Amazon v. Brian Hall - Declaration of Matt Wood, AWS VP of AI

Download as pdf or txt
Download as pdf or txt
You are on page 1of 24
At a glance
Powered by AI
The document discusses ML (machine learning) and how AWS and Google compete in providing ML cloud services to customers. It also details concerns around a former AWS employee, Brian Hall, taking a role at Google based on the confidential information he learned at AWS.

ML involves using algorithms and data to enable computers to learn and improve. AWS offers customers ways to incorporate ML technologies into their own businesses for use on their own data sets. This is an important business opportunity that influences which cloud provider (AWS, Google, Microsoft) customers choose.

AWS and Google provide competing ML cloud services in a growing multi-billion dollar market. They compete for thought leadership, mind share and customers by showing leadership in ML technologies. Studies show customers often choose their provider based primarily on its ML capabilities.

1 FILED The Honorable Sean O’Donnell

2020 JUN 09 09:00 AM


2 KING COUNTY
SUPERIOR COURT CLERK
3 E-FILED
CASE #: 20-2-08977-0 SEA
4

7 SUPERIOR COURT OF THE STATE OF WASHINGTON


IN AND FOR KING COUNTY
8
)
9 AMAZON.COM, INC., )
) No. 20-2-08977-0 SEA
10 Plaintiff, )
) DECLARATION OF
11 v. ) MATT WOOD, PhD
)
12 BRIAN HALL, ) Confidential Motion to Seal/Redact
) Pending
13 Defendant. )
14

15 I, MATT WOOD, declare:

16 1. Identity of Declarant. I have worked for Amazon Web Services, Inc. (“AWS”)

17 for over 10 years. My current title is Vice President of Artificial Intelligence, a position I have

18 held since September 2019. Before I became AWS’s Vice President of Artificial Intelligence, I

19 was its General Manager of Deep Learning and Artificial Intelligence. I held that position from

20 January 2017 until September 2019. Before that, I was AWS’s General Manager of Product

21 Strategy, which position I held from October 2012 until January 2017, where I acted as a script

22 writer for Andy Jassy, CEO of AWS.

23 2. Basis for Declaration. In my position and as part of my employment, I work on

24 issues that give me familiarity with and access to Amazon business records and emails. I am

25 very familiar with Brian Hall’s work for Amazon from when he joined Amazon in June 2018

26 until his resignation in February 2020. We worked together on numerous projects and

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 1 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
1 frequently corresponded and attended in-person meetings together. Except as otherwise

2 indicated, all facts set forth in this declaration are based on my personal knowledge, which is

3 based on my experience and my review of relevant documents. The facts and documents set

4 forth below are maintained in the ordinary course of business at or near the time of the act,

5 condition, or event. If called upon to testify regarding such matters, I could competently do so.

6 BACKGROUND

7 3. My Position at AWS. At AWS, I lead all product development across AWS’s

8 Machine Learning (“ML”) efforts, described below. I partner closely with all the AWS

9 stakeholder teams for ML cloud, such as finance, HR, partner, sales, business development, and

10 marketing, among others.

11 4. ML. ML is a computer science field. At its core, ML is a way to teach a

12 computer to improve its ability to perform tasks and to mimic the ways by which humans learn,

13 such as repetition and experience. Organizations use ML to address a host of business

14 challenges, from product recommendations and pricing predictions, to tracking disease

15 progression and demand forecasting.

16 5. AWS ML. Many of AWS’s customers use software that incorporates ML

17 algorithms. AWS ML offers Amazon’s cloud computing customers ways to incorporate ML

18 technologies into their own businesses for use on their own data sets. Given the breadth of its

19 impact across a wide range of applications and industries, ML is an important long-term

20 business and revenue opportunity for Amazon. ML has emerged as a beachhead as to how

21 Amazon customers choose their cloud provider (i.e., AWS, Google Cloud, or Microsoft).

22 6. AWS ML’s Competition with Google ML. AWS and Google provide ML cloud

23 services that compete head-to-head in a market that is growing by billions of dollars per

24 year. For example, like AWS, Google provides a host of pre-programmed cloud ML

25 applications, including those related to computer vision and speech recognition, in direct

26 competition with AWS. AWS competes with Google for ML thought leadership and mind

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 2 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
1 share, which requires a showing that AWS has deep, broad, and groundbreaking technologies.

2 Studies show that customers often choose their cloud provider primarily based on that

3 provider’s ML capabilities. Mr. Hall was a crucial part of AWS’s discussions regarding what

4 and where to push AWS’s capabilities forward, why and for how long to defer where AWS had

5 opportunities for improvement, and how to develop AWS’s thought leadership in ML,

6 especially vis-à-vis Google.

7 7. Declarant’s Relationship with Mr. Hall. In June 2018, Mr. Hall became

8 AWS’s Vice President of Product Marketing. From June 2018 to February 2020, I worked

9 continuously and closely with Mr. Hall. Among other things, I worked closely with Mr. Hall in

10 the development of marketing plans for AWS’s yet-to-be-launched ML cloud products.

