Will Summerlin @WSummerlinAI Twitter profile | Pikagi

Pikagi

Will Summerlin

@WSummerlinAI

23,736

Followers

187

Following

41

Media

964

Statuses

Managing Partner at Autopilot ( @hqautopilot ), a research-driven investment firm. Previously investing in AI across the public and private markets @ARKInvest .

https://t.co/YLZMBfz51W

Joined March 2021

Don't wanna be here? Send us removal request.

Pinned Tweet

@WSummerlinAI

Will Summerlin

4 months

Introducing Autopilot @maxmhfriedrich and I have spent the last several years together researching and investing in fintech and artificial intelligence across the public and private markets. We have witnessed extraordinary founders build category-defining companies - Block

6

0

30

Last Seen Profiles

@CrotJarik1

@jxhn555

@romulopassos_

@Patino_____

@Reiy69

@Revanth_sai2508

@Prettytootsy1

@RakshanVJ

@NC010203

@suntihill

@elenalingshao

@richyieze

@sitiazieraa

@shemale_tanvi

@tanu_unat_

@baz_christ

@luvrpow

@sevimbensevim

@tsukku0727

@bntaldm76323677

@same_nyan_sen

@sckrgllr

@tomerbrurian

@sweetiebelleth

@stakewolle

@Alexand92840726

@sserhatkeskin

@bokeplokalmalam

@UCVR_

@soonwoophile

@sipotexxx

@sambuduris_05

@riri_coeota1616

@tladpwl03

@CityscootMilano

@raysucassio

@WSummerlinAI

Will Summerlin

1 year

Game theory on the Bing announcement… Microsoft is trying to lower Google’s search margins which will make it harder for Google to continue running Cloud and other competitive businesses at a loss. Google processes ~8.5 billion searches per day. The noise around Bing/AI

117

468

5K

@WSummerlinAI

Will Summerlin

3 years

The Vision vs. Lidar debate is reminiscent of the AC vs. DC battle of the late 1880s. Direct Current (DC) systems, as proposed by Thomas Edison, worked reasonably well but could not transfer power over long distances. (1/5)

110

239

2K

@WSummerlinAI

Will Summerlin

3 years

So much to unpack from Tesla AI day. Here are a few key takeaways:

23

318

2K

@WSummerlinAI

Will Summerlin

3 years

How much would it cost to train a neural network the size of a human brain? (thread)

46

89

1K

@WSummerlinAI

Will Summerlin

3 years

But neural network training efficiency has been doubling every 16 months. If this trend continues, the cost of training will likely decline as follows: 2024: $325mm 2028: $40mm 2032: $5mm

36

54

997

@WSummerlinAI

Will Summerlin

3 years

MIT CS PhD: “I want to solve cancer with AI!” FAANG: "Good for you. Here's $1.2mm/yr. Make people look at their screen for an additional 0.85 seconds."

11

112

938

@WSummerlinAI

Will Summerlin

3 years

The evolution of competitive advantages: 1900s: Economies of Scale 2000s: Network Effects 2020s: Data Advantages 1/14

30

216

921

@WSummerlinAI

Will Summerlin

3 years

Telsa won the currents battle with AC. And I think Tesla will win the FSD battle with Vision. In both cases, scalability matters. (But ironically, Tesla superchargers are DC systems!) (5/5)

50

62

907

@WSummerlinAI

Will Summerlin

3 years

GPT-3 has 175B parameters (3.14E+23 flops) and it cost ~$4.6mm to train (2020). The human brain has at least 100T synapse (parameters). Assuming linear scaling of compute requirements, the cost to train a “human brain” in 2021 would be more than $2B.

32

78

868

@WSummerlinAI

Will Summerlin

3 years

hello world

189

18

832

@WSummerlinAI

Will Summerlin

2 years

AI is the next great platform shift. It will create an order of magnitude more value than mobile, cloud, etc. Companies that embrace it will likely see massive upside. Companies that don’t will be disrupted. Surprisingly few companies are investing accordingly.

