Lädt...


🔧 OpenAI o3 - Thinking Fast and Slow


Nachrichtenbereich: 🔧 Programmierung
🔗 Quelle: dev.to

OpenAI has teased the o3 model today—a further development of the "reasoning" model and a successor to o1.

I was impressed by how much it improved on the ARC-AGI-1 benchmark - a supposedly unbeatable benchmark by the current generation of LLMs. o1 's high-score was 32% while o3 jumped right at 88%. The authors of the Arc Challenge ($1mil reward for beating ARC-AGI) were quite confident that transformer-based models won't be successful in their benchmark - they were not impressed with o1. Yet, the o3 blog post is a completely different sentiment with words such as "surprising", "novel" and "breakthrough".

o3 has reminded me of an analogy often mentioned when discussing LLMs. No matter the complexity of the task GPTs use the same amount of compute/time per token as if they are streaming info from their subconscious without ever stopping to think. This is similar to how the "Fast" System 1 of the human brain.

A quick recap, "Thinking Fast and Slow" is a 2011 book by Daniel Kahneman. He argues that functionally (based on empirical research) our brain has 2 departments (or modes):

  • System 1, Fast - effortless, autonomous, associative.
  • System 2, Slow - effortful, deliberate, logical.

The 2 systems work together and shape humans' thinking processes. We can read a book out loud without any stress, yet we might not remember a single word. We can read the book and be focused, constantly replaying the scenes and pictures in our mind, keeping track of events and timelines, and be exhausted after a short period—yet we might acquire new knowledge.

As Andrew Ng once noted, "Try typing a text without ever hitting backspace" - seems like a hard task, and that is how LLMs work.

Well, that's how they worked until recently. When o1 (and later Deepseek Q, QwQ, Gemini 2.0 Flash Thinking) appeared the models learned to make a break and operate in a mode similar to the "Slow" system.

Recently there has been a lot of talk of LLM pre-training plateauing, exhausting training data, AI development hitting a wall.

We might be seeing a forming trend on what comes in 2025 - combining resigning/thinking models with traditional models, interconnecting them as Slow and Fast minds: planning (Slow) and taking action (Fast), identifying (Fast) and evaluating (Slow) etc.

Here's one of the recent examples from Aider AI coding assistant which shows how combining QwQ as Architect and Qwen 2.5 as a Coder (there's a 2-step "architect-code" mode allowing to choose different models for each of the steps) increases coding performance.

Whether it will play out - it's hard to say. There are plenty of challenges that we haven't seen a lot of progress lately, even with Slow models. It's unclear how models such as o3 will be tolerant to hallucinations. The context window is still too small. The prices are going up... The slow models, while they hit the next levels of different "isolated" evals, are far from practical application at scale (doing large projects on their own OR simulating a junior intern). Additionally the Fast models, the actors, it doesn't seem they have shown progress in computer use and Moravec's paradox is still a challenge when it comes to automating a computer clerk.

...

📰 Slow iPhone: Check How Slow Your Device Is


📈 20.9 Punkte
🖥️ Betriebssysteme

🪟 Windows Insider Fast and Slow rings rebranded as Dev and Beta channels


📈 18.34 Punkte
🪟 Windows Tipps

📰 How to Fix “Too Many Requests, Please Slow Down” on ChatGPT OpenAI


📈 17.12 Punkte
📰 IT Security Nachrichten

🔧 Interviewing, Fast and Slow: Use Behavioral Science to Ace Tech Interviews


📈 17.12 Punkte
🔧 Programmierung

🔧 JavaScript on the Desktop, Fast and Slow


📈 17.12 Punkte
🔧 Programmierung

🪟 Windows 10 build 18362.30 hits Fast and Slow ring Insiders with three fixes


📈 17.12 Punkte
🪟 Windows Tipps

🪟 Windows 10 20H1 build 19033 rolls out to Fast and Slow ring Insiders


📈 17.12 Punkte
🪟 Windows Tipps

🪟 Windows 10 build 19035 rolls out to Fast and Slow ring Insiders


📈 17.12 Punkte
🪟 Windows Tipps

🪟 Windows 10 20H1 build 19037.1 heads to Fast and Slow ring Insiders


📈 17.12 Punkte
🪟 Windows Tipps

🍏 Why Is My Mac Running Slow and How To Make It Fast Again


📈 17.12 Punkte
🍏 iOS / Mac OS

🔧 Fast and Slow Pointers, Coding Interview Pattern


📈 17.12 Punkte
🔧 Programmierung

🔧 Cracking the Coding Interview: Part 4 – The Fast and Slow Pointer Technique


📈 17.12 Punkte
🔧 Programmierung

📰 Fast & Furious Filmreihe: 2 Fast 2 Furious alias Fast 2 startet im Free-TV


📈 16.32 Punkte
📰 IT Nachrichten

📰 Fast & Furious Filmreihe: Fast 5 und Fast 3 im TV und Stream sehen


📈 16.32 Punkte
📰 IT Nachrichten

📰 Microsoft: Office Insider jetzt auch in Fast und Slow unterteilt


📈 15.89 Punkte
📰 IT Nachrichten

📰 Microsoft: Office Insider jetzt auch in Fast und Slow unterteilt


📈 15.89 Punkte
📰 IT Nachrichten

📰 Intel adopts Orwellian irony with call for fast Meltdown-Spectre action after slow patch delivery


📈 15.89 Punkte
📰 IT Security Nachrichten

📰 'Flippy,' the Fast Food Robot, Turned Off For Being Too Slow


📈 15.89 Punkte
📰 IT Security Nachrichten

📰 Microsoft stellt neue Windows Insider-Builds im Fast & Slow Ring bereit


📈 15.89 Punkte
📰 IT Security Nachrichten

📰 Microsoft stellt neue Windows Insider-Builds im Fast & Slow Ring bereit


📈 15.89 Punkte
📰 IT Nachrichten

🪟 Windows 10 Insider Build 19033 steht für Fast und Slow-Ring bereit


📈 15.89 Punkte
🪟 Windows Tipps

📰 Windows 10 20H1 wird Version 2004: Build 19033 im Fast und Slow Ring verfügbar


📈 15.89 Punkte
📰 IT Nachrichten

📰 Windows 10 20H1 wird Version 2004: Build 19033 im Fast und Slow Ring verfügbar


📈 15.89 Punkte
📰 IT Nachrichten

🪟 Windows 10 Insider Build 19035 für Slow und Fast Ring veröffentlicht


📈 15.89 Punkte
🪟 Windows Tipps

📰 Windows 10 Insider Build 19035 für Slow und Fast Ring veröffentlicht


📈 15.89 Punkte
📰 IT Nachrichten

📰 Windows 10: Neues für die Insider im Fast und Slow Ring


📈 15.89 Punkte
📰 IT Nachrichten

📰 Windows 10: Neues für die Insider im Fast und Slow Ring


📈 15.89 Punkte
📰 IT Nachrichten

📰 Too slow! Booking.com fined for not reporting data breach fast enough


📈 15.89 Punkte
📰 IT Security Nachrichten

🪟 Fix: Slow Internet on One Computer But Fast on Another


📈 15.89 Punkte
🪟 Windows Tipps

📰 Chechnya Is Banning Music That's Too Fast Or Slow


📈 15.89 Punkte
📰 IT Security Nachrichten

matomo