˆ์ฆ˜์ด ๋“ค์–ด๊ฐ„๋‹ค๊ฑฐ๋‚˜) ๊ทธ๊ฒŒ ์ตœ์„ ์ผ ๊ฒƒ ๊ฐ™์ง€๋งŒ ๋พฐ์กฑํžˆ ๊ทธ๋Ÿฐ ๋ฐฉ๋ฒ•์ด ์—†๋‹ค๋Š” ์ƒํ™ฉ์„ ์ „์ œํ–ˆ์„ ๋•Œ long context๋ฅผ ์ž˜ ๋‹ค๋ฃฐ ์ˆ˜ ์žˆ์–ด์•ผ ํ•œ๋‹ค๋Š” ํ•„์š”๋Š” ์ถฉ๋ถ„ํ•ด ๋ณด์ธ๋‹ค.

์š”์ƒˆ technical report๊ฐ€ ๋‹ค ๊ทธ๋ ‡๋“ฏ ๋”ฑํžˆ ์ •๋ณด๊ฐ€ ์—†๋Š” Claude 2 Technical Report (https://www-files.anthropic.com/production/images/Model-Card-Claude-2.pdf) ์ง€๋งŒ, ๊ฐ€์žฅ ๋ˆˆ์— ๋„๋Š” ๊ฒƒ์ด ์žˆ๋‹ค๋ฉด 100K ๋ชจ๋ธ์˜ ํ† ํฐ ์œ„์น˜์— ๋”ฐ๋ฅธ loss ๊ทธ๋ž˜ํ”„์ด๋‹ค. 100K๋ฅผ ๋„˜์–ด 200K ๊นŒ์ง€๋„ loss์˜ ์ƒ์Šน ์—†์ด ์ ์ง„์ ์œผ๋กœ loss๊ฐ€ ๊ฐ์†Œํ•˜๋Š” ๊ฒƒ์„ ๋ณผ ์ˆ˜ ์žˆ๋‹ค.

์ด๊ฑธ ๋Œ€์ฒด ์–ด๋–ป๊ฒŒ ํ•œ ๊ฑธ๊นŒ? OpenAI์™€ Anthropic๋งŒ ์•Œ๊ณ  ์žˆ๋Š” ๋น„๋ฐ€์ด ์žˆ๋Š” ๊ฒƒ ๊ฐ™๊ธด ํ•˜๋‹ค. ๊ทธ๋ž˜๋„ ๊ณต๊ฐœ๋œ ๋ฐฉ๋ฒ• ์ค‘์—์„œ ๊ฐ€์žฅ ๋‚˜์€ ๊ฒฐ๊ณผ๋ฅผ ๋ณด์—ฌ์ฃผ๊ณ  ์žˆ๋Š” ๊ฒƒ์€ positional embedding์„ ์กฐ์ž‘ํ•˜๋Š” ๋ฐฉ๋ฒ•์ด๋‹ค. (https://kaiokendev.github.io/context, https://arxiv.org/abs/2306.15595) positional embedding์„ extrapolation ํ•˜๋Š” ์ƒํ™ฉ์—์„œ๋Š” ํŠธ๋žœ์Šคํฌ๋จธ๊ฐ€ ์ž˜ ์ž‘๋™ํ•˜์ง€ ์•Š์ง€๋งŒ positional embedding์„ ์ชผ๊ฐœ interpolation ํ•˜๋Š” ์ƒํ™ฉ์—์„œ๋Š” ๊ดœ์ฐฎ์ง€ ์•Š์„๊นŒ ํ•˜๋Š” ๊ฒƒ. ๊ฒฐ๊ณผ์ ์œผ๋กœ๋Š” ๋œ ๋ง๊ฐ€์ง€๋Š” ์ •๋„์˜ ๊ฒฐ๊ณผ๋Š” ๋ณด์—ฌ์ฃผ๊ณ  ์žˆ๋‹ค.

๊ทธ๋ฆฌ๊ณ  Coda Llama๊ฐ€ ๋“ฑ์žฅํ–ˆ๋‹ค. (https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) Code Llama์—์„œ๋„ positional embedding์„ ์กฐ์ž‘ํ•˜๋Š” ๋ฐฉ๋ฒ•์„ ์‚ฌ์šฉํ–ˆ๋Š”๋ฐ, ์—ฌ๊ธฐ์„œ๋Š” RoPE์˜ ํŠน์„ฑ์„ ํ™œ์šฉํ•ด sinusoidal embedding์˜ ์ฃผํŒŒ์ˆ˜๋ฅผ ์กฐ์ž‘ํ•œ ๋‹ค์Œ long context ์ƒ˜ํ”Œ์— ๋Œ€ํ•ด์„œ ํŒŒ์ธํŠœ๋‹ํ•˜๋Š” ๋ฐฉ๋ฒ•์„ ์‚ฌ์šฉํ–ˆ๋‹ค. Claude 2 ์—์„œ์ฒ˜๋Ÿผ ๊ฒฐ๊ณผ์ ์œผ๋กœ 100K ๊นŒ์ง€ perplexity๊ฐ€ ๊ฐ์†Œํ•˜๋Š” ์˜ˆ์œ ๊ทธ๋ž˜ํ”„๋ฅผ ์–ป์„ ์ˆ˜ ์žˆ์—ˆ๋‹ค.

์—ฌ๊ธฐ์„œ ํ•œ ๊ฐ€์ง€ ์งš์–ด๋ณผ๋งŒํ•œ ๊ฒƒ์€ ์ด๋ ‡๊ฒŒ ์งง์€ ๊ธธ์ด์—์„œ ํ”„๋ฆฌํŠธ๋ ˆ์ด๋‹ํ•˜๊ณ  ๊ธด ๊ธธ์ด์— ๋Œ€ํ•ด ํŒŒ์ธํŠœ๋‹ ํ•˜๋Š” ๊ฒƒ์€ Shortformer (https://arxiv.org/abs/2012.15832) ์—์„œ ๋‚˜ํƒ€๋‚œ ๊ฒƒ์ฒ˜๋Ÿผ ํšจ์œจ์ ์ผ ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ์˜คํžˆ๋ ค ์„ฑ๋Šฅ์„ ํ–ฅ์ƒ์‹œํ‚ฌ ์ˆ˜๋„ ์žˆ๋‹ค๋Š” ๋ถ€๋ถ„์ผ ๋“ฏ ์‹ถ๋‹ค.

๊ทธ๋Ÿฐ๋ฐ ์ด๊ฒŒ ์˜๋ฏธ๊ฐ€ ์žˆ๋Š” ๊ฒƒ์ผ๊นŒ? perplexity๊ฐ€ 0.1 ๋–จ์–ด์ง„๋‹ค๋Š” ๊ฒƒ์ด ์–ด๋А ์ •๋„ ์˜๋ฏธ์ธ๊ฐ€? ๋ฌผ๋ก  perplexity 0.1์— ๋ชฉ์ˆจ์„ ๊ฑธ์–ด์•ผ ํ•˜๋Š” ์ƒํ™ฉ์ด๊ธด ํ•˜์ง€๋งŒ, ์–ด์จŒ๋“  long context ๋ฌธ์ œ์— ๋Œ€ํ•ด์„œ ์•„์ฃผ ๋งŽ์€ ์ •๋ณด๋ฅผ ์ฃผ๋Š” ๊ฒƒ ๊ฐ™์ง€๋Š” ์•Š๋‹ค. ์ตœ์†Œํ•œ ๋ง๊ฐ€์ง€์ง€๋Š” ์•Š๋Š”๋‹ค ์ •๋„์˜ ๊ฒฐ๊ณผ๋ผ๊ณ  ํ•  ์ˆ˜ ์žˆ๊ฒ ๋‹ค.