85

104

625

@WSummerlinAI

Will Summerlin

3 years

Training a neural network

Tweet media one

15

30

545

@WSummerlinAI

Will Summerlin

3 years

My take: AI is likely advancing far faster than most people realize. And Tesla appears to be the leader when it comes to real-world AI. The visual perception stack and training pipeline are extensible to many use cases beyond FSD (Tesla Bot?), and the data advantages are real.

10

33

513

@WSummerlinAI

Will Summerlin

3 years

Many claimed DC was the only "safe" way to transfer electricity. Similarly, many today claim that Lidar is the only "safe" option for FSD. Both were/are wrong. (4/5)

13

25

455

@WSummerlinAI

Will Summerlin

2 years

Our Zoom open-source model is now available on GitHub: You can read about our thesis and assumptions in our blog post: Our base case suggests Zoom’s share price could reach $1,500 by 2026.

Tweet card media

ARK’s Expected Value For Zoom Video In 2026: $1,500 Per Share

ARK’s updated open-source Zoom model yields an expected value per share of $1,500 in 2026. Please explore and test our assumptions.

www.ark-invest.com

120

90

413

@WSummerlinAI

Will Summerlin

1 year

@Kashish__Kumar_ Satya is playing 3D chess

6

7

393

@WSummerlinAI

Will Summerlin

3 years

Whoever is in charge of @ChickfilA ’s drive-thru operations should be put in charge of COVID testing.

13

35

378

@WSummerlinAI

Will Summerlin

2 years

AI training costs are declining at 60% YoY. It cost ~$5mm GPT-3 in 2020. By 2030, it will likely cost ~$500 to train a model to the same level of performance. AI training is Moore’s law on steroids.

@nealkhosla

Neal Khosla

2 years

If you look at some of the best investors in SV it basically comes down to their ability to “trust the exponential.” Turns out this is easy to say and almost impossible to do. Humans just not wired to understand power laws.

6

9

84

16

70

369

@WSummerlinAI

Will Summerlin

3 years

Vector Space: Tesla is predicting in vector space, not image space (most computer vision systems operate in 2D image space). It makes sense, given cars operate in a 3D world, but it's very hard to build.

3

8

352

@WSummerlinAI

Will Summerlin

3 years

Nikola Tesla proposed an Alternating Current (AC) system which transferred power over long distances at high voltage and then stepped down the voltage to make it safe and usable. Similar to Vision-based systems, AC was easier and less expensive to scale. (3/5)

4

11

324

@WSummerlinAI

Will Summerlin

3 years

As a result, power generation in a DC system had to be decentralized, increasing cost and slowing expansion. Lidar-based systems are similar in that they work reasonably well, but they are slow and expensive to scale. (2/5)

5

10

302

@WSummerlinAI

Will Summerlin

3 years

AI is probably just a few years away from a perfect SAT score.

@gdb

Greg Brockman

3 years

We've built an AI system that solves grade school math word problems. Starting to achieve decent accuracy & represents significant progress on top of GPT-3:

Tweet media one

35

126

693

13

41

267

@WSummerlinAI

Will Summerlin

3 years

1996: “The internet is going to collapse”

@PessimistsArc

Pessimists Archive

3 years

🗓️ This day in 1993 the World Wide Web is open sourced. By 1996 its 'inevitable' collapse was predicted and made headlines...

Tweet media one

13

154

448

21

45

263

@WSummerlinAI

Will Summerlin

3 years

Our research suggests that AI tools could make engineers, designers, lawyers, etc., 140% more productive by 2030. The impact on software spend could be profound...more to come in Big Ideas 2022.

@sama

Sam Altman

3 years

I have now gotten enough of a taste of AI-powered creative tools to know that they're going to be much better than even the AI optimists think. So cool to just think of ideas and iteratively have the computer implement and build on them.

71

184

2K

18

36

255

@WSummerlinAI

Will Summerlin

1 year

The lag between technology adoption and productivity gain is known as the “productivity paradox. Total factor productivity doesn’t really increase until organizations and processes are redesigned to take full advantage of a new technological capability. The same will be true for

10

73

243

@WSummerlinAI

Will Summerlin

3 years

Memory: Tesla has inserted a video module and feature queue model into their NN architecture. This effectively gives the NN short-term memory Short-term historical context is useful when driving. For example, did the lane marking seen 5 seconds ago say this is a left turn lane?

4

7

252

@WSummerlinAI

Will Summerlin

3 years

Legacy companies are going to struggle with AI for two reasons: 1/8

13

36

242

@WSummerlinAI

Will Summerlin

3 years

Fusion: Per-camera detection then fusion works fine in image space but doesn't translate well to vector space. Accurate depth per pixel is critical for vector space predictions.

1

5

245

@WSummerlinAI

Will Summerlin

3 years

An iPhone 12 with an A14 chip is capable of 11 TFLOPS (trillion float operations per second). It's 68,750x faster than the first supercomputer! Bonus: The A14 chip is ~898,325,847x faster than the Apollo 11 computer (12,245 FLOPS) that guided us to the moon.

@MIT_CSAIL

MIT CSAIL

3 years

45 years ago today the first major supercomputer was installed. The Cray-1 had a 64-bit processor and a top speed of 160 MFLOPS. For comparison, today’s top supercomputers are 20 billion times faster.

Tweet media one

8

141

464

10

33

239

@WSummerlinAI

Will Summerlin

3 years

Tesla solves the above fusion problems by fusing image data before detections. See the improvement in the screenshot below. Bottom left: fusing images after detections Bottom right: fusing images before detections. Much better!

Tweet media one

2

7

240

@WSummerlinAI

Will Summerlin

3 years

Data (images, GPS, etc.) from multiple vehicles can align to reconstruct roads and 3D obstacles for labeling. The result is that many roads and 3D obstacles can be auto-labeled. More customers = more labels = better product = more customers. Data advantage flywheel spins!

3

6

224

@WSummerlinAI

Will Summerlin

3 years

But they will happily pay a management consulting firm $25 million for an 85-page PPT on the importance of AI! 8/8

11

10

219

@WSummerlinAI

Will Summerlin

3 years

Another problem with per-camera detection is that object prediction is hindered when target objects are only partially visible. For example, no single camera has a complete picture of the truck in the screenshot below:

Tweet media one

2

4

215

@WSummerlinAI

Will Summerlin

3 years

Spatial RNN: The spatial RNNs allow multiple traversals (trips) to collaborate and effectively build a pseudo-HD map in a space of features in a NN.

Tweet media one

4

3

213

@WSummerlinAI

Will Summerlin

3 years

The visual perception system can predict depth as accurately as radar without the complexity of additional sensor fusion. The blue line represents depth prediction with video, the green line with radar.

Tweet media one

4

8

208

@WSummerlinAI

Will Summerlin

3 years

Big Ideas 2022: From 2022 -> 2030, AI training costs could decline by 60% annually. Within the next eight years, AI could boost the output of knowledge workers by 140% (+$56T in global output). By 2030, AI companies could be worth $87T. (w/ @downingARK and @wintonARK )

4

46

196

@WSummerlinAI

Will Summerlin

2 years

OpenAI charges $0.03 for DALL·E to generate an image. By our estimates, a graphic designer would charge ~$150. We estimate inference costs ~$0.005 per image (~22 seconds per image on an A100 with hourly rate of $0.87). >80% gross margins, consistent with SaaS

@OpenAI

OpenAI

2 years

DALL·E 2 is officially now in beta. We’ll be inviting 1 million people from our waitlist over the coming weeks.

213

625

3K

11

34

185

@WSummerlinAI

Will Summerlin

3 years

The above updates also improve object detection when temporary occlusion occurs. Notice how the updated system (blue cars) knows there are cars behind the truck, even though they are temporarily blocked.

Tweet media one

1

4

190

@WSummerlinAI

Will Summerlin

2 years

"Over four years, the bank has automated over 250 critical business operations and saved more than 400,000 hours (with UiPath)."

Tweet card media

UiPath Empowers Sumitomo Mitsui Trust Bank | UiPath

UiPath announced that Sumitomo Mitsui Trust Bank has deployed the UiPath Platform, resulting in savings of more than 400,000 total hours in labor-intensive tasks.

22

26

180

@WSummerlinAI

Will Summerlin

3 years

Simulation: Real-world data is better, but simulation can be useful where situations are difficult to source (i.e. people running down a highway) or difficult to label (i.e. 100 pedestrians).

2

4

181

@WSummerlinAI

Will Summerlin

3 years

The hardware advancements deserve their own thread, so I'll stop here.

3

3

177

@WSummerlinAI

Will Summerlin

3 years

(All figures are estimates) Sources: @lexfridman

Tweet card media

Measuring the Algorithmic Efficiency of Neural Networks

Three factors drive the advance of AI: algorithmic innovation, data, and the amount of compute available for training. Algorithmic progress has traditionally been more difficult to quantify than...

6

15

173

@WSummerlinAI

Will Summerlin

3 years

Labeling: Manual labeling is done in 4D vector space and projected to image space — massive improvement in throughput.

Tweet media one

1

2

169

@WSummerlinAI

Will Summerlin

3 years

Planning: Planning is currently done in a less-than-optimal way. There is an opportunity to apply NNs as heuristics to improve search through action space. Doing so will prevent the search from getting trapped in local minima.

Tweet media one

2

7

169

@WSummerlinAI

Will Summerlin

3 years

Bears: “Zoom is losing market share to Microsoft.” Data: Zoom has held market share between 45-50% since Q1 2020 (according to traffic estimates from a third-party source). And according to Gartner, customers like Zoom.

Tweet media one

21

12

154

@WSummerlinAI

Will Summerlin

3 years

@facebookai recently announced research related to self-supervised learning and Visual Transformers (DINO and PAWS). The coupling of self-supervised learning and Visual Transformers is quite exciting. Here’s why: (1/9)

DINO and PAWS: Advancing the state of the art in computer vision

Working with Inria researchers, we’ve developed a self-supervised image representation method, DINO, which produces remarkable results when trained with Vision Transformers. We are also detailing...

7

31

141

@WSummerlinAI

Will Summerlin

3 years

The future is going to be cool

@Google

Google

3 years

What is Project Starline? It combines research in computer vision, machine learning, spatial audio and real-time compression with a breakthrough 3D display. The effect is the feeling of a person sitting just across from you, like they are right there.

27

173

790

7

19

134

@WSummerlinAI

Will Summerlin

1 year

A simple framework for identifying high-value applications of AI: 1. AI can partially or fully automate a task, resulting in 10x better, cheaper, and/or faster outcomes 2. Humans are currently paid >$10B annually to perform that task 3. Usage of the product generates

4

21

125

@WSummerlinAI

Will Summerlin

2 years

Our research suggests that by 2030, AI coding tools could double the productivity of human software developers. We might be underestimating both the timing and the magnitude of impact. (1/4)

Tweet media one

@GoogleDeepMind

Google DeepMind

@GoogleDeepMind

2 years

Introducing #AlphaCode : a system that can compete at average human level in competitive coding competitions like @codeforces . An exciting leap in AI problem-solving capabilities, combining many advances in machine learning! Read more: 1/

Tweet media one

175

2K

8K

4

25

124

@WSummerlinAI

Will Summerlin

3 years

It’s my fourth day living in LA…and someone just walked into Starbucks with their pet pig

18

1

121

@WSummerlinAI

Will Summerlin

10 months

Important announcement: Today is my last day @ARKInvest . I couldn't be more grateful to have had the opportunity to work alongside @CathieDWood , @wintonARK , and the rest of the ARK team. We've had an incredible front-row seat as the AI revolution has accelerated over the past two

13

3

121

@WSummerlinAI

Will Summerlin

3 years

Will data advantages drive the next 70% of value creation in tech? Maybe. 14/14

14

2

117

@WSummerlinAI

Will Summerlin

3 years

Translation: We believe that within the next decade, AI will boost the productivity of knowledge workers (lawyers, software engineers, etc.) by nearly 140%. Working without AI = walking Working with AI = riding a motorcycle

@wintonARK

Brett Winton

3 years

We’re entering the age of Motorcycles for the mind

8

3

58

5

14

116

@WSummerlinAI

Will Summerlin

3 years

Meet our newest Analyst @downingARK ! Prior to @ARKInvest , he built infrastructure/software to support billions of users at some of the largest companies in the world.

@downingARK

Frank Downing

3 years

Hi, I’m Frank!

176

16

633

9

7

108

@WSummerlinAI

Will Summerlin

3 years

In addition, the convergence of technologies means the implication of AI will likely be more far-reaching than we can imagine.

Tweet media one

5

24

104

@WSummerlinAI

Will Summerlin

3 years

Data is the current limiting factor in AI. Infra is pretty good, and models simply need more training data to push the limits of what is possible. 1/4

9

8

104

@WSummerlinAI

Will Summerlin

3 years

A fascinating observation from AI gameplay (MuZero, etc.) is that an AI system usually wins by employing unique, unexpected strategies. 1/3

5

17

100

@WSummerlinAI

Will Summerlin

3 years

“Move fast and break things.” “Wait, no, don't break that!”

0

6

99

@WSummerlinAI

Will Summerlin

3 years

Go follow our new FinTech Research Associate! @nishitaARK Total underachiever with a double major from @Penn + experience at @Apple , @Facebook , @McKinsey , @PanteraCapital

@nishitajain_

nishita

3 years

Hello friends of Twitter!

98

8

383

10

8

97

@WSummerlinAI

Will Summerlin

9 months

I went to a party yesterday with a bunch of non-tech people, expecting everyone to be following the OpenAI drama and discussing governance structures. Turns out we live in a silo on tech twitter.

7

6

98

@WSummerlinAI

Will Summerlin

3 years

Most people go to college because it’s “what you’re supposed to do.” But that seems like a crazy reason to spend >$50k and four years of your life. 4/4

10

11

91

@WSummerlinAI

Will Summerlin

1 year

IMO @Replit has a shot at building one of the most valuable companies on earth. Probably a good time to join!

@amasad

Amjad Masad

1 year

Replit is at an insane inflection point. The technology is about to get powerful enough to power entire businesses, revenue is scaling rapidly, and we're riding multiple massive tech trends. Yet it's early enough to have an insane impact and upside

Tweet media one

7

4

104

3

20

88

@WSummerlinAI

Will Summerlin

2 years

Just a reminder that in Q4 FY22 Zoom reported a $5.4B cash position with a 38% adjusted free cash flow margin on >$4B ARR

15

6

89

@WSummerlinAI

Will Summerlin

3 years

Great conversation with @heydave7 and @downingARK !

@heydave7

Dave Lee

3 years

New episode 🔥 Tesla AI Day Unpacked w/ ARK Analysts @summerlinARK & @downingARK (Ep. 415)

9

29

274

1

3

86

@WSummerlinAI

Will Summerlin

3 years

The companies with the best training data will generally produce the most performant models. Those companies will attract more customers who will generate even more data. 7/14

1

3

90

@WSummerlinAI

Will Summerlin

3 years

It just shows how humans have explored very little of the possible solution space. Can't wait to see how AI systems discover unexpected but optimal solutions to problems outside of gaming. 3/3

6

8

88

@WSummerlinAI

Will Summerlin

3 years

Network Effects: A study by @nfx found that 70% of value creation in tech is driven by network effects. 4/14

1

3

84

@WSummerlinAI

Will Summerlin

1 year

Human knowledge workers are paid ~$32T annually. It appears AI will be capable of automating most knowledge work within the decade (likely a lot sooner). The adoption rate of automation is very high, given the often extraordinary ROI. I.e., ChatGPT grew from 0-100mm users in 60

8

16

86

@WSummerlinAI

Will Summerlin

3 years

Data Advantages: Contrary to common belief, most data is useless or of little value. But in a few circumstances, data is gold. As products embed AI, large quantities of high-quality training data are needed to train models. 6/14

3

4

86

@WSummerlinAI

Will Summerlin

2 years

Low margin but high quality hardware has become a distribution channel for high margin recurring software. Similar to the old model of Kuerigs & K-Pods, razors & razor blades, etc.

