t2d2

let the wookiee win

[t2d2 @a4travis]

⬅️shot on iPhone
➡️shot on Samsung

[t2d2 @a4travis]

@dunemovie 👀

[t2d2 @a4travis]

@MetaQuestVR is it oven safe?

[t2d2 @a4travis]

@nemchan_nel @JumiusV After all, IRL, we’re often bound by compromises and rules that stop us from fully being ourselves. So maybe metaverse is closer to the "original universe."

[t2d2 @a4travis]

@nemchan_nel @JumiusV in your survey, avatar (VR identity) represents a better version of yourself. Could this version actually be the true and authentic you?

[t2d2 @a4travis]

@nemchan_nel @JumiusV I love 超宇宙 and 超界. They give off such cool manga vibes!

When it comes to 元宇宙, I feel like it leans more toward a "digital twin" philosophy. Why bother creating another universe, right?

[t2d2 @a4travis]

💕

[バーチャル美少女ねむ/Nem⚡アバター国際標準化委員系VTuber @nemchan_nel]

【速報】
なんとHTC様運営の中国語メディア「VIVE 後浪潮」に定性調査 #メタバースでのアイデンティティ を大きく取り上げて頂きました！ 謝謝！メタバースは「元宇宙」なんですね

可愛是重點？2024年元宇宙身分認同調查報告出爐，5個重點整理這邊看 https://t.co/U8vxiSoJVV

[t2d2 @a4travis]

@nemchan_nel @yuyake_hino @VR_Girl_NoraCat @BredikhinaL @kusahaelv @maguraeneR18 @Kerin_Vtuber @KeiroKamioka @nekomantle @JumiusV @patsupyon 宇宙 means "universe," so we combined "meta" and "verse" to create this new term. Still, it can be a bit tricky to grasp the concept in Chinese. BTW, I’m the editor from VIVE 後浪潮. Thanks for sharing—this is a great survey!

[t2d2 @a4travis]

Star Realm | Join My World: https://t.co/aO03PcxIon https://t.co/EcaGBDsgIf

[t2d2 @a4travis]

Star Realm | Join My World: https://t.co/aO03PcxIon https://t.co/2LNmNOmQ67

[t2d2 @a4travis]

@Miulahung 如果左派就等於共產主義

[t2d2 @a4travis]

Interesting times.
Apple has published its research on the LLM model MM1, is considering Gemini as the default AI assistant on iPhones, and is collaborating with NVIDIA on the Vision Pro.

[t2d2 @a4travis]

Apple considered buying Tesla, but the deal fell through when Elon wanted Tim Cook's job at Apple.

Elon attempted to bring OpenAI under Tesla's umbrella. When they refused, he stopped the funding and left OpenAI.

Imagine a parallel universe where he is the boss of...🤯

[t2d2 @a4travis]

@NASAWebb looks like a wormhole to me

[Jim Fan @DrJimFan]

In my decade spent on AI, I've never seen an algorithm that so many people fantasize about. Just from a name, no paper, no stats, no product. So let's reverse engineer the Q* fantasy. VERY LONG READ:

To understand the powerful marriage between Search and Learning, we need to go back to 2016 and revisit AlphaGo, a glorious moment in the AI history.
It's got 4 key ingredients:

1. Policy NN (Learning): responsible for selecting good moves. It estimates the probability of each move leading to a win.

2. Value NN (Learning): evaluates the board and predicts the winner from any given legal position in Go.

3. MCTS (Search): stands for "Monte Carlo Tree Search". It simulates many possible sequences of moves from the current position using the policy NN, and then aggregates the results of these simulations to decide on the most promising move. This is the "slow thinking" component that contrasts with the fast token sampling of LLMs.

4. A groundtruth signal to drive the whole system. In Go, it's as simple as the binary label "who wins", which is decided by an established set of game rules. You can think of it as a source of energy that *sustains* the learning progress.

How do the components above work together?

