What's even better than #GPT4 ? Your own personal fine-tuned GPT-4! OpenAI released some new fine-tuning features...read more in my latest InfoQ news.

The Blackwell superchip from #NVIDIA is the largest GPU ever built, with over 200 billion transistors, and can train large language models (LLMs) up to 4x faster than previous generation hardware. Read more in my latest InfoQ news!

Meta is working on the next iteration of their Llama #LLM , and they've designed a giant GPU cluster to train it. Read more about the design in my latest InfoQ news!

Worried about #jailbreak attacks on your #LLM ? SafeDecoding outperforms baseline jailbreak defenses without incurring significant computational overhead. Read more in my latest InfoQ news!

They say that something like 99% of ancient literature has been lost. Maybe the Vesuvius Challenge can bring some back to life...learn how they are using AI to read ancient scrolls in my latest InfoQ news!

Another 'small' language model: Eagle 7B. Interestingly, it uses an attention-free Transformer that combines the benefits of both Transformers and RNNs...for example the model has no maximum input context length. Read more in my latest InfoQ news!

Amazon's new BASE TTS text-to-speech AI supports voice-cloning and outperforms baseline TTS models when evaluated by human judges. Read more in my latest InfoQ news!

Google's new forecasting AI, TimesFM, has zero-shot forecasting performance comparable to or better than supervised-learning models. Learn more in my latest InfoQ news.

As *THE* Bard once said: 'What's in a name?' Google's #Bard is now #Gemini ! Read more in my latest InfoQ news

There's more to AI than chatbots and image generation. MIT scientists are using it to discover new antibiotics to fight MRSA. Read more in my latest InfoQ news!

Lots of new updates from #OpenAI , including an improved #GPT4 model. Read more in my latest InfoQ news!

When humanity does create AGI, it will be named Untitled14.ipynb

Forget #LLM large language models! The new trend is *small* language models, like Stability AI's 1.6B parameter Stable LM, which is 'easy to fine tune on your MacBook.' Read more in my latest InfoQ news!

Check out our new InfoQ podcast, where Roland Meertens and I talk about AI, generally...this first episode is about #LLMs !

Looking for an open-source #LLM that can outperform GPT-3.5? Learn more about Mixtral 8x7B in my latest InfoQ news

Google Announces Video Generation LLM VideoPoet bit.ly/3O3VOOy by Anthony Alford

How is OpenAI handing risks posed by their models? Read about their Preparedness Framework in my latest InfoQ news!

Are autonomous vehicles safe? #Waymo publishes a report showing that their AVs have fewer crashes per million miles than human drivers. Read more in my latest InfoQ news?

OpenAI's #PromptEngineering guide gives you six strategies for eliciting better responses from their GPT models. Read more in my latest InfoQ news!

Microsoft's new (small) #LLM , Phi-2, outperforms models which are up to 25x larger. Read more in my latest InfoQ news!