10

9

86

@WSummerlinAI

Will Summerlin

3 years

A network effect exists when one user creates value for many other users in a system (think Facebook, Twitter, etc.). The larger the network, the more value it has, and the harder it is to replicate. This exponential value creation can be modeled with Reed's Law: 5/14

Tweet media one

2

3

82

@WSummerlinAI

Will Summerlin

2 years

The adoption rate of AI products is incredible. The AI supercycle has started, and it’s likely going to create an order of magnitude more value than past supercycles.

@sama

Sam Altman

2 years

ChatGPT launched on wednesday. today it crossed 1 million users!

1K

3K

49K

5

9

82

@WSummerlinAI

Will Summerlin

3 years

E.g. in autonomous driving, companies that are already collecting real-world data will likely have the most performant models in the short-term. That performance will probably attract more customers who collect more data, improve the models, and widen the competitive moat. 9/14

1

4

79

@WSummerlinAI

Will Summerlin

3 years

2. Knowledge is now widely available at a low cost. The content on Coursera, etc. might be better than most college courses. 3. The cost of tuition is increasing with no limit in sight. 3/4

Tweet media one

4

6

76

@WSummerlinAI

Will Summerlin

3 years

Two things to consider when thinking about data advantages: 1) Is the data hard to acquire? 2) Where is the data-value asymptote? 10/14

2

3

80

@WSummerlinAI

Will Summerlin

3 years

Investor relations people - check out @twilio ’s new earnings call format. Call was Q&A only. A document with prepared remarks was available before the call. Much more time-efficient approach.

6

4

79

@WSummerlinAI

Will Summerlin

3 years

@elonmusk Marv from @OpenAI thinks he’s funny

9

6

75

@WSummerlinAI

Will Summerlin

3 years

Conversely, I doubt that legacy companies will pay 27 y/o engineers $1mm+. Nor will they give them unlimited vacation or the autonomy to work on problems they find interesting. 7/8

2

0

75

@WSummerlinAI

Will Summerlin

3 years

Economies of Scale: In the 1900s, the dominant companies benefited from scale. Standard Oil's size enabled Rockefeller to negotiate railroad rebates, acquire early tank cars, etc. He could profitability sell oil at a price point lower than his competitors could produce it. 2/14

1

3

75

@WSummerlinAI

Will Summerlin

3 years

This creates a flywheel effect where the robustness and performance of a system increase exponentially over time. The result is a competitive moat that strengthens with scale. 8/14

1

3

73

@WSummerlinAI

Will Summerlin

3 years

This will inevitably confuse and anger the risk-averse middle managers at legacy companies. They will default to “no” because their incentive structure rewards stable, low-risk execution. 3/8

1

1

75

@WSummerlinAI

Will Summerlin

3 years

Autonomous driving data is hard to acquire - you need many cars on the road operating in various conditions. Facial images (for face recognition) are easy to acquire from public datasets (labeled faces in the wild, etc.), so the value of facial data is implicitly lower. 11/14

1

1

72

@WSummerlinAI

Will Summerlin

3 years

Economies of scale aren't as powerful in software because the underlying infrastructure components - internet, servers, etc. - are generally shared resources. The cost of compute isn't that much different for BigCo and SmallCo. And scale is available instantly. 3/14

1

1

74

@WSummerlinAI

Will Summerlin

3 years

Reducing system complexity is hard but usually worth it in the long run. Most impressive fact from @karpathy recent presentation is that Tesla was able to shadow test vision-only on existing fleet, generating 1,000 years worth of driving data.

