• AIPressRoom
  • Posts
  • 11 Main AI Developments: RT-2 to ‘100X GPT-4’

11 Main AI Developments: RT-2 to ‘100X GPT-4’

September 14, 2023

There have been 11 main developments in AI in simply this week, from RT-2 – a major step on the trail to robotic AGI – to ‘100x GPT 4 in 18 months’, and from the uplifting information of TranscribeGlass and AI Barbie-heimer to dramatic revelations about OpenAI within the Atlantic.

We’ll additionally glimpse Secure Beluga 2 – the ‘first true competitor’ to ChatGPT, based mostly on the open supply Llama 2, hear about Common Jailbreaks, study that OpenAI has surrendered on Textual content Detection, plus I’ll cowl the highlights of the Senate listening to on AI.

Chapters:0:18 – RT-22:46 – 100X GPT-43:57 – AI Video4:29 – Altman Atlantic8:41 – Jan Leike Interview10:02 – Speech Transcription + Technology11:07 – OpenAI Textual content Give up11:43 – Secure Beluga 212:51 – Common Jailbreaks14:19 – Senate testimony: Bio16:59 – Senate Testimony: Safety

RT-2: https://www.deepmind.com/blog/rt-2-new-model-translates-vision-and-language-into-actionRT-2 Paper: https://robotics-transformer2.github.io/assets/rt2.pdfRT-2 NYT: https://www.nytimes.com/2023/07/28/technology/google-robots-ai.htmlSuleyman Barron’s Interview: https://www.barrons.com/articles/ai-chatbot-siri-alexa-inflection-pi-fa1809f8OpenAI Atlantic: https://www.theatlantic.com/magazine/archive/2023/09/sam-altman-openai-chatgpt-gpt-4/674764/TranscribeGlass: https://twitter.com/igorsushko/status/1684645290132000768Barbie-heimer: https://twitter.com/terhavlova/status/1684265329277403149Runway Gen 2: https://twitter.com/runwayml/status/1666429706932043776ElevenLabs: https://elevenlabs.io/speech-synthesisJan Leike Interview: https://www.lesswrong.com/posts/bsNXqHgiDA6dAKNun/24-superalignment-with-jan-leikeSecure Beluga 2: https://stability.ai/blog/stable-beluga-large-instruction-fine-tuned-modelsOpen LLM Leaderboard: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboardQuick Takeoff: https://www.alignmentforum.org/posts/YgNYA6pj2hPSDQiTE/distinguishing-definitions-of-takeoff#:~:textual content=associatedpercent20withpercent20testing.-,Quick,casepercent20inpercent20Yudkowsky’spercent20foompercent20scenario.OpenAI Textual content Detection Give up: https://techcrunch.com/2023/07/25/openai-scuttles-ai-written-text-detector-over-low-rate-of-accuracy/Common Jailbreaks: https://www.nytimes.com/2023/07/27/business/ai-chatgpt-safety-research.htmlhttps://llm-attacks.org/https://twitter.com/mustafasuleymn?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5EauthorSenate Listening to Amodei: https://www.youtube.com/watch?v=hm1zexCjELoAltman Query from Philip @ AI Defined in Melbourne: https://www.youtube.com/watch?v=7SMzkBKzsQsWired Bio Knowledgeable: https://www.wired.com/story/have-a-nice-future-podcast-15/