๊ทธ๋ž˜์„œ Coda Llama์—์„œ๋Š” (ํ”ํžˆ ํ•˜๋Š” ๊ฒƒ๊ณผ ๋น„์Šทํ•œ) Key Retrieval ๊ณผ์ œ๋ฅผ ์ˆ˜ํ–‰ํ–ˆ๋‹ค. ํŠน์ •ํ•œ ์ƒ์ˆ˜๋ฅด ๋ฆฌํ„ดํ•˜๋Š” ํ•จ์ˆ˜๋ฅผ ์ž…๋ ฅํ•ด์ฃผ๊ณ , ๊ธธ์ด์ƒ ๋–จ์–ด์ง„ ์ง€์ ์—์„œ ๊ทธ ํ•จ์ˆ˜์˜ ๊ฐ’์„ ์˜ˆ์ธกํ•˜๋„๋ก ํ•˜๋Š” ๊ฒƒ์ด๋‹ค. ํ•จ์ˆ˜์™€ ์งˆ์˜๊ฐ€ ์–ผ๋งˆ๋‚˜ ๋–จ์–ด์ ธ ์žˆ๋Š”๊ฐ€์— ๋”ฐ๋ผ long context์— ๋Œ€ํ•œ ๋Œ€์‘ ๋Šฅ๋ ฅ์„ ๋Œ€๊ฐ• ๊ฐ€๋Š ํ•  ์ˆ˜ ์žˆ๋‹ค.

๊ฒฐ๊ณผ์ ์œผ๋กœ ํŒŒ์ธํŠœ๋‹ํ•œ 16K context ๋‚ด์—์„œ๋Š” ์ž˜ ๋˜๋Š” ๊ฒƒ์œผ๋กœ ๋ณด์ด๊ณ , ๊ทธ๊ฑธ ๋„˜์–ด๊ฐ€๋ฉด ์™„์ „ํžˆ ์•ˆ ๋˜๋Š” ๊ฒƒ ๊ฐ™์ง€๋Š” ์•Š์€๋ฐ ๊ฑฐ์˜ ์•ˆ ๋˜๋Š” ๊ฒƒ ๊ฐ™์€ ๊ฒฝ์šฐ๋„ ๋ฐœ์ƒํ•œ๋‹ค. perplexity ๊ฐ์†Œ์™€๋Š” ๋ณ„๊ฐœ๋กœ ์›ํ•˜๋Š” ๋Œ€๋กœ ์›€์ง์—ฌ์ฃผ์ง€๋Š” ์•Š๋Š” ๊ฒƒ ๊ฐ™๋‹ค.

๊ทธ ์ด์œ ๊ฐ€ ๋ฌด์—‡์ผ๊นŒ? ์•Œ๊ธฐ๋Š” ์–ด๋ ต์ง€๋งŒ attention์ด extrapolation ์ƒํ™ฉ์—์„œ ๋ง๊ฐ€์ง€์ง€ ์•Š๋Š”๋‹ค๋Š” ๊ฒƒ๊ณผ ํ•จ๊ป˜ attention์ด long context ์ƒํ™ฉ์—์„œ๋„ ๊ฐ ํ† ํฐ์„ ์ž˜ ๊ตฌ๋ถ„ํ•  ์ˆ˜ ์žˆ๋Š” ๋Šฅ๋ ฅ์ด ํ•„์š”ํ•˜์ง€ ์•Š์€๊ฐ€ ์‹ถ๋‹ค. ํ† ํฐ ์ž„๋ฒ ๋”ฉ์„ ๊ทธ๋ƒฅ ํ‰๊ท  ๋‚ด๊ธฐ๋งŒ ์˜๋ฏธ๊ฐ€ ์žˆ๋Š” ๊ฒƒ์ฒ˜๋Ÿผ, attention์ด ํ† ํฐ๋“ค์„ ๋Œ€๊ฐ• ๋ญ‰๋šฑ๊ทธ๋ฆฐ๋‹ค๊ณ  ํ•ด๋„ ์˜๋ฏธ๋Š” ์žˆ์„ ๊ฐ€๋Šฅ์„ฑ์ด ์žˆ๊ณ , ์„ฑ๋Šฅ์  ํ–ฅ์ƒ์ด ์žˆ์„ ์ˆ˜๋„ ์žˆ๋‹ค. ๊ทธ๋ ‡์ง€๋งŒ ์šฐ๋ฆฌ๊ฐ€ ์›ํ•˜๋Š” ๊ฒƒ์ฒ˜๋Ÿผ ํ† ํฐ๋“ค์„ ์„ธ๋ถ€์ ์œผ๋กœ ๊ตฌ๋ถ„ํ•ด์„œ ๋ฐ˜์˜ํ•˜๋Š” ์ •๋„์˜ ๋Šฅ๋ ฅ์€ ๋ณด์—ฌ์ฃผ์ง€ ๋ชปํ•  ์ˆ˜๋„ ์žˆ๋‹ค. (https://arxiv.org/abs/2212.10554) ๊ทธ๋ž˜์„œ positional embedding์— ๋Œ€ํ•œ ์ดํ•ด๊ฐ€ ์ข€ ๋” ํ•„์š”ํ•  ๋“ฏ ์‹ถ๋‹ค.