@TashaARK

Tasha Keeney

3 years

Just wrote about Tesla's decision to remove radar in our latest newsletter. Would love to hear anyone's thoughts/ experiences with the tradeoff of including additional sensors to fuse into an autonomous tech stack.

41

39

363

1

6

70

@WSummerlinAI

Will Summerlin

3 years

Other AI problems require little training data. The classic "handwritten number recognition" problem can probably be solved with just a few thousand labeled images. In this case, the data-value asymptote is low. Adding more data would only marginally increase performance. 13/14

3

1

71

@WSummerlinAI

Will Summerlin

3 years

@TrumanFalkner I think so!

3

0

66

@WSummerlinAI

Will Summerlin

3 years

First, they are risk-averse. Deep learning models are black-box, and it’s kind of impossible to explain why a model produced a specific output. Sometimes the output is totally unexpected. Sometimes it’s offensive. Remember Tay, Microsoft’s offensive chatbot? 2/8

2

0

67

@WSummerlinAI

Will Summerlin

3 years

Is the value of a college degree increasing or decreasing? 1/4

Increasing

332

Decreasing

2097

17

22

65

@WSummerlinAI

Will Summerlin

2 years

The consensus opinion (at least for institutional investors) appears to be “sell tech near the potential low” and “buy value/energy near the potential high.” 🤷🏻‍♂️

6

5

65

@WSummerlinAI

Will Summerlin

2 years

To build a great AI model, you need data, compute, and AI/ML talent. You can buy compute, you might be able to collect data, but you probably can’t recruit top-tier AI/ML talent. Top-tier talent flows to @Tesla , @MetaAI , @DeepMind , @OpenAI , etc.

6

9

66

@WSummerlinAI

Will Summerlin

3 years

Interesting thought by @sama on AI-enabled Moore's law of everything: "Imagine a world where, for decades, everything–housing, education, food, clothing, etc.–became half as expensive every two years." 1/11

5

11

65

@WSummerlinAI

Will Summerlin

3 years

Founder-led tech companies tend to be more aggressive and risk-tolerant. They will push the limits and reap the benefits. 4/8

1

0

65

@WSummerlinAI

Will Summerlin

3 years

Some AI problems, like autonomous driving, require massive amounts of training data. It will probably be a long time before autonomous driving data reaches a point of diminishing returns. 12/14

2

1

64

@WSummerlinAI

Will Summerlin

3 years

1000x more powerful than BERT, the last transformer-powered search model

@Google

Google

3 years

Multitask Unified Model (MUM) — our latest AI milestone — has the potential to transform how Google helps you with complex information tasks. #GoogleIO

27

257

1K

2

9

65

@WSummerlinAI

Will Summerlin

1 year

The plow increased the productivity of farmers by 5-6x. That productivity uplift enabled significant portions of the population to shift from food gathering (the primary “job” at the time) to producing specialized goods and services. It set the stage for significant economic

7

13

63

@WSummerlinAI

Will Summerlin

3 years

Second, legacy companies will struggle to attract top-tier talent. There are (maybe?) a few thousand great AI practitioners. Top tech companies know how to recruit, retain, and maximize the productivity of top-tier talent. Legacy companies don't. 5/8

1

0

61

@WSummerlinAI

Will Summerlin

2 years

The first operational US fighter jet was designed and built in 143 days. The 1,700 mile Alaska highway was built in 234 days. The city of SF has been building a bus lane on Van Ness for 7,600 days....

3

4

58

@WSummerlinAI

Will Summerlin

3 years

@wintonARK I have been told I am human

6

1

57

@WSummerlinAI

Will Summerlin

9 months

@heydave7 ARK venture fund is the closest with SpaceX, Tesla, and X.

14

0

58

@WSummerlinAI

Will Summerlin

3 years

Top tech company culture promotes autonomy. They pay very well (over $1mm for good engineers) and have cool brands. And networks are insular in that great talent follows great talent. 6/8

1

0

59