Mohamad’s Newsletter

Share this post

8: Should we be worried about AGI?; Are social platforms programming us?; and more..

mohamadahmad.substack.com

8: Should we be worried about AGI?; Are social platforms programming us?; and more..

Mo
Sep 13, 2020
1
Share this post

8: Should we be worried about AGI?; Are social platforms programming us?; and more..

mohamadahmad.substack.com

Hi all!

Today we’ll talk about two recent documentaries, and share other cool stuff i’ve come across last week:

We need to talk about A.I

Narrated by 2001: A space odyssey's Keir Dullea, we need to talk about AI is an invitation to join the global conversation about AI so we can ensure a safe future for humanity. Top AI researchers ,AI builders, and even James Cameron weigh in and share their views about the topic. They discuss the goal alignment; when we build an AI it is essential that we imbue it with our values and what's okay to do in order to achieve a specific goal: Say you program an AI system with the goal of achieving peace in the world. For all we know, it can go about killing all humans, for, according to its predictions and calculations, that is the only way this goal could be achieved. This for instance is an example of goals unaligned between humans and machines.

Currently, we have narrow AIs, capable of achieving a specific goal in one area, through computational prowess, and processing, but outside of this area they're good for nothing. The interviewees in the doc talk about the future of AGI, a future AI that is competent in all areas, with infinite degrees of freedom, emulating reality; an AI as intelligent as a human, and just like a human, adapts to any context/environment. The researchers argue that it shouldn't necessarily be the case that a brain can not to be emulated in silicon, cause at the fundamental level,all matter is made of common elemental particles, and there is nothing special about carbon based life forms. The laws of physics don’t prevent the advent of an AGI. Narrow AIs have been gaining new degrees of freedom every few years. IBM’s deep blue beat Kasparov in 1997, than one degree of freedom later, Watson beat Jeopardy champions in 2011, Alpha go beat Lea Sedoul in 2016, then Alpha zero came out… and now we have Gpt3. The computation power of computers is doubling every 5 years and the cost is going down. At this rate, one researcher argues, we'd be able to emulate the brain in 50 years. The thing with AGI is that once it happens,we’re past the time horizon; we reach the singularity, a point where the AI, by itself, can get incredibly smarter over night, what we call an intelligence explosion occurs, the result of which no one could predict. Fears about AI taking over the world have been going on since forever, but this time it is different because of the internet. It is not in the offing that the development of AI will take a break, cause there are huge investments being poured in and a lot of smarter people joining the field. You can not bet against human ingenuity and say that AGI is impossible,says one researcher. Another interviewee who’s been building AI for 25 years feels offended by the claims of Elon musk and others about AI being an existential threat, and argues that every time we think we're close to a breakthrough we discover that there a lot we still don’t understand. He says that we should not look at AI as good vs bad. AI could produce a lot of good in the world. One researcher raised the concerns about killer robots AI used as weapons to do the dirty work of killing. Russia in one conference says that whoever wins the AI race will rule the world, China is of the same opinion, which means that no nation is willing to put the brakes on to think about safety and getting things right. It is a race to the top. The risk is that AI builders are building things fast,with the sole purpose of succeeding, heedless of the safety issues that need to be addressed and thought through. The transhumanists won’t leave it for chance and propose that we merge with AI as a long term solution. You don't wanna be devalued or looked at as ants by the next generation of machines. Could we be the species whose purpose is to build these far smarter species that will eventually rule the planet and occupy the whole universe? Could be. When asked when AGI might happen, the answers varied from a decade to hundreds of years. Even if we have a 100 years to think about how we can make AGI a boon to humanity, we need to address the safety concerns and the goal alignment on a global scale starting now. Who is to say that there are no highly secretive schemes out there to build AI for use in vandalism ? People were saying that figuring out nuclear energy is impossible and then suddenly we had nuclear weapons. Max Tegmark from the doc compares Apollo 11 to AGI; it is not fear mongering thinking about the future of AI, but safety engineering. We have to think about everything that could go wrong so it goes right.

Are social media platforms programming us?