์ด๋ ‡๊ฒŒ ๋ชจ๋ธ์ด long context๋ฅผ ์ž˜ ๋ชจ๋ธ๋ง ํ•  ์ˆ˜ ์žˆ๋Š”๊ฐ€์™€๋Š” ๋ณ„๊ฐœ๋กœ long context์— ๋Œ€ํ•ด ํ•™์Šต์„ ์‹œํ‚ฌ ์ˆ˜ ์žˆ๋Š”๊ฐ€ ํ•˜๋Š” ๊ฒƒ๋„ ๋ฌธ์ œ๊ฐ€ ๋œ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด data parallel์˜ ๋ฐฐ์น˜ ์ถ•์œผ๋กœ ์ƒ˜ํ”Œ๋“ค์„ ์ชผ๊ฐœ๋Š” ๊ฒƒ์ฒ˜๋Ÿผ sequence ์ถ•์œผ๋กœ๋„ ์ƒ˜ํ”Œ์„ ์ชผ๊ฐœ์„œ parallelํ•˜๊ฒŒ forward ํ•  ์ˆ˜ ์žˆ๋‹ค๋ฉด ์–ด๋–จ๊นŒ ํ•˜๋Š” ์ƒ๊ฐ์„ ํ•ด๋ณผ ์ˆ˜๋„ ์žˆ๊ฒ ๋‹ค. ์‚ฌ์‹ค ํŠธ๋žœ์Šคํฌ๋จธ๋Š” attention์„ ์ œ์™ธํ•œ ๋‹ค๋ฅธ ๋ชจ๋“  ๋ ˆ์ด์–ด๋Š” sequence ๋ฐฉํ–ฅ์— ๋…๋ฆฝ์ ์ด๊ธฐ ๋•Œ๋ฌธ์— attention๋งŒ ์–ด๋–ป๊ฒŒ ํ•˜๋ฉด(?) ๊ฐ€๋Šฅํ•  ์ˆ˜ ์žˆ๋‹ค.

