Must read if you’re interested in AI and its implications; Tyler’s commentary on the recent explosion of AI into the popular consciousness (driven in large part by ChatGPT) has been, in my view, the most realistic+pragmatic:
“I don’t have a prediction for the rate of improvement, but most analogies from the normal economy do not apply. Cars get better by some modest amount each year, as do most other things I buy or use. LLMs, in contrast, can make leaps.”
“I’ve started dividing the people I know into three camps: those who are not yet aware of LLMs; those who complain about their current LLMs; and those who have some inkling of the startling future before us. The intriguing thing about LLMs is that they do not follow smooth, continuous rules of development. Rather they are like a larva due to sprout into a butterfly.”
He doesn’t read the news
Likes trolling on Twitter…”Twitter’s fun”
“Twitter’s gonna be fine”
HIS INVESTMENTS
400 personal investments, a few thousand including YC
All the companies he’s added value to are those he thinks about in his free time – while hiking, texting the founder an idea
Most successful investment? Stripe on a multiples basis
Helion – fusion energy
Personally invested $375M (!)
Other thing besides OpenAI that he spends a lot of time on
New energy system that works on super low cost Hardest challenge is how to replace all (current) generative capacity on Earth really quickly
“Who can deliver energy the cheapest, and enough of it”
Simple machine, affordable cost, reasonable size
If fusion works…will change dynamics of what’s possible – enables more downstream (eg, more powerful planes)
Hermeus – supersonic jet company
Led $100M round
Was also involved with a competitor Boom – but different tech and approach
Huge market with multiple needs
Worldcoin
He’s a cofounder, on the board, but not day to day involved
Will tell its story soon – believes it will go over well (unlike earlier negative media coverage) We give up more privacy to Facebook than Worldcoin
Phenomenal team
Launch in months
Interested in any tech to experiment with global UBI (versus what one country can do)
Re: crypto — “honestly not super interested” “Love spirit of web3, but don’t intuitively feel why we need it”
Inception Fertility
In-vitro gametogenesis
In shadow of AI Next 5-7 years of biotech will be remarkable
Human life extension – “yeah maybe that’s gonna work”
Investing for 20 years, president of YC for 5-6 years
Garry (new YC president) will do a lot of things differently and be wildly successful
Last few years were really hard for YC YC can remake itself now – tourists are leaving now
—
ARTIFICIAL INTELLIGENCE
OpenAI has pulled together “most talent dense” AI team
“Gonna be tremendously good”
Why did ChatGPT and DALL-E so surprise people?
“Don’t know…reflected on it a lot”
If you make a good UX on top of something – believed users wanted to interact via dialogue
Pieces were there for awhile
Standard belief was AI would take over low skill / truck driving / generic white collar
Going exact opposite direction – taking over creativity where we thought humans might have special sauce
It’s not an intuitive finding
Released GPT-3 three years ago – thought ChatGPT would be incremental, was surprised by public reaction
ChatGPT will cause societal changes – eg, academic integrity “Stakes are still relatively low”
Covid did show us society can update to massive changes faster than he expected
Given expected economic impact – “more gradual is better”
GPT-4 will come out when we’re confident we can do it safely and responsibly
Will release tech much more slowly than people will like
GPT-4 rumor mill is a ridiculous thing
Re: ChatGPT – built a thing, couldn’t figure out how to monetize it, put it out via an API, and users figured out how to use it
Would like to see AI super democratized, have several AGIs in the world Cost of intelligence and energy trends down and down
“Massive surplus…benefits all of us”
Believes in capitalism – best service at lowest price
Society will need to agree on what AGI should never do
Broad absolute rules of the system
Within that, AI can do different things – safe for work one, edgier creative one – different values they enforce A user can write up a spec of what they want, and AI will act according to it – “should be your AI…to serve you”
Microsoft – only tech company he’d be excited to partner with this deeply
Satya, Kevin Scott, Mary McHale
Values-aligned company
“We’re very much here to build AGI”
“We wanna be useful to people”
Re: Google’s AI – hasn’t seen it, assume they’re a competent org
We’re in a new world – generated text is something we all need to adapt to, like we adapted to calculators
“I’d much rather have ChatGPT teach me…than read a textbook”
Anthropic – rival AI, stressing an ethical layer
Very talented team
Multiple AGIs in the world is better than one
Society decided free speech is not quite absolute – in similar ways AI / LLMs will need to have bounds too
Video is coming… no confident prediction about when Legitimate research project – could take awhile
AUDIENCE Q&A
When fusion online?
By 2028, could be plugging fusion generators into grid (pending regulators)
Re: AI worst and best case?
“best case is so unbelievably good that it’s hard to imagine, discovering new knowledge in a year instead of 70K years”
“Bad case is lights out for all of us” More worried about accidental mis-use in short term, less about the AI itself being evil
How far away is AGI? Much blurrier and gradual transition than people think
Re: state of San Francisco
Real shame we treat people like this
How elected leaders don’t fix the problem
Tech has some responsibility for it
But other cities do better than this Super long in-person work and Bay Area
Re: ChatGPT reaction
Expected one order magnitude less hype, users, of everything
Less hype is probably better The tech is impressive, but not robust Use it 100x, see the weaknesses
How Sam uses ChatGPT? Summarize super long documents, emails
For translation
Re: Google code red, threat to search
When people talk about new tech being end of a giant company, they’re usually long
Change is coming (for Google), but not as dramatically as people think
Before Google, memorizing facts was important – and now we’ll change again – and we’ll adapt faster than most people think
Prefers hybrid work, like YC – few days at home, few days in office
Skeptical that fully remote is thing everyone does
Most important companies will still be heavily in-person
Safety engineering for AI is different from standard safety engineering – consequences are greater, deserves more study
Raising capital now is hard, especially later stages – but other things easier – easier to rise above noise, hire, get customers What he’d do now — recommends for founders – “do AI for some vertical”
Advice for AI startups
Differentiate by building deep relationships with users, some moat like network effect
Plan for AI models continually improving
OpenAI is a platform, but also wants to do a killer app (platform + killer app) to show people what’s possible
Hosts: Daniel Whitenack (Data scientist), Chris Benson (Lockheed Martin)
All about ChatGPT
Chris – feels collaborative, like having a partner
Lots of structuring in the output – bulleted lists, paragraphs
Humans get things wrong / incomplete all the time – yet we’re holding AI to a higher standard
Shifting to more open access – maybe in response to open source AI products like Stable Diffusion
Chris – expect to see more “fast followers” to ChatGPT soon
TECHNICALS
GPT language models – it’s a “causal language model” not a “mass language model”
Trained to predict next word in sequence of words, but based on all previous words Auto-regressive – predicts next thing based on all previous things, and so forth
That’s why the text develops as if a human were typing
Few shot learning – can change answer style based on your questions and prompts
Zero shot = input that a model has never seen before
Few shot = provide a small number of inputs / prompts to guide the model
ChatGPT – trained with “reinforcement learning from human feedback” (RLHF)
Human preference is key part of this
How does it scale?
Human feedback is expensive 3 steps:
1. Pre-train a language model (aka a “policy”) – not based on human feedback
2. Gather human preference data to train a reward model – outputs prediction of human preference
3. Fine tune (1) based on (2)
eg for ChatGPT,
(1) is GPT3.5 (for ChatGPT)
(2) outputs data based on (1), add human labels of preference, and train a “reward model”
GPT3 is 100B+ parameters ChatGPT reward model is 6B parameters
For (2), goal is to reduce harm by adding human feedback into the loop
For (3), will penalize if it strays too far from (1), and score output according to (2)
Try only to make small iterative changes (adiabatic)
What’s next?
Open research questions – (2) architecture and process hasn’t been fully optimized, lot to explore there
Will be new language models coming (eg, GPT4, Microsoft, Google) – trying different (1) and (2)
Chris Albon tweet:
—
Sci-fi got it wrong.
We assumed AI would be super logical and humans would provide creativity.
But in reality it’s the opposite. Generative AI is good at getting an approximately correct output, but if you need precision and accuracy you need a human.
—
A little something we’re working on, you can try it here: http://galerie.ai/
It does two things
1. Returns results from a growing library of the best AI art
2. Lets you create AI art from multiple generative models (including Stable Diffusion 1.5, 2.0, 2.1, and MidJourney)
Let me know what you think!
We auto-suggest some of the trending prompts and popular art results on the home page, try it out!
For example here is ” photorealistic portrait, a young beautiful woman Goddess wearing Echo of Souls Skull Mask armor, skeletal armor bones made from 24k gold and silver metal intricate scroll-work engraving details on armor plating, skeletal armor, gemstones, opals, halo, aura, intricate details, symmetrical,”:
Not yet trillion dollar “take on Googles” startups yet – but will be a serious challenge to Google for first time
eg, Human level chatbot interface that actually works – new medical services, new education services
Idea of language interface where you say in natural language, dialogue, and computer just does it for you
Very powerful models will be one of genuine new tech platforms since mobile
How to create an enduring differentiated business
-small handful of large base models will win – skeptical of startups doing small models
–middle layer will become really important – take large models, tune it, create model for medicine, or model for AI friend – will have data flywheel
Lots of AI experts think these models won’t generate net new knowledge for humanity – thinks they’ll be wrong and surprised
AI in science:
1. Science dedicated products eg Alpha Fold – will see a lot more, bio cos will do amazing things
2. Tools that make us more productive – improve net output of scientists and engineers – eg, CoPilot
3. AI that can be an AI scientist to self improve – automate our own jobs, go off and test new science and research – teaching AI to do that
What is Alignment Problem?
A powerful system that has goals in conflict with ours
How do we build AGI that does things in best interests of humanity
How to avoid accidental or intentional mis-use
AI could eventually help us do alignment research itself Reid: will be able to tell agent “don’t be racist” and let it figure out
AI moonshots?
-language models will go much further than people think – so much algorithmic progress to come, even if we run out of compute or data
–true multi modal models – every modality, fluidly move between them
-continuous learning models
These above 3 things will be huge victory
OpenAI – focus on next thing where we have high confidence, let 10% of company go and explore
Can’t plan for greatness, but sometimes breakthroughs will happen
AI will seep in everywhere Marginal cost of intelligence and energy will rapidly trend towards zero – will touch almost everything
Metaverse will become like iPhone – a new container for software AI will be new technological revolution – more about how metaverse will fit into AI then vice-versa
Low cost + fast cycle times is how you compete as a startup
In bio – simulators are bad, AI could help
What are best utopian sci-fi universes so far
-Star Trek is pretty good
-The Last Question is incredible short story
-Reid: Ian Banks – Culture series
-tried to write his own sci fi story, was a lotta fun
Having a lot of kids is great – wants to do it
Won’t be doing prompt engineering in 5 years
Will be text / voice in natural language to get computer to do what you want
eg, Be my therapist and make my life better; Teach me something I want to know
Reid: great visual thinker can get more out of DALL-E — will be an evolving set of human talents going that extra mile
How to define AGI Equivalent of a median human that you can hire as a coworker – be a doctor, be a coder
Meta-skill of getting good at whatever you need
Super intelligence = smarter than all of humanity put together
Economic impacts will be huge in 20-30 years Society may not tolerate that change – what is the new social contract
How to fairly distribute wealth
How to ensure access to AI systems (“commodity of the realm”)
Not worried about human fulfillment – we’ll always solve it
But concepts of wealth and access and governance will all change
Running largest UBI experiment in world – 5 year project
Tools for creatives — will be the great application for AI in short-term
Mostly not replacing, but enhancing their jobs
How do these LLMs differentiate from each other?
The middle layer is what will differentiate – the startups fine-tuning the base models, about the data flywheel, could include prompt engineering
