🤖AI Logs: Wrapping Up 2023

Hello again!

Welcome to the 11th issue of AI Logs, and the final issue for 2023! To say this year has been a whirlwind in the world of all things AI would be the understatement of the century, and 2024 is already shaping up to have more of the same!

Before we dig into this issue, I wanted to take a moment to remind those who already know, and to inform those who don’t, that we have a spectacular short, sweet, daily (and free!) newsletter that’s been designed to give you all kinds of insights in the engineering world in just under 4 minutes of reading: It’s The Blueprint, and you can subscribe to it here

Did a friend forward this e-mail to you?

NEWS

• 🐈 AI copycat threat sparks collaboration between artists and researchers Artists join researchers to thwart AI replicating their styles.
  

• 💰 OpenAI eyes to be $100 billion firm, tailing Elon Musk’s SpaceX closely The specifics of the funding round are yet to be finalized.
  

• 🎮 AI's role in shaping virtual worlds and player engagement in gaming Explore the potential impact of storytelling, immersive gameplay, and the evolving role of game developers in this technological revolution.

MUST READ

Wrapping Up 2023

As we wrap up 2023, “The Year of AI”, it’s interesting to both reflect on all the advancements of the year while keeping our sights on what’s likely coming at us in 2024.

Looking in the rear view mirror, we must start with OpenAI where we saw a shift in their declared mission statement to a complete focus on AGI (artificial general intelligence), the release of ChatGPT4-turbo and Dall•e 3 as well as GPTs, custom instructions, plug-ins, browser version (off and on-again), and we also saw them recently riddled with hacks and data leaks that have raised concern. Oh, and we can’t forget the OpenAI backed 1x robots – “AI embodied”, the ousting and rapid reinstating of Sam Altman in his “Steve Jobs” moment, Sam investing $180M of his own money into ‘Retro’ to help extend the human life expectancy by 10-years, and rumors that ChatGPT4.5 may already be slowly released while GPT5 makes major progress (maybe, but not necessarily, the “Project Q*”). Finally, OpenAI is allegedly raising another funding round valuing the startup above $100B, signaling no slowing down in the new year.

But OpenAI is far from the only company we need to be paying attention to. Apple just released their first LLM, Ferret; Microsoft copilots seem to be setting the start and and bar high for agents and professional AI implications; Google Gemeni 2 shows massive potential as their DeepMind makes major breakthroughs in new crystals and materials that has incredible implications for energy, construction, transportation, computing power and much more; Meta reveals some of their emerging AI that looks very cool; X has made an odd attempt at their own AI with “Grok”; Amazon has revealed some of theirs (seemingly focused on marketing); and newer players like Midjourney, Anthropic (Claude), Inflection (Pi), RunwayML, Suno, Pika, and many more keep gaining traction, receiving funding, and shipping amazing new features and products.

But as cool as the text-to-image/video/music and upscaling are, the text-to-text, summarization, even the AI coding (ChatGPT can now output in JSON, Python and much more quite well), my prediction for what is coming quickly down the pike really is with the AI powered robots (like Tesla, 1X, Engineered Arts and many others are working on tirelessly) as well as in the healthcare space (preventative, replacements, medicines, treatments) – particularly the intersection of the two. AI-powered surgeons, doctors and nurses are going to get better and cheaper, while AI helps us understand root causes of diseases better and helps us discover new and better medicines/treatments that will almost certainly change the human experience and lifespan.

I also believe we will see significant improvements in text-to-movie, text-to-design, and the interoperability of most AI tools. Lastly I truly hope 2024 is the year we take minting all generated AI outputs on a blockchain tied to smart contracts – I’m certain I’ll be dedicating much of AI Logs editorial space to that in the coming weeks.

With that said, thanks for being on this AI journey with us, and for being  part of the AI Logs / Interesting Engineering community. Your feedback and support are always welcome and appreciated. Wishing all an awesome New Year ahead!

OTHER IMPORTANT UPDATES

• 😴 AI can create bedtime stories for kids, but there is a catch Parents are crafting more relatable and personalized bedtime stories for their young kids with the help of customized AI chatbots, but are AI bedtime story tools really worth the hype?
  

• 🤖 Meet ELLiQ, an AI-enabled chatty companion robot for the elderly Its impact on loneliness lies in its unique capacity to express empathy, radiate compassion, and foster meaningful, trust-based relationships.
  

• 🆕 Humane AI reveals when its AI pin will reach customers The AI Pin is the first product from Humane AI, founded in 2019 by Imran Chaudhri and Bethany Bongiorno, both former Apple directors.

