In Conversation with OpenAI (ChatGPT) CEO, Sam Altman - Melbourne, Australia

Welcome to my recap of A Conversation with OpenAI CEO, Sam Altman, hosted by The Startup Network (formerly Startup Victoria), with Dr Nora Koslowski of 澳大利亚墨尔本大学 - 墨尔本商学院 and Vicki Stirling sharing the stage.

You know Sam and OpenAi because you know ChatGPT, which acquired its first million users in 5 days, and got to 100 million in 3 months.

Depending on your experience with technology, startups, business and investing you may also know Loopt which Sam co-founded or Y Combinator , an early-stage startup investment firm, in which he was a partner, becoming President in 2014, and announced they had over $65BN in valuations across the portfolio.

Y Combinator has launched over 4,000 companies including Airbnb , Coinbase , DoorDash , Dropbox , Quora , Stripe and Twitch .?He was briefly the CEO of Reddit, Inc. for about 8 days, which also came out of Y Combinator.

Melbourne was Sam's last stop having visited 22 countries in 4 weeks, and straight off the plane from speaking to our heads of state in Canberra, a level of discussion it was clear from his time on stage was the norm, as he asked world leaders and anybody that had the chance to front a microphone, what they wanted help with, what they were worried about and how they wanted to use Artificial Intelligence (AI).

I hadn't looked into Sam before attending this event, and when being asked if I was excited I honestly didn't know.. what was about to happen???

Was I going to witness a new tech messiah march us towards a promised land with a touch of scorched earth policy behind him??Was the spectre of Artificial General Intelligence (AGI) being unleashed on the world???

Was a blend of Black Mirror, Cambridge Analytica, everyday Facebook / META, The Social Dilemma and Johann Hari's Stolen Focus about to crash land into society, displacing all white-collar jobs through chatty robots that were fusions of Tesla Bots plugged into ChatGPT personas?

Turns out, I felt more like I was standing in Kitty Hawk, North Carolina, 120 years ago, hearing Orville and Wilbur Wright excitedly discuss inventing flight.

Sam was eloquent, deeply confident, conscientious, affable, optimistic, and considered and didn't miss a beat scaling between the absurdly mundane (what classic car is this photo competition) and will extremists have access to open source LLMs (Large Language Models) they can chat to and get coached on biological warfare with???

"It's a ute" and "Yes", if you were wondering.

Extremes aside, there is an enormous transition going on around us, and an accelerating pace of change that isn't slowing down - Dr Nora Koslowski of the Melbourne Business School held the stage, asking a series of inspired questions that I've done my best to note and summarise for you below;

Nora asked Sam to speak to Australia's slow adoption rates and being noted as incremental improvers, and how we should tackle creating innovation ecosystems as a country;

Sam ".. I think one of the exciting things about this AI revolution that has just started is it is a full reset for everyone.?Everybody gets to start over, we're all at the starting gun and everything is going to be very different now.

People, countries, areas, whoever and whatever missed the last innovation boom, everybody is in the same place.?It's a reset for all of us and that's an exciting time.?

I think this is probably the best time to start a startup ever, it's at least the best time to start a startup since the internet.

This is going to be bigger than mobile and on the order of the internet, maybe bigger than that, we'll see.?

For new companies or existing companies that really want to innovate, all the rules just changed. And this is when people can make moves in innovation.?

Don't worry about the past, look forward, and embrace this.?The innovation potential in front of us globally is quite exciting. I knew it was happening super globally, but the visceral feel of that on this trip has been very cool.."?

On Australia making a contribution to the AI revolution;

".. It's all going to come down to who embraces the new technology the fastest and the deepest.?The whole world wants this technology, the whole world needs the benefits of this.?

Asking 'what ways are Australians different from the rest of the world?' is somewhat the wrong question.

Ask 'What can we go do for the whole world?'?

Before OpenAI I ran Y Combinator, one of the things I would notice is startups around the world that were not in the US would say 'We're gonna do this regional version, then we'll do it for the world'.?And I would always say 'Why don't you just do the thing for the whole world?'

And I think that's really important. Many startups have now gone on to do that. But rather than ask "What can we do unique to Australia?" ask "What can we go do better than anybody else in the world?"

Nora - "So you're encouraging us to think bigger?"

Sam - "Sure, why not? That's always more fun"

On tackling the productivity challenge faced by Business Leaders in Australia (it's not going up), and more so the Western world at large;

".. AI is coming just in time. This is not just an Australian phenomenon. This is happening globally, it's happening really broadly - in the US we've had over a decade of totally lost productivity gains.?In some ways, it doesn't make sense, though if you look at all the pieces, it does.??

And the productivity growth that we are already getting from AI is astronomical.?An area we're most familiar with is computer programmers because we have been working with that area much earlier than other technologies.?

Computer programmers are telling us they are two to three times more productive, sometimes they are going up to 20 or 30 times more productive.

When you have a 20x or 30x productivity gain, it becomes a qualitative change in what someone can do, when someone has not just the ability to programme 20 or 30 times more productively, but do all the other aspects of their job, it's remarkable.?For people running their own medium-sized companies, these gains can really help with a lot of things.?

So I think one of the most exciting parts of this technology, and part of why people want it so much is the productivity growth that we can all already see happening.."

Talking to the leaders in the room wondering about how to transition their workforce, and encourage learning, reskilling or upskilling at scale;

".. the thing that works best is companies saying "we're going to have everybody start using ChatGPT, everybody start building on the API.?We don't what's going to make sense, though we're going to have a very high rate of experimentation.

As the models get better and the capabilities go up, we're going to keep trying things that didn't work before."

For the handful of companies that have worked closely with us and gone all in saying 'we're going to (try) using AI for everything' - they report incredible results.

It may take a few weeks to figure out exactly what workflows are going to change and how you're going to think about which people are allowed to use it or not.?What we're seeing is leaning in all the way and trying this for everything, then seeing where it works.?That seems to be the best approach."

Nora - ".. And in this version of future events, if we're all more productive, wealthier, ideally healthier, more educated, to me a potential downside is you're removing friction from life. And we need friction. We need challenge, we need adversity as humans in order to create meaning..??

Sam - ".. We'll find it! We're not going to run out of problems.?Every problem we solve is a greater and greater triumph that leads to greater, more interesting problems.?

I've never believed we're going to build AGI and then all sit around and play video games - some of us will, and that's totally fine - if that's what you want to do, I think you should be allowed to do that and society should make that happen for you.?Though most people want to experience adversity, or at least experience challenge.?

And I think the problems won't go away, as we have better tools, our capabilities will increase, but the expectations will increase as much or maybe even more.?

Nora - ".. If and when superintelligence becomes a thing, how will it change our definition of humanity as a species?"

Sam - ".. I think it's not as simple as saying, we have this system now that can compute really well and therefore humans have no role.

We'll have this tool that'll do a lot of the things that we do now or did in the past and we will go on to do and care about new things. And there are versions of this that have been happening for a long time.?

If you think about the previous technological revolutions, we used to spend a lot of our time hunting and gathering and probably defined ourselves by our ability to outsmart, certainly to outrun over long distances, animals.

And then we didn't have to do that as much and had new things that we defined our time with.?And then the industrial revolution came along and that changed things quite dramatically.?Then came the computer revolution which changed things dramatically again.

I somehow feel that this will be bigger than others, but that's probably all of us just feeling that our moment is particularly special, or maybe it really will be different, history thousands of years from now will be the judge, though also, maybe humans are not all that special.."?

In closing, regarding his tour of 22 countries in 4 weeks and what he hasn't been asked;

Sam - ".. It's been a real pleasure to get to talk to so many people.?The one thing I'm always happy to talk about more, that naturally does not get asked as much as the downside is, what does the real upside scenario look like??