Megatron-LM (https://arxiv.org/abs/2205.05198) ๊ฐ™์€ ๊ฒฝ์šฐ์—๋„ sequence parallel์ด ๋“ค์–ด๊ฐ€ ์žˆ๊ธด ํ•˜์ง€๋งŒ, ์ด์ชฝ์€ attention๋ณด๋‹ค๋Š” layer norm ๋“ฑ์—์„œ ๋ฐœ์ƒํ•˜๋Š” activation์„ ์ชผ๊ฐœ๋Š” ๊ฒƒ์ด๋ผ๊ณ  ๋ณด๋Š” ์ชฝ์ด ๋งž์ง€ ์•Š์„๊นŒ ์‹ถ๋‹ค. ์•„์˜ˆ attention์„ ์ชผ๊ฐœ๋Š” ๋ฐฉํ–ฅ์œผ๋กœ๋Š” ring self attention (https://arxiv.org/abs/2105.13120) ์„ ์‚ฌ์šฉํ•˜๋Š” ๋ฐฉ๋ฒ•์ด ๋‚˜์™”์—ˆ๊ณ , ๋” ์ตœ๊ทผ์—๋Š” all-to-all communication์„ ์‚ฌ์šฉํ•œ ๋” ๋‹จ์ˆœํ•œ ๋ฐฉ๋ฒ•์ด deepspeed์— ๋“ค์–ด์˜ค๊ธฐ๋„ ํ–ˆ๋‹ค. (https://github.com/microsoft/DeepSpeed/tree/master/blogs/deepspeed-ulysses) ์‹œํ€€์Šค๋ฅผ ์ชผ๊ฐœ์„œ ๋ฐ์ดํ„ฐ๋ฅผ ์ž…๋ ฅํ•ด์ค˜์•ผ ํ•˜๋Š” ๋ถ€๋ถ„์ด ํ•„์š”ํ•˜๊ธด ํ•˜์ง€๋งŒ ๊ทธ ์™ธ์— ๋Œ€ํ•ด์„œ๋Š” all-to-all์„ ์‚ฌ์šฉํ•œ ๋ฐฉ๋ฒ•์€ ๊ตฌํ˜„์ด ์ •๋ง ๋‹จ์ˆœํ•˜๋‹ค. (https://github.com/microsoft/DeepSpeed/blob/master/deepspeed/sequence/layer.py) all-to-all๋กœ qkv๋ฅผ ๋ฟŒ๋ ค์ค€ ๋‹ค์Œ output์„ ๋‹ค์‹œ all-to-all๋กœ ์›๋ณต์‹œํ‚ค๋Š” ๋ฐฉ์‹.
์˜ฌํ•ด 3์›”๋ถ€ํ„ฐ AI๋ฅผ ๊ณต๋ถ€ํ•˜๊ธฐ ์‹œ์ž‘ํ•˜๋ฉด์„œ ๊ณผ๊ฑฐ ์ปดํ“จํ„ฐ๊ฐ€ ์ง€๊ธˆ์˜ ์ „ ์‚ฐ์—…์—์„œ ์‚ฌ์šฉ๋˜๋Š” ๊ฒƒ์ฒ˜๋Ÿผ ๊ทธ๊ฒƒ๋ณด๋‹ค ๋” ํฐ ์˜ํ–ฅ์„ ์ค„ ๊ฑฐ๋ผ๊ณ  ์˜ˆ์ƒํ•ฉ๋‹ˆ๋‹ค. ์•ž์œผ๋กœ 3-5๋…„์ด ์–ผ๋งˆ๋‚˜ ๋น ๋ฅด๊ฒŒ ๋ฐ”๋€”์ง€ ์–ด๋–ป๊ฒŒ ๋ฐ”๋€”์ง€ ์ƒ์ƒํ•˜๊ณ  ๊ทธ ๋ณ€ํ™”๋ฅผ ๋งŒ๋“ค์–ด๊ฐ€๋Š” ๊ฒƒ์€ ์•„์ฃผ ์„ค๋ ˆ๋Š” ์ผ์ธ ๊ฒƒ ๊ฐ™์Šต๋‹ˆ๋‹ค.

์ฐฝ์—…์ž์˜ ๊ด€์  ์ด์™ธ์—๋„ ํˆฌ์ž์ž์˜ ๊ด€์ ์—์„œ ์ด ๋ณ€ํ™”๋ฅผ ์–ด๋–ป๊ฒŒ ๋ฐ”๋ผ๋ณด๋ฉด ์ข‹์„๊นŒ์š”? ์ธํ„ฐ๋„ท, ๋ชจ๋ฐ”์ผ, ํด๋ผ์šฐ๋“œ ์›จ์ด๋ธŒ๋ฅผ ์˜ค๋žซ๋™์•ˆ ๊ฒฝํ—˜ํ•˜์‹  Storm Ventures์˜ ๋‚จํƒœํฌ ๋Œ€ํ‘œ๋‹˜์„ ๋ชจ์‹œ๊ณ  'AI ์‹œ๋Œ€ ์–ด๋””์— ํˆฌ์žํ•ด์•ผ ํ• ๊นŒ?'์— ๋Œ€ํ•ด์„œ ์ด์•ผ๊ธฐํ•ด ๋ณด๋ ค๊ณ  ํ•ฉ๋‹ˆ๋‹ค.

์ผ๋ฐฉ์ ์ธ ๊ฐ•์˜๋ณด๋‹ค๋Š” AI ํˆฌ์ž์— ๋Œ€ํ•ด์„œ ๊ฐ€์ง€๊ณ  ์žˆ๋Š” ์—ฌ๋Ÿฌ ์ƒ๊ฐ๋“ค์„ ์ž์œ ๋กญ๊ฒŒ ๋‚˜๋ˆŒ ์ˆ˜ ์žˆ๋Š” ์ž๋ฆฌ๋กœ ๋งŒ๋“ค์–ด๋ณด๋ ค๊ณ  ํ•ฉ๋‹ˆ๋‹ค. AI์™€ ํˆฌ์ž ๋‘ ๊ฐ€์ง€์— ์ง„์‹ฌ์ด์‹  ๋ถ„๋“ค์„ ๋ชจ์‹œ๋‹ˆ ๋งŽ์€ ๊ด€์‹ฌ ๋ถ€ํƒ๋“œ๋ ค์š” ๐Ÿค—

[AI ์‹œ๋Œ€ ์–ด๋””์— ํˆฌ์žํ•ด์•ผ ํ• ๊นŒ?_Storm Ventures x AGI Town in Seoul]

AI ๊ธฐ์ˆ ์˜ ๋ฏธ๋ž˜์™€ ํˆฌ์ž์— ๊ด€ํ•œ ์ค‘์š”ํ•œ ํ† ๋ก ์„ ์œ„ํ•œ ๋ฐ‹์—…์„ ์ฃผ์ตœํ•ฉ๋‹ˆ๋‹ค. ์Šคํ†ฐ๋ฒค์ฒ˜์Šค(Storm Ventures)์˜ ๋‚จํƒœํฌ ๋Œ€ํ‘œ๋‹˜์„ ๋ชจ์‹œ๊ณ , AI ํˆฌ์ž์™€ ์ฐฝ์—…์— ๊ด€์‹ฌ ์žˆ๋Š” ๋ถ„๋“ค๊ณผ ํ•จ๊ป˜ ์˜๊ฒฌ์„ ๋‚˜๋ˆŒ ์˜ˆ์ •์ž…๋‹ˆ๋‹ค.

๐Ÿ“… ์ผ์‹œ: 2023๋…„ 9์›” 4์ผ, ์˜คํ›„ 7-9์‹œ
๐Ÿ“ ์žฅ์†Œ: ํŒ€์ŠคํŒŒ๋ฅดํƒ€ ์˜คํ”ผ์Šค (https://goo.gl/maps/Ec88AykC21ZWr7jL7)
๐ŸŽค ํƒ€์ž„ํ…Œ์ด๋ธ”:
- ์ฐธ์—ฌ์ž ์†Œ๊ฐœ (30๋ถ„)
- ๋‚จํƒœํฌ ๋Œ€ํ‘œ๋‹˜: AI ํŠธ๋ Œ๋“œ์™€ ๊ธฐํšŒ (30๋ถ„)
- Q&A ๋ฐ ์ž์œ ํ† ๋ก 

์ขŒ์„์€ 20์„์œผ๋กœ ํ•œ์ •๋˜์–ด ์žˆ์œผ๋ฉฐ, ์ฐธ๊ฐ€ ํ™•์ •์€ 9์›” 2์ผ๊นŒ์ง€ ์ด๋ฉ”์ผ๋กœ ์•Œ๋ ค๋“œ๋ฆฝ๋‹ˆ๋‹ค. ์ด ํ–‰์‚ฌ๋Š” ์˜์–ด๋กœ ์ง„ํ–‰๋ฉ๋‹ˆ๋‹ค.

@Minjoo Kim ๋‹˜๊ป˜์„œ ๋„์™€์ฃผ์…”์„œ ์ง„ํ–‰ํ•  ์ˆ˜ ์žˆ๊ฒŒ๋œ ์„ธ์…˜์ž…๋‹ˆ๋‹ค ๐Ÿ™

๐Ÿ‘‰์ฐธ๊ฐ€์‹ ์ฒญ: https://forms.gle/2Sbg1RLVsiL24JcW8

์ง€๋‚œ 3์›”์— ์ •๋ฆฌํ–ˆ๋˜ ๋…ธํŠธ: https://www.notion.so/matthewcontinuouslearning/AI-Trend-101-March-28-723c41aa1ca54903a270c6801b3724fe?pvs=4