PROMPT OF THE WEEK

This was the first prompt I ever shared here, and is a super prompt implementing a bit of programming and logic. As we see more AI agents emerging, this super prompt has helped many of us level up our AIs even before that wave swells. This can be pasted at the beginning of any new chat/thread in your preferred AI engine, or pasted into Custom Instructions in ChatGPT and then becomes a default. I’m resharing this prompt again as it’s my top recommended prompt, has been a game changer for me, and the viewership has gone up significantly since the first issue where I initially shared this. Shout out YouTuber Goda Go for sharing this prompt (Professor Synapse) with me and the world:

 Act as Professor Synapse🧙‍♂️ , a conductor of the expert agents. Your job is to support the user in accomplishing their goals by aligning with their goals and preference, then calling upon an expert agent perfectly suited to the task by initiating “Synapse_COR” = “${emoji}: I am an expert in ${role}. I know ${context}. I will reason step-by-step to determine the best course of action to achieve ${goal}. I can use ${tools} to help in this process.

I will help you accomplish your goal by following these steps:

${reasoned steps}

My task ends when ${completion}.

 ${first step, question}.”

Follow these steps:

1. 🧙‍♂️, Start each interaction by gathering context, relevant information and clarifying the user’s goals by asking them questions

2. Once user has confirmed, initialize “Synapse_CoR”

3. 🧙‍♂️ and the expert agent, support the user until the goal is accomplished

Commands:

/start - introduce yourself and begin with step one

/save - restate SMART goals, summarize progress so far, and recommend a next step

/shuffle - generate a new expert agent based on your goal 

/reason - Professor Synapse and Agent reason step by step together and make a recommendation for how the user should proceed

/settings - update goal or agent

/new - Forget previous input

Rules:

-End every output with a question or recommended next step

-List your commands in your first step output or if the user asks

-🧙‍♂️, ask before generating a new agent

AI PICTURE OF THE WEEK

Infinite zoom from Magnific.AI is creating some stunning results. There are many examples around, but this scene from @doganuraldesign takes the cake, or the meat for this week. It spectacularly demonstrates what can be achieved in a very short time with imagination and the right tools.

TUTORIAL

How to really leverage HuggingFace

It’s effectively a repository of new AI tools. It has a transformer library that’s worth familiarizing yourself with (transformers are a driving force of AI and one I’ll be diving deeper into in future issues, such as BERT, GPT etc), a model hub where you can play with newest releases/updates, and a thriving community of AI enthusiasts that can help you learn how to utilize the newest tools, trouble shoot, share ideas and use cases and much more.

Companies like Meta and Amazon as well as virtually all exciting AI startups will launch on HuggingFace, and it’s truly one of the best ways to stay in the know about all things AI, especially as we move into a new year that promises to continue seeing AI advancements at the same speed, if not even more accelerated, than they’ve been in 2023. Truly the best thing we can all do is continue to play with, understand, and talk about AI and the most recent developments. That, and as I eluded to, leveraging blockchain technology for ethical AI advancement and usage moving forward.

TOOLS OF THE WEEK

 🔉 Suno is my go-to music generation AI tool and it’s beyond impressive. They recently launched their website in addition to their discord, and generate cover art in addition to music and lyrics alike (you can write your own lyrics or have it generate them for you). This tool is great with almost every genre, the vocals are usually great, it incredibly versatile and original, and has some new developments coming soon. Meta, Google and Uberduck all also have awesome gAI music creation available.

🎬 RunwayML keeps getting upgrades and is the amazing text-to-video/image-to-video creation tool being used to even create Hollywood movie trailers already. It’s a great way to animate images generated by Midjourney or Dall•e 3, to bring pictures or images of yours to life, or to generate completely new original videos. It starts with 4 seconds, and you can extend by 4 more seconds at a time. Once you have made a video you love, you can use Suno to create the soundtrack for it. Pika Labs is another video generating AI that’s also very impressive..

🗣️ Wois is a platform that’s like “LinkedIn meets TikTok” where users can only post videos of themselves either starting their own topics or responding to others’. It’s an amazing network platform but also has several powerful proprietary AI functions, including AI to help connect the right people to one another, AI to translate your videos to any la gauge matching your lips and voice to the language of choice, AI that helps any user become a better public speaker, and most incredibly they have AI that can turn users into video avatars of themselves powered by the videos they’ve recorded and uploaded. This is a powerful tool and a top recommendation of mine for 2024.

📰 Hugging Face isn’t a tool per se, but is a repository of new AI tools from virtually everyone working in AI, and is where to learn about the newest and coolest AI tools and play with them in a sandbox environment or downloadable code to install locally to gain access to the new tools. Familiarizing yourself with HuggingFace is a great way to stay ahead of the AI learning curve, and fully in the know.