Happy Monday and welcome to another edition of Artificial Insights, your weekly guide to analytic imagination.
This week I am excited to announce our brand-new website over at Envisioning.io. Over the years we have often struggled with finding the right tone and messaging to explain our vision and mission. After much iteration and fine-tuning, our new website finally explains the work we have done for clients, the services we deliver, and features an extensive overview of our signal collection around emerging technologies. Please share any feedback or bugs with us. 🙌
We have loads of improvements rolling out in the coming weeks and months, and hope to develop the platform we have always envisioned through this new website.
In other news: AI remains a big deal.
Few arguments can be made against the certainty of increased machine intelligence in our lives. Individually we can always push back – apply a luddite / minimal / Amish approach to our personal technology usage – but collectively the direction is fairly universal. AI will keep bleeding into every part of our institutions. If so many organizations were already struggling with digital transformation, just imagine the number of opportunities ahead when everything we’ve created starts acting intelligent. Few people are making the argument that LLMs are in any way conscious – but we don’t require consciousness to perceive intelligence. Being responsive is enough.
There are many ways of making sense of this transformation. After foundational knowledge (in ML, data science, statistics, etc.), you arguably need three things: keeping track of current developments, applying your skills, and having a map of where the field is heading. We are working on that last part, to be shared soon.
Until next week,
MZ
P.S. My very favorite feature of the new website is having reintroduced Futureshock - an ongoing collection of images of shocking reminders we live in the future. I started collecting these a decade ago and find them a powerful way of feeling the future. Please send additions my way on email or on our 💬 WhatsApp.
Hackers Guide to LLMs
This is an awesome talk about having a systematic approach to prompting and natural language programming:
Where you should be spending most of your time really is just changing and finding new approaches to solve things because that makes a huge, huge difference.
Your daily reminder that LLMs are very different from computers.
AI's Impact in the Gaming Industry
Gaming with generative tools seems like a big deal. Conversations, environments, development - all quickly changing with AI.
It's not just a chatbot experience but it actually fits within what you feel like a gamer should be—a gamified type of experience.
Transform your Organization with AI
Dmitry Shapiro, CEO of Mind Studio, offers spot-on insights about organizational adaptation to AI. Valuable observations, regardless of the tools used.
Our employees are still doing work, but the work they're doing now is sort of wrangling AI to try to do the work for them, and they could have done it themselves but now they're spending their time prompt engineering.
Demystifying AI: From Mythology to Practical Applications
This lecture by Leon Bottou of Meta on "Borges and AI" starts in mythology and goes deep.
This story of hallucination of language models—they're not hallucinations. It's not that the language model views things that do not exist; it's just that it's free to do that. It's totally allowed for the language model to invent a continuation as long as it's coherent; it follows the narrative necessity, truth doesn't matter.
If we keep confusing the fiction machine for an AI that knows so much and reasons so well that we don't need to learn and think, then there's going to be trouble. The trouble is not the machine; the trouble is if we give up learning and thinking ourselves, we're going to have problems.
-
🏅 Forwarding this issue to colleagues and friends.
-
🎯 Commenting with your favorite talks and thinkers.
You are receiving this newsletter because you signed up on envisioning.io or Substack.