AlphaGo does self-play, i.e. playing against its own older checkpoints. As self-play continues, both Policy NN and Value NN are improved iteratively: as the policy gets better at selecting moves, the value NN obtains better data to learn from, and in turn it provides better feedback to the policy. A stronger policy also helps MCTS explore better strategies.

That completes an ingenious "perpetual motion machine". In this way, AlphaGo was able to bootstrap its own capabilities and beat the human world champion, Lee Sedol, 4-1 in 2016. An AI can never become super-human just by imitating human data alone.

-----
Now let's talk about Q*. What are the corresponding 4 components?

1. Policy NN: this will be OAI's most powerful internal GPT, responsible for actually implementing the thought traces that solve a math problem.

2. Value NN: another GPT that scores how likely each intermediate reasoning step is correct.
OAI published a paper in May 2023 called "Let's Verify Step by Step", coauthored by big names like @ilyasut @johnschulman2 @janleike: https://t.co/iAvXNjjhcK
It's much lesser known than DALL-E or Whipser, but gives us quite a lot of hints.

This paper proposes "Process-supervised Reward Models", or PRMs, that gives feedback for each step in the chain-of-thought. In contrast, "Outcome-supervised reward models", or ORMs, only judge the entire output at the end.

ORMs are the original reward model formulation for RLHF, but it's too coarse-grained to properly judge the sub-parts of a long response. In other words, ORMs are not great for credit assignment. In RL literature, we call ORMs "sparse reward" (only given once at the end), and PRMs "dense reward" that smoothly shapes the LLM to our desired behavior.

3. Search: unlike AlphaGo's discrete states and actions, LLMs operate on a much more sophisticated space of "all reasonable strings". So we need new search procedures.

Expanding on Chain of Thought (CoT), the research community has developed a few nonlinear CoTs:
- Tree of Thought: literally combining CoT and tree search: https://t.co/KM1P2ZJrjG @ShunyuYao12
- Graph of Thought: yeah you guessed it already. Turn the tree into a graph and Voilà! You get an even more sophisticated search operator: https://t.co/5ncT5tuTOY

4. Groundtruth signal: a few possibilities:
(a) Each math problem comes with a known answer. OAI may have collected a huge corpus from existing math exams or competitions.
(b) The ORM itself can be used as a groundtruth signal, but then it could be exploited and "loses energy" to sustain learning.
(c) A formal verification system, such as Lean Theorem Prover, can turn math into a coding problem and provide compiler feedbacks: https://t.co/vpOBOI2FR5

And just like AlphaGo, the Policy LLM and Value LLM can improve each other iteratively, as well as learn from human expert annotations whenever available. A better Policy LLM will help the Tree of Thought Search explore better strategies, which in turn collect better data for the next round.

@demishassabis said a while back that DeepMind Gemini will use "AlphaGo-style algorithms" to boost reasoning. Even if Q* is not what we think, Google will certainly catch up with their own. If I can think of the above, they surely can.

Note that what I described is just about reasoning. Nothing says Q* will be more creative in writing poetry, telling jokes @grok, or role playing. Improving creativity is a fundamentally human thing, so I believe natural data will still outperform synthetic ones.

I welcome any thoughts or feedback!!

[t2d2 @a4travis]

But Oedipus, far from dying, ends up killing his father and marrying his mother. (Sam, far from being down, leads all employees to Microsoft, and eventually AGI really does doom humanity) If only they hadn't abandoned their son... (If only they hadn't fired Sam...)

[t2d2 @a4travis]

The whole OpenAI saga is straight out of Oedipus: The king, heeding the oracle's prophecy, fears his son Oedipus will kill him. (Ilya worries Sam will cause a AGI uprising) So, he abandons his son in the wild to die. (So, a coup is staged to oust Sam from the board)

[t2d2 @a4travis]

OpenAI is nothing without @sama

[t2d2 @a4travis]

@emilychangtv I think they should hire @elonmusk