```yaml
---
slug: tycho
published: 2024-01-09T23:22:00-07:00
description: A Little Prince, lost in the era of diffusion models.
changelog: ~
---
```

Tycho was my first cat. There were cats before him of which I was party to their excellent time on this planet, but Tycho was my first cat. I saw him at the emergency vet. I'm afraid it might be my last time. My phone is on me. I dread the inevitable call. It's cancer. Again. It's just a matter of time now.

That this one year old lovable Maine Coon made it to month five and a half of his chemo is a testament to his stubbornness. I'd like to think he complained loudly at the cancer until it went away.

Sucks the cancer collapsed his trachea.

I'm going to wake up with wracking sobs.

I was working on a "family photo" in DALL-E. Not because I wanted the output, but because I wanted to test the limitations of AI diffusion models. What did they "understand" and what shapes output more than others. To get good output, do I forever need to tag myself as a "German-American out of shape Bruce Willis?" Could the AI ever make Tycho an all-black Maine Coon mix without slipping that smoke color into his ruff?

Now, do I want to try?

Previously, I was crafting a moment. A moment that could happen. A photo that might have been. Now it's a photo that never will be. I could keep refining the prompt, escaping the reality for a fabrication, never letting go. He could fight a dragon. He could be human-sized. He could be a lot of things, but only in my mind.

Now, only in my mind.

It's probably not healthy to want to create for a pet you cannot cherish in the present anymore. You move from an expression of ideas to memorializing really quick. And it's an easy escape. I feel it already. I want the AI to make up scenes. I want us playing with the cat wand again. I crave seeing him standing on my chest until I wake up, judging my sleeping in. I'm going to miss tomorrow; the moment every morning he senses me finally stirring and slides up the bed to be little spoon and start his day proper. Moments as I remember them.

I have my photos. But making more feels wrong. They're a poor substitute for my love.

But he's forever my little prince. This is as the AI last imagined him.

![The family portrait project, envisioned by DALL-E, 2024](https://github.com/jakobo/codedrift/assets/1795/f781653e-9c9e-4424-a398-1f5d5046ed73)



A Little Prince, lost in the era of diffusion models.

Document

A Eulogy for a Pet Where LLMs Persist

```yaml
---
slug: too-many-males
published: 2015-09-15T00:00:00.000Z
description: Women are 50% of the workforce, but only 15% of LinkedIn's new spotlight
changelog:
  - on: 2016-09-01
    note: One year later, there is significantly broader representation in LinkedIn's Next Wave program. We're not at equal representation, but it's at least trending in the correct direction.
---
```

To much fanfare and coverage, LinkedIn [recently announced](https://www.linkedin.com/pulse/meet-linkedin-next-wave-150-professionals-35-under-who-daniel-roth?trk=ranking-overview-f-link) their Next Wave influencer program. Browsing the list of people selected, I started feeling uneasy with the number of men who were also listed as CEOs. Back in 2014, Harvard Business Review [put out a fact sheet](https://www.americanprogress.org/issues/women/report/2014/03/07/85457/fact-sheet-the-womens-leadership-gap/) showing only 14.6% of executive roles are held by women despite being more than 50% of the workforce.

Putting the 150 new influencers into a spreadsheet, 23 of them were women holding executive roles. So let's talk about that

## An Analysis of LinkedIn's Next Wave

LinkedIn's Next Wave is a quarterly list of professionals that the company wants to highlight. Many hold leadership roles in their companies, although there are also contributors, activists, and other individuals who make up the total collection.

Digging deeper into the numbers doesn't offer any additional comfort.

- 66% of the C-level writers selected are men. Over 60% of all leadership positions highlighted in the Next Wave are men. **Law** has seven highlighted men, all of whom are C*O individuals. Women in Law are represented by a single CEO and one Partner. Despite having a an even split in the **Retail** and **Fashion** category, four male CEOs were highlighted while only one female CEO was. Almost half of the categories presented were 70% male.
- A bright spot in the data is the **Energy** category. Not only are the number of highlighted individuals balanced, but there is a much better 5:4 mix of men to women in leadership positions. Furthermore, the majority of the C-level positions highlighted in the Energy category are women.
- But when you put it all together, 23 of 150 individuals are women in a C-level position. 15% of LinkedIn's Next Wave are women in an executive role.

## What Went Wrong?

Unfortunately, without any insight into LinkedIn's editorial process, we can only speculate what happened. However, the blog article indicates through tidbits around MBAs and volunteer work that some amount of aggregate analysis was done. While it's possible nobody spoke up, it's also possible LinkedIn never really asked.

It's unfortunate this is the long shadow given the company's [recent diversity efforts](http://blog.linkedin.com/2015/06/08/linkedins-2015-workforce-diversity/). Coming up short on something billed to "highlight different professionals and companies that are doing extraordinary work" feels like people weren't paying attention.

The *next* Next Wave needs to be better.

## The Next Next Wave

As the default professional network of record, LinkedIn has the ability to affect change in both their own technology industry as well as the larger economic workforce. However, the list in its current form reinforces our unconscious perception of what it means to be a thought leader (male, mostly white). Those who are underrepresented benefit the most from an effort to be more inclusive.

According to LinkedIn, these lists of individuals are curated as a quarterly exercise. While it's unrealistic to go back and remove individuals from the current list, the list itself can be improved. Here's a few suggestions.

- Basha Rubin and Mirra Levitt founded Priori Legal to help businesses find the lawyers they need and would be great additions to the **Law** category.
- **Retail** and **Fashion** could be improved by asking Ellie Williams, President of Tog+Porter to contribute.
- Rachel Holt holds the East Coast of Uber together and would add new perspectives to **Automotive**.

There are countless other diverse voices that could make LinkedIn's Influencer program better. Improving the influence program is two steps. First, share the influencer demographics. Just like how LinkedIn and many other companies share their employment numbers, a program designed to promote thought leaders should also be transparent about their program's makeup. Second, continue to improve those influencer demographics. Since LinkedIn is committed to bringing new lists of people to follow every quarter, there's 3 more opportunities this year to make these lists better and improve the overall diversity of the influencer program.

Dan Roth and team have done a fantastic job in creating the momentum and energy that defines LinkedIn's Influencer program. It's up to LinkedIn to take this energy and convert it into a force for change.

You can view the raw spreadsheet [here](https://docs.google.com/a/felocity.com/spreadsheets/d/1ZsFZrRNV8BVyfiJAjnvilf5y7J1O0oxkUxq5y-uiw74/edit?usp=sharing).


One year later, there is significantly broader representation in LinkedIn's Next Wave program. We're not at equal representation, but it's at least trending in the correct direction.

Women are 50% of the workforce, but only 15% of LinkedIn's new spotlight

Link

Too Many Males

```yaml
---
slug: scd-diversity-fund
published: 2016-10-01T00:00:00.000Z
description: CSUMB is getting high marks for minorities in STEM, but they need support
changelog:
  - on: 2021-03-01
    note: The fund's five year tenure has concluded. I've updated the post to include out findings and learnings. I still encourage you to support this university's mission to increase under represented groups in computer science.
---
```

It's no secret that [tech lacks diversity](https://gigaom.com/2014/08/21/eight-charts-that-put-tech-companies-diversity-stats-into-perspective/). For most companies, less than 20% of individuals in technical roles identify as female. Less than 20% identify as [non-white, non-asian](http://www.nytimes.com/2014/10/05/opinion/sunday/silicon-valleys-diversity-problem.html?_r=0). These are just reported categories. It should be obvious. We, Silicon Valley, have a diversity problem.

A small Cal State fifty miles south of San Jose is doing something about it. As of 2018

CSUMB accounts for 5.8% of all CS graduates in the CSU system (23 schools), but represents 9% of all graduates in the CSU system identifying as female. The university represents 7.6% of all graduates in the CSU system identifying from a Latinx background.

The university is creating opportunity in computer science at a rate higher than any other CSU in the system, thanks to the work of the School of Computing and Design (SCD). Established through alumni donations in early 2015, the SCD Diversity Fund represents a two part plan to create a more diverse and inclusive industry.

1. To encourage opportunity for students currently enrolled at CSUMB, the **Diversity Conference Program **will send individuals in good academic standing to conferences and events such as Grace Hopper, Self.Conference, and Out For Undergrad
2. Giving back to the communities that surround the university, the **Technology Ambassador Program** encourages current students and CSUMB alumni to return to their schools within the Monterey/Salinas area and inspire the next generation of students to overcome the challenges of the Digital Divide

This is a program I am excited to have a hand in both founding and funding.

## Get Involved

The SCD Diversity Fund is an ambitious five year plan. While funded through the CSUMB alumni working in Silicon Valley, the program can only reach a fraction of the university's students. To grow the effort, we are calling upon the alumni, community, and people who believe in the Vision of CSUMB to continue to develop the fund. 100% of the donations are allocated towards the university's commitment to diversity in technology.

On the [CSUMB Donation page](https://donate.csumb.edu), you can designate **SCD Diversity**. (The fund's 5-year program has concluded)

![CSUMB Logo. An Otter holding a circular frame, depicting California and the location of Monterey Bay](https://user-images.githubusercontent.com/1795/146831579-ab9945e2-ea99-4a6c-a9ab-c75a45ae4456.png)

## Fund Details

The SCD Diversity Fund is allocated in the following manner:

- 65% of contributions go towards the Diversity Conference Program
- 20% of contributions go towards the Technology Ambassador Program
- 15% of contributions enable the School of Communication & Design to pursue additional diversity efforts outside of the above two programs

You can find [CSUMB's enrollment fact sheet](https://csumb.edu/iar/enrollment-fast-facts-most-recent-term) online.

## 2017 Update

The fund is live and contributions from the Alumni community have helped make an impact in student lives and included sending over 30 students to their first conferences in 2016. To maximize the number of students exposed to a conference environment, students took busses to the Game Developer Conference in San Francisco, while the fund subsidized the admission costs.

## 2018 Update

Students underrepresented in Computer Science attended Adobe MAX in 2017, while the School of Computing and Design launching the Technology Ambassador Program in the local tri-county area. More than a dozen students and faculty visited eighteen different high schools to share their experiences and expose students to opportunities in computer science.

## 2019 Update

In addition to a return to Adobe MAX, the fund also enabled students to attend the Game Developer Conference and the International Collegiate Programing Competition as part of the 2018 conference program. The technology ambassador program visited an additional eight schools, reaching out to hundreds of underrepresented students considering careers in computer science. Finally, with expanded funding, the SCD Diversity fund also sponsored the SCD Game Jam, a 30-hour weekend-long event for CS game design students across CSUMB's diverse student body.

## 2020 Update

The School of Computing and Design continued their pattern of enabling both design and computer science students in SCD to attend conferences, while the TAP program returned to every Monterey, Santa Cruz, and San Benito county school visited in prior years. The computer science department's [revised mission](https://csumb.edu/scd/cs-mission) outlines their diversity-centric goals:

- Increase the percentage of women in the CS program to over 30%
- Increase the percentage of underrepresented students in the CS program to over 30%

## 2021 Update

A lot happened in 2020. With the fund's focus on enabling success for underrepresented groups in computer science, it made sense to fund students' at-home success through technology. This meant webcams, modems and routers, and even funding repairs of student equipment when necessary. The digital divide was on display more than ever during the pandemic, and it required incredible effort to help students succeed. The success is best measured in retention; funds enabled more than a dozen students to stay enrolled in an environment where the pandemic would have put a stop to their education goals.


The fund's five year tenure has concluded. I've updated the post to include out findings and learnings. I still encourage you to support this university's mission to increase under represented groups in computer science.

CSUMB is getting high marks for minorities in STEM, but they need support

strong

SCD Diversity Fund (concluded)

```yaml
---
slug: in-memoriam
published: 2014-06-11T00:00:00.000Z
description: For Eric, For Kat, For the Meyerses
changelog:
  - on: 2020-02-23
    note: This was rescued from the archives of the old [felocity.org](https://web.archive.org/web/20050507121849/http://www.felocity.org/journal). As a result, many links in this piece were left pointing to archive.org destinations.
  - on: 2020-02-24
    note: Also, fuck cancer.
theme:
  primary: "#663399"
---
```

I have conversed with Eric Meyer twice in my life. The first time was an oddball question on his reset.css file. The second was in person at my first ever An Event Apart. In only two conversations, the haze surrounding CSS and specificity was removed and I finally saw a beautiful web. He doesn't know it, but those two conversations were when my career moved away from server code in favor of the browser. For many of us in the web industry, Eric is a mentor and an inspiration.

Since the middle of last year, Eric and his wife Kat have shared with the world [a parents' nightmare: their battle with cancer](https://web.archive.org/web/20141009223744/http://meyerweb.com/eric/thoughts/category/personal/cancer). Their second child Rebecca was diagnosed in August with grade 3 (anaplastic) astrocytoma. Day by day we watched, hoped, and prayed for their daughter to win a battle that wasn't of her choosing.

Rebecca Alison Meyer passed away June 7, 2014. She was six years, eleven and a half hours old.

On June 12 and days after, [the internet will be awash in purple](https://web.archive.org/web/20141009223744/http://www.zeldman.com/2014/06/10/the-color-purple/). The color #663399 to be exact. For those of us who the Meyer family has touched deeply, it is our symbol of mourning. This page, all the tweets, and all that will be said in person and online cannot begin express the collective sadness we all share. Eric and family, may the pages, the blogs, [the tweets](https://web.archive.org/web/20141009223744/https://twitter.com/search?f=realtime&amp;q=%23663399Becca), and the comments all find you. I wish the best for you all as you make it through the the days to come.