You turn on the news and it is Mad Max fury road out there; the world has gotten crazy, and we, oblivious to what the cause is,are just ineffectual nodes on a large unfathomable network. The new Netflix doc, The social dilemma, gathers together former engineers and designers at Twitter, Facebook ,Youtube, Pinterest, and Instagram in a confrontational session about what’s got us here. What is the root of the problem? Spearheading the rebellious movement is Tristan Harris, The guy who started it all. Tristan was a google engineer(design ethicist), but he got sick of where things were heading, and felt that he should free himself of blame and tell the whole world what is happening on the other side of their devices. I have been following Tristan’s work for two years. It started when I came across this talk of his, then I listened to several podcasts of his on the Tim Ferries show and the making sense podcast. Really changed how i look at social media and the connected world, but now, with this doc, he’s not alone, he broughts some insiders who, for the first time, are lifting the curtain on the social media conglomerates. Tristan argues that our brains are equipped with the tools necessary to survive in small tribes, in hunter gatherer times. Never in the history of humanity have 30 tech designers been able to control the lives of 2.7 billion people. The tech designers confess that the ultimate goals of the company are to increase profit, engagement, and growth, never caring about what’s good for humans, and all of these goals are reflected in the design of their products, nothing is hapless. Only few people in the tech world know how these systems really work. Once released into the world, the system takes a life of its own. These systems designed to “connect” the world aim at gradually changing people’s behavior. What makes them successful is their abiliy to make perfect predictions, and sell certainty to advertisers. Every user of Facebook gets a curated version of reality, targeting his psychological nature. You have to be aware that whatever you do on the internet is recorded, watched, and used to make your avatar version of yourself more manipulative to these companies. These tech giants make a huge amount of money because they're trading the future of humanity for money; the avatars of users on their platforms are manipulated by algorithms that feed on their personalities. It felt unsettling seeing the impact of using these social platforms on teenagers’ mental health, their dwindling willingness to take risks, and the increase in suicide rates among them. The wreck these platforms wrought goes further: they create more polarization and echo chambers and produce people that don’t wanna to have any conversation with each other; opposing parties filled with hate and animosity toward the other. You know, when the genie is out of the bottle, it is hard to put it back in; all the tech people can do now is tweak their systems to salvage the good left in them. The solution to this dilemma will happen when humans induce, collectively, a massive public pressure to coax governments into regulating these platforms in a way that protect people’s rights. Unsurprisingly, None of the interviewees let their children use these platforms, because, being the designers, they're aware of the dangers of these devices. The design of Facebook and Instagram is inspired by Las Vegas’s slot machines. Unexpected rushes of dopamine manipulate the brain’s circuitry. The tagging, the likes buttons, and the comment sections are all examples of the tricks used to boost engagement. It is not as if social media is a tool that is sitting at rest waiting to be used, like a bicycle, say. It is a tool that is controlling the masses’ behavior and affecting even people who don’t use it through the culture it creates. The continuing use of these platforms heralds an existential risk. AI will get better and these tech are following the same strategies, treating people as nodes on large network of collective brain, with one goal: increasing profit. Left as is, these platforms might be our undoing. We need to voice our concerns and recognize that better AI won’t solve the problem; AI predictive models can’t distinguish between what the truth is and what fake news is. Fake news naturally attract more attention owing to how the brain is wired. The truth is boring to the brain. It is the era of attention economy. What happens when these platforms act as mediums to distort the truths and spread conspiracy theories ? Better recognize the danger of these devices before it is too late.

What I’ve been reading

  1. This week, I finished reading the book Future Perfect: the case for progress in a networked age . Check out my review here.

What I’ve been watching

  1. World's Fastest Pitch - Supersonic Baseball Cannon: Fascinating video about what happens when you fire a baseball that travels with a supersonic speed into a wall.

  2. Is dust mostly dead skin: We see dust everywhere but how much of it is actually shed human skins, and what other constituents dust is made of?

  3. Joe Rogan- Naval: Finally got around to watching Naval’s episode on the Joe Rogan’s podcast. Here are the main takeaways:

  • 1- Naval: I think young geeks are attracted to my ideas because they never saw a bear on a tricycle. Message, combine skills in new ways

  • To have a clear mind avoid politics

  • Reading the best 100 books again n absorbing them well is better than reading everything

  • The trifecta are rich, happy, Calm. Rich can be taught. Don’t have many desires. Have a limited number of desires and let go of the others to increase happiness.

  • AGI is overblown. You cannot simplify the brain by what a neuron does. Something is going on inside the cell of the neuron that you're not accounting for. All the jobs that are being replaced are countered with new opportunities and ways for creativity. Naval is against Universal basic income because it could bankrupt the government by people asking for more.

  • Everyone could get rich

  • Be authentic to escape the competition

  • In a world inundated with knowledge go back to the basics. Wanna understand economics, read the wealth of nation. Physics? Read Feynman’s six easy pieces…

  • Do something you love which seems like work to other people. No one will beat u at that.

Films:

  1. I’m thinking of ending things: Most of the movie is a couple driving,snow outside, with only brief stops; the most crucial of which is a visit to Jake( the driver)’s parents house, where things get confusing and time is distorted. We don’t exactly know whether everything Jake is experiencing is a mirage or true. I liked the movie because its soothing ambience, and the deep conversations had throughout. Also, that scene at the end where everyone is aged for no obvious reasons is interesting. I think it is one of those movies that require multiple vieweings to appreciate.

  2. Cat people: Cat people are people who turn into cats when triggered by certain human behaviours. Watchable.

Series:

  1. Undone season 1:

Undone is a series done in the animated style of waking life., and Scanner Darkly(rotoscoping?). The series centers around a a girl who‘s on a mission to undone her father’s accident and alter the timeline. One time the girl is driving her car when, like a bolt from the blue, her dead father appears on the side of the road, she looks at him in a moment of shock, and just as she is looking at him, a speeding car crashes into hers at a road intersection. From here, the series takes on an unexpected turns and reality blends into dreams, and no longer is it clear what is what. Could all this we’re seeing be delusions the mind of the girl created like in Shutter island? Turns out the father had distracted his daughter deliberately so that she arrives at that astral plane sort of, where he could communicate with her and make her do his bidding in finding out who kills him by going back in time and prevent the murder. The dad gives the girl two choices: either continue this boring life linearly, which will follow a normal course of events: aging, kids, and death. Or, live in the world of infinite possibilities where you could alternate between timelines and shift state of consciousness at the drop of a hat. She accepts the offer and the madness begins: one time she is in a class tutoring children, suddenly she is in the backseat of a car with her dad in the driver’s seat; another time, she finds herself floating in space among the stars. All the while, she was receiving helpful inputs and instructions from her father in order to be able to train herself to go back in time to whatever time she wants; her training included making the key on a table disappears by accelerating the passing of days and go to the time when the key isn’t there. When she first succeeded in going back in time to the time the father had left her upon receiving a call from work. Watching from a distance, father besides her, that her mother had stormed in her father’s lab, and was rummaging through the drawers. The father had a girl assistant so the mother thought that a kinda of affair was going on. At last, after the girl discovers that her dad was the perpetrator of the crime and that he deliberately swerved into the his demise, girl sat besides him in the car, the girl tells him that he’s not the same person from the past and that he should go back with her and prevents the accident. So he does this by going back to that time he left his daughter, and he merges present( now wise) self with his past self, which could now make the good decision of ignoring the call from work and spend time with his daughter. At the end of season 1 the girl was sitting opposite a cave with her sister who was trying, as her boyfriend and mother before her did to convince the girl that all of what she’s experiencing is the working of her mind. The final scene showed her face widening as the light coming out of the cave opposite engulfed her face. Could be that the father is back and what she is experiencing is real. the creativity and inventiveness is top-notch in this show.

Music:

  1. Moderat- therapy

  2. Milk

  3. Damage done

Twitter:

Twitter avatar for @waitbutwhy
Tim Urban @waitbutwhy
If you ever have a business idea and get discouraged thinking the market is saturated and it’s all been done, remember that Google was the 21st search engine to enter the market in 1998. A+ products make waves in almost any market, no matter how saturated. https://t.co/CRmiycRn3X
Twitter avatar for @ValaAfshar
Vala Afshar @ValaAfshar
22 years ago, Larry Page and Sergey Brin filed for incorporation of @Google (September 4, 1998). Google was the 21st search engine to enter the market in 1998. A look inside the company’s garage office in 1998, with intros to the co-founders https://t.co/5AWxbmv0Df
2:58 PM ∙ Sep 7, 2020
3,912Likes804Retweets
Twitter avatar for @waitbutwhy
Tim Urban @waitbutwhy
.@isaboemeke continues her crusade to teach people about the case for nuclear energy without them realizing what’s happening
Twitter avatar for @isaboemeke
isabelle 🪐 @isaboemeke
in which I talk about the density of uranium https://t.co/42downmuMx
2:07 PM ∙ Sep 7, 2020
1,465Likes171Retweets
Twitter avatar for @michael_nielsen
Michael Nielsen @michael_nielsen
The installed capacity for nuclear power generation. It basically stops circa the mid-1980s, at about 400 reactors worldwide. Changing this will require extraordinary marketing / messaging, and (likely) changed technology / safety. I suspect the former is the biggest bottleneck
Image
12:53 AM ∙ Sep 8, 2020
76Likes9Retweets
Twitter avatar for @veitchtweets
James Veitch @veitchtweets
Two thousand years of global temperatures in twenty seconds
Image
12:50 PM ∙ Aug 19, 2020
18,857Likes11,148Retweets
Twitter avatar for @TrungTPhan
𝚃𝚛𝚞𝚗𝚐 𝙿𝚑𝚊𝚗 @TrungTPhan
When @elonmusk sold Paypal (and cashed $165m), one of the first things he did was cold call a rocket expert (Jim Cantrell). Below is the text of the greatest cold outreach ever.
Image
2:40 PM ∙ Sep 2, 2020
2,593Likes541Retweets
Twitter avatar for @rivatez
Riva @rivatez
In a world that’s trying to keep you melancholic and docile, the ultimate rebellion of our times is to ignore it all and dare to be happy
3:30 PM ∙ Sep 2, 2020
3,271Likes597Retweets
Twitter avatar for @Schuldensuehner
Holger Zschaepitz @Schuldensuehner
The Tech Bubble could get even bigger, Barron’s says in cover story: Recent correction doesn't mean that the bubble is ready to burst. The forces that have inflated Apple, Amazon, Zoom, and other tech stocks could be w/us for a while. That suggests this market could keep climbing
Image
8:56 AM ∙ Sep 12, 2020
301Likes104Retweets
Twitter avatar for @TalebWisdom
Nassim Nicholas Taleb's Wisdom @TalebWisdom
"Daily news and sugar confuse our system in the same manner." - Nassim Nicholas Taleb
5:18 PM ∙ Sep 12, 2020
267Likes61Retweets
Twitter avatar for @george__mack
George Mack @george__mack
Rule of thumb: 1. The more uncomfortable the activity, the more likely it will lead to growth. 2. The more comfortable the activity, the more likely it will lead to stagnation. 1000 uncomfortable hours > 10,000 comfortable hours
5:00 PM ∙ Sep 12, 2020
231Likes30Retweets
Twitter avatar for @balajis
balajis.com @balajis
We wanted flying cars, but we’ll start with folding chairs.
Image
5:55 AM ∙ Sep 12, 2020
2,621Likes446Retweets