11 8. As with all AWS products, Mr. Hall was responsible for strategically developing

12 the “story” around AWS’s ML products prior to their launch. Mr. Hall and I worked together

13 on marketing strategies for numerous AWS ML products that AWS has yet to launch and

14 deciding where marketing efforts would be most effective in driving AWS’s message and the

15 product’s adoption. By virtue of Mr. Hall’s involvement in creating related marketing plans,

16 Mr. Hall possesses significant amounts of highly confidential non-pubic information regarding

17 these yet-to-be launched AWS ML products. This includes technical specifications, intended

18 customers, messaging strategies, pricing strategies, and anticipated launch dates; the amount of

19 marketing effort AWS anticipates is necessary; and which customers have requested such

20 products.

21 MR. HALL’S EXPOSURE TO AWS’ ML STRATEGY

22 9. Mr. Hall also led development of the overall product marketing strategy for

23 AWS ML, including but not limited to his role in: (a) analyzing AWS’s operational planning

24 documentation (“OP1”), which sets out the strategy for AWS’s ML business through the end of

25 2020; (b) participating in weekly business reviews of key metrics and status reports on AWS’s

26 overall ML strategy; and (c) analyzing monthly update reports regarding the status and

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 3 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
1 competitive positioning of the cloud ML business and participating in presentations of these

2 reports to AWS’s CEO, Andy Jassy. Examples of business reviews in which Mr. Hall

3 participated are attached as exhibits. Exhibit A is a true and correct copy of an AI Weekly

4 Business Review (“WBR”) Summary; Exhibit B are true and correct copies of AWS’s

5 December 2019 and February 2020 ML Marketing Monthly Business Review (“MBR”)

6 Summaries.

7 10. By way of example, Mr. Hall’s involvement in high-level ML strategy meetings

8 is reflected in Exhibit C, which is a true and correct copy of a February 5, 2020 email that both

9 Mr. Hall and I received regarding a high-level ML strategy meeting to discuss AWS’s overall

10 business plan for AWS ML in 2020. Exhibit D is a true and correct copy of the highly

11 confidential 2020 AWS Machine Learning Category Market Plan attached to that email. This

12 document contains extremely sensitive AWS information, including AWS’s customer analysis,

13 AWS’s strategic business goals for 2020, specific AWS initiatives to be launched in 2020, and

14 AWS’s analysis as to how it is stacking up against its competitors in the marketplace—

15 specifically including Google.

16 11. Mr. Hall’s Role in Competitive Analysis of Google. A large part of Mr. Hall’s

17 role with AWS ML was product positioning—i.e. ensuring AWS’s ML products were

18 competitive in the market. Through

19

20

21

22

23

24
25

26

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 4 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
1 AWS developed an ML powered camera, DeepLens, t

2 in pushing AWS forward as an ML innovator. Microsoft and Google both copied this effort –

3 the day after AWS’s announcement of DeepLens, Google even announced that it would soon

4 have a camera available too, but the damage was done and AWS was at the forefront. With

5 Mr. Hall’s leadership, AWS has since released multiple ML products

10

11

12

13 .

14 13. Mr. Hall’s team constantly worked on pushing my team to innovate beyond

15 what was available commercially and made it a priority to stay ahead of Google in marketing

16 efforts. During the course of Mr. Hall’s work on marketing plans for these products, he learned

17 large amounts of highly confidential and proprietary information regarding the products and

18 associated marketing strategies. I describe three examples of these efforts, set to launch this

19 year and into 2021, below. These are just are three examples out of more than

20 across AWS’s other product categories such as compute, storage,

21 databases, security, gaming, and others.

22 EXAMPLES OF AWS ML PRODUCTS WITH WHICH


MR. HALL WAS SUBSTANTIALLY INVOLVED
23
14. Amazon SageMaker. Amazon SageMaker is AWS’s fully managed ML
24
service, released publically in November 2017. SageMaker loads customer data from AWS’s
25
data warehouses, analyzes the data automatically for processing (including numerical and
26

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 5 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
1 categorical data), and then trains a preloaded ML algorithm on the customer’s dataset in order

2 to create, evaluate, and interpret a custom ML model.

3 15. SageMaker was groundbreaking for AWS and

6 AWS CEO Andy Jassy’s last keynote for re:Invent—

7 AWS’s annual conference for all things cloud computing, Mr. Jassy devoted a half hour to

8 SageMaker.

9 16. Because of its success, many companies, including Google, attempt to imitate

10 SageMaker and to stay ahead, AWS must have a very clear vision of how to continue to

11 innovate with SageMaker.

12 Through his job responsibilities at AWS, Mr. Hall also learned

13 AWS’s roadmap for SageMaker. Attached as Exhibit F is a true and correct copy of an

14 October 28, 2019 email that both Mr. Hall and I received regarding a meeting to discuss

15 SageMaker attaching two highly confidential documents which provide a comprehensive list of

16

17 Google could easily use this information to target

18 its messaging regarding competing products to these AWS customers in order to undercut

19 AWS’s current market advantage and to steal AWS’s customers.

20 17. Most recently, one of the “stories” that Mr. Hall helped AWS develop was

21 regarding how

22

23

24
25

26

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 6 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
1 . A few examples of these unlaunched products are described

2 below.

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24
25

26

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 7 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
1

10

11

12

13

14

15

16

17
MR. HALL’S EMPLOYMENT WITH GOOGLE
18 WILL ENABLE GOOGLE TO COMPETE UNFAIRLY WITH AMAZON
19 21. Mr. Hall’s Proposed Temporary Position Working for Google Cloud Next. I

20 have reviewed Google’s proposal that Mr. Hall work on Google Cloud Next. Google Cloud

21 Next is Google’s premiere cloud product marketing event of the year and where Google makes

22 key announcements regarding its capabilities, positioning (such as what releases it has in beta

23 or pre-release and what is generally available), and targeting. My team watches Google Cloud

24 Next incredibly carefully as this conference has the potential to upstage our anticipated

25 launches and interrupt our positive press for the next year. I am especially sensitive to the

26 impact of Google Cloud Next given my previous work in product strategy and speech writing

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 8 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
1 for AWS CEO Andy Jassy. I know first-hand how a CEO’s speech can have a real and

2 concrete impact on the business due to a shift in perception and press coverage. If Mr. Hall

3 provided Google with his perspective on Google’s narratives, it would be devastating,

4 especially with regard to my business in ML.

7 .

8 22. Further, there is no way to prevent Mr. Hall from doing so. Even the most

9 conscientious of us must draw on the full extent of our knowledge and experiences in drafting,

10 revising, and editing speeches that will drive the vision of a cloud computing company. This is

11 equally true with Mr. Hall providing play-by-plays of what Google announces. Mr. Hall would

12 essentially be crafting sound bites for the press based on his judgment and determination of

13 what is important. For example, Mr. Hall knows the weaknesses and strengths of AWS’s yet to

14 be launched ML products. He could influence thought leadership (such as major customers and

15 analysts) into believing areas where AWS is anticipating launching a key product are not as

16 important as the areas where Google has an offering, knowing areas where AWS is not

17 planning a competing product launch. He has the benefit of AWS’s confidential roadmap,

18 strategy, narratives, and launch information. Mr. Hall’s editorial decisions cannot be distanced

19 from the two years Mr. Hall spent determining this strategy on behalf of AWS nor the many

20 months he specifically spent working on re:Invent, AWS’s own version of Cloud Next.

21 23. Mr. Hall’s Position at Google as VP of Product Marketing. I have also

22 reviewed Google’s letter describing Mr. Hall’s proposed role at Google as VP of Product

23 Marketing for Cloud. This role does not seem to be distinguished any meaningful ways from

24 Mr. Hall’s work at Amazon. Google proposes that Mr. Hall have immediate responsibility for

25 the narratives and marketing strategy for Google’s brand and products, working with analysts,

26 and creating public facing materials. This position competes directly with the position Mr. Hall

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 9 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
1 held at Amazon across all categories of AWS cloud products. As to ML, Mr. Hall’s

2 knowledge of AWS’s plans to advance its thought leadership and mind share as well as his

3 knowledge of AWS’s years of development toward ML product suite “stories,” and his

4 detailed, comprehensive, and highly confidential knowledge regarding over a dozen anticipated

5 launches of AWS products over the next two years would provide Mr. Hall a roadmap for how

6 to improve customers’ and analysts’ perception of Google to Amazon’s direct detriment.

7 Google could also use the information in Mr. Hall’s possession to spur Google’s development

8 of competing Google products. Similarly, Google could use this information to launch

9 competitive marketing messaging before AWS actually launches its cloud products according

10 to the cloud roadmap and marketing plans Mr. Hall developed for AWS before leaving to work

11 for Google. Again, the competition between Amazon and Google is fierce, and even the

12 slightest inference (or enhancement or deviation in a pre-drafted narrative) from a former AWS

13 Vice President who was just recently deeply immersed in AWS’s roadmap would undoubtedly

14 (and unfairly) affect a customer’s decision to migrate workloads to Google over AWS. Hall, as

15 Vice President of Product Marketing for Google, will necessarily make this his primary

16 objective.

17

18 I declare under penalty of perjury of the laws of the state of Washington that the

19 foregoing is true and correct.

20
EXECUTED at ____Seattle__________ this ___8th__ day of June, 2020
21

22

23

24

25

26

27

Davis Wright Tremaine LLP


L AW O FFICE S
920 Fifth Avenue, Suite 3300
WOOD DECLARATION - 10 Seattle, WA 98104-1610
(206) 622-3150  Fax: (206) 757-7700
DECLARATION OF MATT WOOD
EXHIBIT A FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT B FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT C FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT D FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT E FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT F FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT G FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT H FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT I FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT J FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT K FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT L FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT M FILED UNDER SEAL
DECLARATION OF MATT WOOD
EXHIBIT N FILED UNDER SEAL

You might also like