Links for 2024-06-05

AI Predictions:

1. Microsoft CTO Kevin Scott says what he's seeing in early previews of forthcoming AI models are systems with memory and reasoning at a level that can pass PhD qualifying exams https://www.youtube.com/watch?v=b_Xi_zMhvxo

2. Leopold Aschenbrenner: “By 2025-2026, we’re going to get models that are basically smarter than most college graduates…By 2027-2028, it’ll get as smart as the smartest experts.” https://www.dwarkeshpatel.com/p/leopold-aschenbrenner (Also: “OpenAI leadership had a plan laid out to fund and sell AGI by starting a bidding war between the governments of the United States, China, and Russia.”)

3. When will the first weakly general AI system be devised, tested, and publicly announced? Median (1.35k forecasters): 2027 https://www.metaculus.com/questions/3479/date-weakly-general-ai-is-publicly-known/

4. When will an AI win a Gold Medal in the International Math Olympiad? Median (256 forecasters): 2027 https://www.metaculus.com/questions/6728/ai-wins-imo-gold-medal/

5. Will an AI be able to reliably construct bug-free code of more than 10,000 lines before 2030? Community prediction (256 forecasters): 89% https://www.metaculus.com/questions/11188/ai-as-a-competent-programmer-before-2030/

6. Takeoff speeds presentation at Anthropic https://www.lesswrong.com/posts/Nsmabb9fhpLuLdtLE/takeoff-speeds-presentation-at-anthropic

7. ‘Everything is Going to Be Robotic’ Nvidia Promises, as AI Gets More Real https://www.youtube.com/watch?v=nxO_t5N82m0

AI Research:

1. Evidence of Learned Look-Ahead in a Chess-Playing Neural Network https://www.lesswrong.com/posts/GfwdBoaLw3ef3zBqe/evidence-of-learned-look-ahead-in-a-chess-playing-neural

2. “We empirically show that a GPT-style transformer exhibits a transition from in-distribution to out-of-distribution generalization as the number of pre-training tasks increases.” https://arxiv.org/abs/2406.02550

3. Show, Don’t Tell: Significantly outperforms few-shot prompting, SFT and other self-play methods by using demonstrations as feedback directly https://github.com/SALT-NLP/demonstrated-feedback

4. To Believe or Not to Believe Your LLM (Quantify uncertainty to detect hallucinations) https://arxiv.org/abs/2406.02543

5. Artificial Generational Intelligence: Cultural Accumulation in Reinforcement Learning https://arxiv.org/abs/2406.00392

6. Learning Temporally Consistent Video Depth from Video Diffusion Priors https://arxiv.org/abs/2406.01493

7. “Codestral solved a problem in two messages that I couldn't resolve with bouncing around between GPT4o, GPT-4 and Claude Opus for an hour.” https://www.reddit.com/r/LocalLLaMA/comments/1d86di0/codestral_solved_a_problem_in_two_messages_that_i/

https://x.com/TimothyDuignan/status/1797960946243223850

Miscellaneous:

1. Spectacular new state of art for Brain-Computer Interfaces: After 5 recording sessions the word error rate was <10% and after 15 sessions, consistent <5% WER for online decoding. Stable performance 250 days after implantation. https://www.medrxiv.org/content/10.1101/2023.12.26.23300110v2

2. “Quantinuum’s H-Series hits 56 physical qubits that are all-to-all connected, and departs the era of classical simulation” https://www.quantinuum.com/news/quantinuums-h-series-hits-56-physical-qubits-that-are-all-to-all-connected-and-departs-the-era-of-classical-simulation

3. How a simple circuit could offer an alternative to energy-intensive GPUs https://www.technologyreview.com/2024/06/05/1093250/how-a-simple-circuit-could-offer-an-alternative-to-energy-intensive-gpus/ [archived: https://archive.is/bRDWQ]

4. how to find a location using only the shadow https://www.youtube.com/watch?v=pQIjDPFgdJA

5. New map shows vast potential for geothermal energy beneath entire US https://thehill.com/policy/energy-environment/4703057-geothermal-energy-potential-beneath-western-eastern-us-map/

Ukraine:

1. Gazprom badly hurt by Ukraine war, says company-commissioned report https://www.ft.com/content/21f8f63f-80d6-455f-abf8-fce269d70319 [no paywall: https://archive.is/Zdd0T]

2. “Russian MinFin's stats shows that the government paid 201,7 billion worth of fuel subsidies to fuel producers in May 2024 or 94,9% more than the 103,5 billion rubles in May 2023. The subsidies are paid so that Russian fuel prices remain low.” https://x.com/delfoo/status/1798302242438885609

3. Ukraine has lost 9.2GW in power generation capacity since Russia's new airstrike campaign began in March. 1.2GW of generation was lost just in Saturday's attack, and infrastructure for transporting gas from underground storage facilities was badly damaged. https://www.ft.com/content/4d583259-7565-4cbc-972e-ea77f4a76175 [no paywall: https://archive.is/l1AwJ]