Bookmarked Using GPT-3 to augment human intelligence: Learning through open-ended conversations with large language models by Henrik Olof Karlsson

Wow, this essay comes with a bunch of examples of using the GPT-3 language model in such fascinating ways. Have it stage a discussion between two famous innovators and duke it out over a fundamental question, run your ideas by an impersonation of Steve Jobs, use it to first explore a new domain to you (while being aware that GPT-3 will likely confabulate a bunch of nonsense). Just wow.
Some immediate points:

  • Karlsson talks about prompt engineering, to make the model spit out what you want more closely. Prompt design is an important feature in large scale listening, to tap into a rich interpreted stream of narrated experiences. I can do prompt design to get people to share their experiences, and it would be fascinating to try that experience out on GPT-3.
  • He mentions Matt Webbs 2020 post about prompting, quoting “it’s down to the human user to interview GPT-3“. This morning I’ve started reading Luhmann’s Communicating with Slip Boxes with a view to annotation. Luhmann talks about the need for his notes collection to be thematically open ended, and the factual status or not of information to be a result of the moment of communication. GPT-3 is trained with the internet, and it hallucinates. Now here we are communicating with it, interviewing it, to elicit new thoughts, ideas and perspectives, similar to what Luhmann evocatively describes as communication with his notes. That GPT-3 results can be totally bogus is much less relevant as it’s the interaction that leads to new notions within yourself, and you’re not after using GPT-3s output as fact or as a finished result.
  • Are all of us building notes collections, especially those mimicking Luhmann as if it was the originator of such systems of note taking, actually better off learning to prompt and interrogate GPT-3?
  • Karlsson writes about treating GPT-3 as an interface to the internet, which allows using GPT-3 as a research assistant. In a much more specific way than he describes this is what the tool Elicit I just mentioned here does based on GPT-3 too. You give Elicit your research question as a prompt and it will come up with relevant papers that may help answer it.

On first reading this is like opening a treasure trove, albeit a boobytrapped one. Need to go through this in much more detail and follow up on sources and associations.

Some people already do most of their learning by prompting GPT-3 to write custom-made essays about things they are trying to understand. I’ve talked to people who prompt GPT-3 to give them legal advice and diagnose their illnesses. I’ve talked to men who let their five-year-olds hang out with GPT-3, treating it as an eternally patient uncle, answering questions, while dad gets on with work.

Henrik Olof Karlsson

Something to aspire to

A few years ago Elmine and I wrote a short e-book on how to organize an unconference as a birthday party (PDF linked on the right). Since then I’ve regularly entertained the idea of writing another e-book, but that never really happened. While I do have some topics I’d like to write about, I find my knowledge of those topics still too limited to be able to come up with a narrative to share anything worthwile. There are also doubts (fears?) about what type of things would have a potential readership

So this week I decided to ask:

What would you like to see me write more or more extensively about?

Already I got a range of responses, and it is an intriguing list. Some suggestions are about aspects of my own journey, others are about topics that I don’t know much (or anything) about, but where apparantly there’s interest in my take on it. Some come close to topics I already want to write more about, but feel I haven’t found an angle yet.

Here’s the list until now. More suggestions and thoughts are welcome.

  • Optimal unfamiliarity (a phrase I coined in 2004 initially to describe what mix of people make a great event audience to be part of, but has become a design principle in how I try to collect information and learn.), suggested by Piers Young
  • An epistolary travel log novella (something that could arise from my 14 years of blogging about my travels and work), suggested by Georges Labreche
  • Open currencies (which Google tells me they have no meaningful results for, but which connects to my experience with LETS, and chimes with free currencies in p2p networks), suggested by Pedro Custodio
  • Moderating sessions with a mix of analog and digital tools (closely connected to my thoughts about fruitful information strategies in social contexts), suggested by Oliver Gassner
  • Fatherhood (as I became one 9 weeks ago, but I don’t think 9 weeks counts as experience), suggested by Dries Krens
  • Motivating others to act on open data (a large chunk of my work), suggested by Gerrit Eicker
  • Being a European in the digital age (which I strongly claim to be), suggest by Alipasha Foroughi
  • Convincing profit oriented organisations of the value of open access and responsible research (comes close to Gerrit’s point), suggested by Johnny Søraker
  • How and why I left my job (being employed by Dries mentioned above), suggested by Rob Paterson
  • The journey from my involvement in knowledge management and early blogging, to where I am now, and how it impacted the way Elmine and I arrange our lives (lots to unpack here!), suggested by Jon Husband (who, like Rob Paterson, has been part and witness of that journey over many years)
  • The proliferation of means of communication versus the quality of communication (for me this points to information strategies on focus, filtering etc.), suggested by Jos Eikhout
  • Personal information strategies and processes using open source tools (something I blogged often about in various shapes and forms), suggested by Terry Frazier, a fellow blogger on knowledge management back when I started blogging in 2002

Looking at who responded is already in a way a manifestation of some of the suggested topics (the journey, the information strategies, the optimal unfamiliarity, facilitating communities).

I can’t promise I’ll write about all of the things suggested, but I appreciate the breadth and scope of this list and the feedback I can unpack from it. More suggestions are very welcome.

That is the topic I am currently working on with Elmine Wijnia and Valeri Souchkov (an expert in the TRIZ methodology).
Elmine and I had been discussing if it would be possible to predict viable social software tools and niches based on the affordances people actually need to be able to consciously learn things. This as a result of our paper for the BlogTalk Reloaded conference where we suggested using community critical success factors, and the work of George Siemens (Connectivism) as design principles for tools.

Illustration from George Siemens book Knowing Knowledge

Opening up this discussion with Valeri Souchkov helped focus on the singlemost important question that surfaced in our exchanges: how do you get to own your own learning path?

This is a bigger question than we started out with, but it is the ultimate conclusion of trying to approach things not in a tech-driven, or functionality focussed way, but by focussing on an individual and on being empowered to reach your own goals.

Starting from the question How to be owner of my own learning path?, we distinguish a couple of prerequisites:
1) Knowing what to learn
2) Knowing when to learn it
3) Knowing how to learn it
4) Knowing when you’ve learned it
5) Being in an environment that allows you to own your own learning path

Building on this quickly branches of in all kinds of directions.
Those branches allow a more detailed look at things like:

Monitoring, evaluating, shaping and balancing your environment as fit for learning. Which connects to my earlier work on social networks as filters, as well as the second BlogWalk on self-directed learning.
Detecting the need, and right time to learn things against the background of continual changes in the world, and in light of your goals.

Shaping your learning steps with interventions that fit your goals, environment and context, and being able to establish when you’ve learned something and can move on to the next learning goal.

Each of these ‘fields’ need to be supported with their own group of skills and tool box.
And this is where I start to see the contours of what existing tools should look like in different contexts of usage, and what type of tools are missing. When it comes to environment think of visualization e.g., and when it comes to monitoring change, look at pattern hunting.
Because it helps integrate and connect different pieces of your efforts and actions (e.g. it already helps Elmine see how her consulting and research work fits together in a new light.), it helps you think about the type of affordances you need and from which type of tools to get them.

That is an entirely different approach than the one that went ’round the blogosphere in the past week on the building blocks of social software. That was more about describing tools, but not starting with the intended usage of those tools.

This is all still a bit vague, and in part deliberately so.
We hope to take this to Reboot at the end of next month, as a conversation with an introductory presentation. So we will keep building on it.
I hope that it will give us a means to proactively define the tools I need, to determine better what type of feedback to give toolbuilders (like I’ve been doing at different BarCamps) and perhaps spot a niche or two for start-ups.

Sticky notes from exploring one branch of our model