Twitter avatar for @moms4nuclear
Mothers for Nuclear @moms4nuclear
Hilariously accurate
Image
5:33 PM ∙ Sep 11, 2020
25,106Likes7,381Retweets
Twitter avatar for @jaltma
Jack Altman @jaltma
Best career advice I ever got: only play games that you really care to win.
7:29 PM ∙ Sep 4, 2020
4,796Likes633Retweets
Twitter avatar for @naval
Naval @naval
Money doesn’t buy happiness - it buys freedom.
9:22 AM ∙ Sep 11, 2020
27,107Likes5,321Retweets
Twitter avatar for @balajis
balajis.com @balajis
It's not Silicon Valley anymore. It's Technology. Tech is now remote-first. Moving to the Bay Area is no longer necessary. And outages, fires, prices, and dysfunction are driving out the people already there. The place is less appropriate as a metonym for the concept.
6:52 PM ∙ Sep 10, 2020
1,141Likes194Retweets
Twitter avatar for @longnow
The Long Now Foundation @longnow
Graduates from National University of Singapore future-proof their city dezeen.com/2020/07/07/nat…
Image
Image
Image
Image
8:33 PM ∙ Sep 9, 2020
15Likes1Retweet
Twitter avatar for @andy_matuschak
Andy Matuschak @andy_matuschak
Fun inversion from Thorndike (1921). The normal angle is: "Why are some people so much better at some things? What are the limits of expertise?" He reframes to: "Why do most people remain so mediocre at things they spend their whole lives doing?" andymatuschak.org/files/papers/T… (p. 178)
Image
12:03 AM ∙ May 4, 2020
220Likes18Retweets
Twitter avatar for @mcgillmd921
Michael McGill @mcgillmd921
Naval Ravikant advice on what to pay attention to and what you can safely ignore. “If it is not making you happier, or healthier, or calmer, or having better relationships, or wealthier, then what good is it? It's useless. You can safely discard it.” @naval
10:34 PM ∙ Sep 6, 2020
4,723Likes867Retweets
Twitter avatar for @jasoncrawford
Jason Crawford @jasoncrawford
Most people see @BillGates's life in two parts: Act 1: make money, Act 2: give it away. The first seen as greed/ambition, the second as generosity. I think this is a mistake. I see much more continuity between the two acts.
6:20 PM ∙ Apr 6, 2020
444Likes124Retweets
Twitter avatar for @jasoncrawford
Jason Crawford @jasoncrawford
I was asked by @rrana03_ about the mindset behind scientific and technological progress. A few brief thoughts:
4:36 PM ∙ Aug 31, 2020
21Likes4Retweets
Twitter avatar for @jasoncrawford
Jason Crawford @jasoncrawford
Thinking clearly and well is a skill that deserves lifelong continuous improvement. (Call it rationality, objectivity, whatever) How do you improve your epistemology? A few things I came up with, would love to hear your ideas:
11:04 PM ∙ Jul 18, 2020
122Likes9Retweets
Twitter avatar for @sama
Sam Altman @sama
Vector theory of impact:
3:04 PM ∙ Aug 24, 2020
2,246Likes533Retweets
Twitter avatar for @JoaquimCampa
Joaquim Campa @JoaquimCampa
Thread of Artists With Unusual Art Mediums 1. Salavat Fidai, Russian sculptor who makes stunning miniature sculptures out of pencil graphite
Image
Image
Image
Image
6:49 AM ∙ Sep 6, 2020
1,438Likes628Retweets
Twitter avatar for @ryanstephens
Ryan Stephens 🥃 @ryanstephens
Picasso created more than 50,000 works of art. How many are considered masterpieces that we still admire today? About a 100. Less than 1% of his creations are still relevant. Stop trying to be perfect. It's a numbers game. Start creating. Be courageous enough to share.
4:00 PM ∙ Nov 27, 2019
11,646Likes3,562Retweets
Twitter avatar for @DrewCoffman
Drew Coffman 𝕚𝕤 𝕠𝕟𝕝𝕚𝕟𝕖 🟢 @DrewCoffman
This VFX work is absolutely incredible, and it becomes even more amazing when I tell you that it’s made in Blender (which is free and open-source) and the work of a SINGLE PERSON (@Mrdodobird, who you should follow)
Image
4:38 PM ∙ Jun 21, 2020
25,077Likes6,645Retweets
Twitter avatar for @rosstraining
Ross Enamait @rosstraining
Whenever I share outdoor footage, someone always misinterprets the message. I'm not saying to cancel your gym membership. I don't care where you train. Just don't come at me with an excuse that you have nowhere to train. You can always do something. instagram.com/rosstraining
Image
5:30 PM ∙ Dec 6, 2019
1,429Likes215Retweets
Twitter avatar for @jasoncrawford
Jason Crawford @jasoncrawford
So, let's say it's the 1840s, and you're tasked with building a suspension bridge at Niagara Falls, spanning a gorge almost 800 feet across and over 200 feet deep. How do you get the *first* metal cable across?
Image
10:54 PM ∙ Jan 5, 2020
555Likes152Retweets
Twitter avatar for @sriramk
Sriram Krishnan @sriramk
What are labs/environments/companies which had outsized output in engineering/science output in the last 100 years? For example: thinking of Bell Labs a few decades ago, PARC, Lockheed Martin's Skunkworks team in the 70s/80s. But also curious about science/other eng fields.
2:38 PM ∙ Sep 30, 2019
43Likes3Retweets
Twitter avatar for @david_perell
David Perell @david_perell
Every discipline has a book that everybody references, but nobody reads: Physics: The Feynman Lectures Philosophy: Godel, Escher, Bach Sociology: Das Kapital English: Infinite Jest Media: Understanding Media Investing: The Intelligent Investor Economics: The Wealth of Nations
5:01 AM ∙ Sep 10, 2020
2,290Likes316Retweets
Twitter avatar for @robgrav3s
Rob Graves @robgrav3s
Nothing is real. Thread.
4:38 AM ∙ Nov 20, 2019
101Likes26Retweets
Twitter avatar for @fermatslibrary
Fermat's Library @fermatslibrary
A demonstration of the conservation of angular momentum using a Hoberman sphere. Angular momentum depends on the rotational velocity of an object, and on its rotational inertia. When an object changes its shape, its angular velocity will also change.
Image
2:34 PM ∙ Nov 9, 2019
43,263Likes7,914Retweets
Twitter avatar for @lDidNotKnowThat
I Didn't Know That @lDidNotKnowThat
This is how to do 48 different exercises
Image
7:13 AM ∙ Jul 12, 2020
26,517Likes7,649Retweets

That’s it for today’s weekly reporting.

Until next week

Share this post

8: Should we be worried about AGI?; Are social platforms programming us?; and more..

mohamadahmad.substack.com
Comments
TopNew

No posts

Ready for more?

© 2023 Mohamad Ahmad
Privacy ∙ Terms ∙ Collection notice
Start WritingGet the app
Substack is the home for great writing