What does it really mean if we have intelligence that is too cheap to metre, if we can all have everything we want??Not only that, if we can have things that we can't even imagine today.?It's not even that we solve climate change and every disease is cured.?

That's all great, but that's boring and obviously certainly going to happen.?Though what's the really great stuff??30 years, 50 years, 100 years into the future??

Dr Koslowski handed back to Vicki Stirling for the audience Q&A, where Sam was asked about his market picks for emerging opportunities;

Sam ".. The other thing that has astonished me about this tour is the creative energy of the world when you give them a new tool.?And that's what's been so fun about this trip, simply the breadth of what people are doing.?

And I don't know what's going to work.?I have my guesses and you have your guesses, all of us voting together as a market will decide.?But the energy is there.?That's exciting.

Another thing going on is there were people that a year and a half ago were all in on crypto that have all pivoted to AI, and two years from now will have pivoted to something else.?And for these tourists, I'm less excited."

On the opportunity in education;

Sam ".. I also think it's going to be a revolution and transformation for education.?I've seen in the US there are big differences between classroom education and one-on-one tutoring, it's two standard deviations.

But most people just can't afford one-on-one tutoring.?And also, we have better one-on-one tutoring in a hybrid way than we can with any one human right now.

So if we can combine one-on-one tutoring for every child with the things that only a human teacher can provide, the support and the excitement, that combination is going to be incredible for education.?

For neurodivergent kids too, it's going to have a big impact.?I'm really excited by the opportunity that AI can bring into the classroom to help level the playing field.."

Q.?Given your experience with Y Combinator, what advice would you give to Australian AI startups seeking global recognition??

Sam - ".. Make something people want.?Of all of the advice that YC has to give, that is the one inviolable tenant I believe in.?Paul Graham started that comment, though it is the only piece of startup advice that I think there are no exceptions.?

Whatever you're thinking, do it a little bit bigger, make something that people want a little bit more and don't worry about too much else.?Any way you solve a problem, if you solve it sufficiently well, you will have all of the global recognition you want.?

And don't do something derivative, though everything is somewhat derivative.?Do something that is very ambitious.?

Even if it takes many years of being misunderstood, it took OpenAI four and a half years to launch our first product, which was quite bad.?

And it took us six and a half to get to something good.??

I think that's totally fine and there should be more of that as long as you are making rapid progress internally.

Q. How did you achieve so much with such a relatively small team compared to much larger tech companies, and what advice would you have for small Australian tech teams?

Sam - ".. I would flip the question and say, how do these companies with very large teams accomplish anything at all??

It's easier to succeed with a small and focused and super talented dense team.?

I'm a huge believer in small focused, super talent-dense companies that have a clear mission that everybody believes in.?

And at a company that has hundreds of thousands of people, that's just hard. They can copy super well, though innovation is easier with a relatively small team that has to make decisive and clear concentrated bets and that doesn't tolerate any mediocre performance. That's it."

There was terrific detail in the rest of the Q&A, including;

- the challenge of whitewashing US cultural norms across the world based on the parameters of ChatGPT and providing localised versions of it

- equity of access to the software, eg 'Hello' in English being 1 token and 'Konichiwa' in Japanese being 7 (brought up in the opening anecdotes)

- a representative of Beyond Blue asking about the challenges of people seeking support and counselling using ChatGPT as a therapist

The last of my notes has a big circle around it saying "ChatGPT is good at tasks. Not good at jobs" in relation to whether we're all going to be replaced, and that creativity and human interactions are the necessary skills of the future.

Sam's summation was "The only thing I'm certain about is the rate of change is going up - figure out how to operate 10x faster".

He also noted it takes 5 years to figure out how technology impacts the world.?

I'll check back with you in 2028 and we can see how we're going hey?

How are you leaning in on ChatGPT in your work??Let me know in the comments, or share what stood out to you from this recap.

Thanks for reading, liking and sharing Smarter Impact, the place for Remarkably Better Conversations!