Science fiction? Far from it. Researchers recently created the Darwin Gödel Machine, which is “a self-improving system that iteratively modifies its own code.” https://
bit.ly/4mvLKND #AI #selfevolvingAI #ArtificialIntelligence #Technology #TechNews
RESEARCH
-
Darwin Gödel Machine: Self-Improving AI System Breakthrough
By
–
-

Meta Restructures AI; Musk Claims Grok 5 Approaches AGI
By
–
Top stories in AI today: – Meta’s massive AI restructure
– Google analyzes Gemini’s environmental footprint
– Automate performance reviews and PIPs
– Musk: Grok 5 has ‘a shot at being true AGI’ – 4 new AI tools, community workflows, and more Read more: https://
therundown.ai/p/metas-major-
ai-restructure
… -

Sharing AI and Robotics Insights at Deep Learning Indaba Kigali
By
–

Had a great time sharing some thoughts with the folks @iafrikan on AI, Robotics and Research yesterday. The atmosphere at @DeepIndaba in Kigali 🇷🇼 is absolutely electrifying! #DLI2025 #Indaba2025 #Urunana
→ View original post on X — @shakir_za, 2025-08-22 08:25 UTC
-

Make Your AI More Sustainable: Energy Efficiency Challenge
By
–
𝐋𝐞ç𝐨𝐧 𝟳 – 𝗘𝗽𝗶𝘀𝗼𝗱𝗲 𝟯 𝗖𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲 : 𝗥𝗲𝗻𝗱𝗲𝘇 𝘃𝗼𝘁𝗿𝗲 𝗜𝗔 𝗽𝗹𝘂𝘀 𝘃𝗲𝗿𝘁𝗲 Les coûts environnementaux de l’IA explose. Il est grand temps de challenger votre IA sur le terrain de la sobriété. Réduire l’empreinte énergétique de l’IA, cela n'est pas
-
Small models could topple AI scaling narrative
By
–
If small language models outperform LLMs in practice, the whole AI scaling narrative might collapse overnight.
-

Google Gemini vs Mistral: Environmental Impact Analysis
By
–
The Verge publishes a piece criticizing some of these measures: https://
theverge.com/report/763080/
google-ai-gemini-water-energy-emissions-study
… And Mistral's audited report for their older model gave larger amounts of water (50 mL) and carbon emissions (1.14g) per average query. https://
mistral.ai/news/our-contr
ibution-to-a-global-environmental-standard-for-ai
… -
Environmental Impact of Large-Scale AI Models and Energy Consumption
By
–
Google paper on Gemini: https://
services.google.com/fh/files/misc/
measuring_the_environmental_impact_of_delivering_ai_at_google_scale.pdf
… Sam Altman post: https://
blog.samaltman.com/the-gentle-sin
gularity
… Google search energy in 2008: https://
googleblog.blogspot.com/2009/01/poweri
ng-google-search.html
… Llama 3.3 power usage: -
Energy consumption metrics for Llama 3.3 and GPT-4 training
By
–
These numbers match independent direct measures: 0.00004 kWh for 400 tokens on Llama 3.3 70B on a H100 node. We do not know the amount of energy required to train these models, which was estimated at a little above 500,000 kWh for GPT-4, about 18 hours of a Boeing 737 in flight.
-

AI Model Energy Efficiency Improves 33x Year-over-Year
By
–
We have data on the environmental impact per AI prompt:
Gemini: 0.00024 kWh & 0.26 mL water
ChatGPT: 0.0003 kWh & 0.38 mL
…the same energy as one Google search in 2008 & 6 drops of water. Seems to be improving, too: Google reports a 33x drop in energy use per prompt in a year. -
Google DeepMind Launches Genie 3: Revolutionary Generative AI Model
By
–
Congrats @jparkerholder
, @shlomifruchter
, & the Genie & Veo teams! If you are interested to know more, the latest episode of the @GoogleDeepMind Podcast with the brilliant @FryRsquared has just dropped, and is all about Genie 3 & its incredible potential: