Podcast notes – ChatGPT goes prime time! – Practical AI (Daniel and Chris)

Practical AI 206: ChatGPT goes prime time! – Listen on Changelog.com

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.
—

Podcast notes – Simone Weil on the need for roots – Philosophize This!

Follow-up to previous notes: https://kevinhabits.com/podcast-notes-philosopher-simone-weil-on-attention-human-dignity-and-factory-work/

Moral sage
Died at 34

If keep life at safe distance, prevents you from seeing world in truly clear way

Marx – if his predictions were wrong, it may be because he never truly lived a worker’s life

After working in factories, feeling affliction firsthand, ideas grow more radical

People hide behind the abstract
While many out there are truly suffering and starving

We’re also starving in “needs of the soul”

14 of these needs
1. Order – need some order to function, just like food
2. Liberty
3. etc
Many of these are in tension with each other – and need balance
Each contributes to peoples’ psychological needs

Biggest need that is sabotaged by modern society – “the need for roots”, like a plant

Any human being that is truly independent?
No, inseparable from society

People get roots from active participation in life of community

Colonialism destroys this culture and solidarity
In her time, France subjugated foreign cultures, tried to replace with French culture
This transplanting is like uprooting a plant, destroying the roots

Codified human rights won’t matter to a person like Hitler

Forced prostitution is wrong and we all agree on it – but it’s far deeper than legal rights

Preferred discussing human NEEDS rather than rights

Separating people from their own culture, roots is de-humanizing

Revolution was the new opiate of the masses – but very little comes out of it

After her factory experience, she discovered a new way to think about ethics

What makes someone sacred?
Is it their body? Their unique personality?

The way forward is a spiritual revolution
Cultivation of a new kind of attention
Renounce the ego
Receive the world in a more universal, impersonal way
De-creating biases, assumptions, that we bring to bear
This is a spiritual transformation
Connect to human beings on a universal shared level, instead of at level of personality or physicality

Similar to Catholic sacrament of Eucharist

This new form of attention is a new lens to see the world

Become an antenna for God to communicate with
An experience to feel closer connection to the universe

“Waiting for God” – her most famous book

Podcast notes – Skeptical take on the AI revolution – Gary Marcus on Ezra Klein: “They’re just autocomplete…it’s mysticism to think otherwise”

Guest: Gary Marcus, NYU professor emeritus

Derived from things humans said, but doesn’t always know connections between the things – which can lead to it saying whacky things

Transforms everything into an “embedding space”

It’s pastiche – imitating styles, cutting and pasting, template aspect
Both brilliance and errors come from this

Humans have some internal model – of physical world, of a relationship
Understanding is also about indirect meanings

AI doesn’t have these internal representations

Sam Altman – “humans are energy flowing through a neural network”

Human neural networks have much more structure than AI neural nets

Neural nets like ChatGPT have 2 problems:
-not reliable
-not truthful
Making them bigger doesn’t solve those problems – because they don’t have models of “reliability” or “truthfulness”
“They’re just auto-complete”
“It’s mysticism to think otherwise”

No conception of truth – all fundamentally bullshitting
It’s a very serious problem

Can use GPT3 to make up stories about covid, spread bullshit and misinformation

No existing tech to protect us from this potential tidal wave of misinformation

People used ChatGPT to make up fake coding answers to put on StackOverflow
StackOverflow had to ban ChatGPT

“Talking about having submachine guns of misinformation”

Ezra – ChatGPT is really good at mimicking styles, but core doesn’t have truth value or embedded meaning

Silicon Valley always good at surveillance capitalism
Now you can use AI to write targeted propaganda all day long

Ezra – this AI will be good for people and companies who don’t care about truthfulness. eg, Google doesn’t care about “truth” but about clicks

One of main AI use cases is SEO copy to drive clicks and engagement

In some aspects, as models get bigger they get better (eg, at generating synonyms); but at truthfulness, wasn’t as much progress

Loves AI and have thought about it his whole life – just wants it to work better, be on better path

Biology has enormous complexity – too much for humans – AI could really help us
Also climate change
Could empower individuals like DALL-E does for artists
“Reason according to human values”
Right now we have mediocre AI – risk of being net negative – polarization of society, growth of misinformation

Parable of drunk looking for keys at night around a street light
The street light in AI is deep learning

Human mind does many things: pattern recognition, use analogies, plan things, interpret language
We’ve only made progress on a tiny part
We need other tools (not just deep learning)

Fight between neural nets and symbols (symbolic systems / symbolic learning)
Find ways to bridge these 2 traditions

Ezra:
Deep learning – give all this data, figure it out itself, but bad at abstraction
Symbolic systems – great at abstraction (eg, teach it multiplication from first principles), abstract procedures that are guaranteed to work

Most symbol manipulation is hard wired
Kids learn to count using generalized rules

AI hasn’t matched humans in capacity to acquire new rules – some paradigm shift needed
We’re worshipping the false god of “more data”
Believes will be genuine new innovation in next decade

Stopped ~50 minutes in

Blitzed: “We snort and we inject!”

Thanks to Sam Parr’s recommendation, I am reading Blitzed, a book about Germany’s widespread manufacture and usage of meth and other drugs after WW1, and how, despite Hitler’s reputation as a strict clean teetotaler, there is plenty of evidence he was addicted to cocaine and opiates.

Just had to share this song that the author references from 1920s Berlin:

—

Once not so very long ago
Sweet alcohol, that beast,
Brought warmth and sweetness to our lives,
But then the price increased.

And so cocaine and morphine
Berliners now select.
Let lightning flashes rage outside
We snort and we inject!

At dinner in the restaurant
The waiter brings the tin
Of coke for us to feast upon—
Forget whisky and gin!

Let drowsy morphine take its Subcutaneous effect
Upon our nervous system—
We snort and we inject!

These medications aren’t allowed,
Of course, they’re quite forbidden.
But even such illicit treats
Are very seldom hidden.

Euphoria awaits us
And though, as we suspect,
Our foes can’t wait to shoot us down,
We snort and we inject!

And if we snort ourselves to death
Or into the asylum,
Our days are going downhill fast—
How better to beguile ’em?

Europe’s a madhouse anyway,
No need for genuflecting;
The only way to Paradise
Is snorting and injecting!

Huberman Lab – notes on Chris Palmer (ketogenic diet) and Kyle Gillett (hormones)

Took notes on parts of both podcasts

Started notes 1 hour+ in

Mitophagy – subset of autophagy
Autophagy is stimulated by fasting states
When body doesn’t have enough food, it hunkers down and recycles dead old parts
Can stimulate through fasting, calorie restriction
Becoming lean and conservative

Mitochondrial dysfunction – associated with many ailments

David Sinclair – Cell journal – mitochondria are unified link in everything we know about aging

Mitophagy – how to get rid of old / defective mitochondria
Most powerful signal is diet – calorie restriction, fasting

Glucose levels in brain need to be reduced
Neurons love glucose – but why is too much glucose bad?

High glucose levels may be a symptom of metabolic dysfunction, possibly mitochondrial dysfunction

Infants – breast milk is most beneficial / healthy food source
If infants have seizures – ketogenic diet can help

Even experts don’t know what causes obesity epidemic
Believes mitochondria is the answer
Maybe epigenetic factors in womb environment

Mitochondrial DNA comes exclusively from moms DNA?
“Unequivocally no”
Mitochondria encoded in 36 genes – but majority of proteins that comprise mitochondria are encoded in nuclear DNA which is inherited from both parents

Diet, drug, alcohol use (including THC), lack of sleep – all impair mitochondrial function

Alcohol disrupts how brain uses all kinds of fuel – including glucose metabolism
If alcoholics switch to ketogenic diet – less alcohol craving, better brain metabolism, reduced brain inflammation

Rats on ketogenic diet – much more responsive to alcohol (“cheap dates”)
5x higher BAC levels!

Hypothalamus – involved in motivation and craving

Brain metabolism deficits can be corrected in short term with ketone supplements (eg, liquid ketone esters / ketone salts)

Ketogenic diet isn’t just about ketones themselves – also lowering glucose levels, ramping mitochondrial biogenesis

If metabolically healthy, ketones themselves *could* have benefit

—

Started 1:08 in

L-carnitine – oral only absorbs 10%
Increases androgen receptor density

Vitamin D – optimizes testosterone if deficient

Free testosterone should be ~2% of total testosterone – reference range is 1-4%

Longjack – up regulates several enzymes

Tongkat – helps synthesize testosterone; 300-1200mg/day; be careful with standardization
Andrew takes it early in day, has stimulant effect – increases free testosterone for him
No reason to cycle it
Can increase DHEA

Fadogia – increases LH (luteinizing hormone), stimulates production of sperm / testosterone
300mg / day as a safe dose or 600mg / every other day

1st trimester pregnant ladies have high HCG levels in urine

Growth hormone / IGF-1 – usually don’t need to optimize it, but sometimes in diabetics it can be dysregulated
Amino acids before bed can help with growth hormone release
Some weight training regimens can increase growth hormone

Helpful to avoid eating 2 hours before bed as general health protocol
While fasting increases growth hormone release, downstream effects are mitigated so not significant net effect