---

[The family requests charitable donations](https://web.archive.org/web/20141009223744/http://meyerweb.com/eric/thoughts/2014/06/09/in-memoriam-2/) be made in Rebecca's name to the Philadelphia Ronald McDonald House or the St. Baldrick's Foundation.


This was rescued from the archives of the old [felocity.org](https://web.archive.org/web/20050507121849/http://www.felocity.org/journal). As a result, many links in this piece were left pointing to archive.org destinations.

In Memoriam

```yaml
---
slug: blind-references
published: 2017-12-09T00:00:00.000Z
description: or... Don't ask random strangers for their "back channel" opinion
changelog: ~
---
```

When faced with a difficult hiring decision, the least productive thing you can do is begin reaching out to your network in search of an "unvarnished" view of a candidate. These blind, or off-list, reference checks are both useless and dangerous.

*"He always left early to do stuff with his kids."*

*"Sally's bossy. It was always 'that time' for her."*

*"Jon was easier to work with once he started taking stuff for his anxiety."*

## The Danger of Off-List References

Let's get the dangerous part out of the way first. In fifteen years of working with companies, not a single blind reference check was ever suggested as a way to transform a "no hire" decision into a "hire" decision. That means that every blind reference check we do is actively looking for a reason not to hire someone. The repercussions of asking strangers for negative feedback are horrifying. Take a look at the above statements again.

Those are **actual blind reference checks**, sitting in three separate ATS systems, by three different companies. Protected categories extend to all your reference checks, but your premise for blind reference checks is going to naturally unearth vendettas, grudges, and other defamatory statements.

Or worse, your blind reference will go ruin your candidate's career. Consider Greg:

> I applied to WidgetSoft back in 2011 and got to the final stage of the interview. The hiring manager saw that I was connected on LinkedIn to Danny, a coworker from a few jobs ago. What the hiring manager didn't know was Danny and my boss were frat brothers. I'm sure Danny said something, because I was fired the next day. I "already had a foot out the door," according to my boss.

What was actually said came out, and it was settled before trial. For the curious, Greg didn't take the job.

So here's a better solution. If you're going to ask for an off-list reference, it's a "no hire" decision. It's that simple. You're done. We can all go back to interviewing candidates on their merits.

Every time you get the urge to make a blind reference check, it reveals a weakness in your interviews. Fix those, and the you'll get the answers you're looking for.

## Fix Your Interview

If at any point you or anyone in your company considers reaching out to their network for a blind reference check, stop the transaction immediately and write on the nearest writable surface. Your goal is to capture the raw thoughts and motivations behind wanting to check other references.

- "They seem to be dispassionate about the product"
- "I think they are a [brilliant jerk](http://www.brendangregg.com/blog/2017-11-13/brilliant-jerks.html)"
- "I wasn't convinced they are a strong coder"

These may be gaps in your interview process. Now is the time to ask the team, "what should we ask to dig into this issue deeper next time?" With a few tweaks to an existing interview or a small adjustment to a process, you can begin addressing these gaps.

Don't be surprised if you encounter a lot of resistance the first time you ask everyone to slam on the brakes. If your company habitually checks off-list references, folks have likely become comfortable with their flawed process. Since blind reference checks aren't done to overturn a unanimous "no hire" decision, that leaves two routes to explore:

**If the team was all "hire" decisions**, then why do we need to ask more folks outside of the interview process? What is it we couldn't glean from our own process? Why don't we trust our interviewers?

**If the team was split between "hire" and "no hire" decisions**, then why wouldn't we respect our interviewers and default to "no hire"? Are we just looking for confirmation bias?

## On-List References Still Have Value

All of this doesn't mean you ignore the references provided. While off-list references have significant dangers, on-list references can provide valuable insights into how to effectively onboard, manage, and grow your employee. Know what you are looking for in your references, and ask for references who can speak to the areas you want to explore. The following three individuals are a good place to begin.

- Someone who managed the candidate directly. This person will help you learn which relationships your new employee will need to prioritize building.
- Someone who was peer to the candidate. This will help you or their direct manager ensure the new employee is in an environment that is set up for success.
- (For Managers) Someone who the candidate managed. This will help you understand both the new employee's management style and the the support they'll need to be successful.

When it comes to questions, Sean Falconer has done tremendous work with his article ["Best Questions to Ask References"](http://blog.proven.com/questions-to-ask-a-reference). Not only is it a comprehensive list of questions, but includes the all-important "why" you'd want to ask these questions in the first place.

If you find yourself needing more references than were provided, the best thing you can do is be honest with the candidate. Your honesty will yield better references, and the candidate has the opportunity to provide individuals who can actually speak to the areas you want to explore.


or... Don't ask random strangers for their "back channel" opinion

blockquote

Blind References

```yaml
---
slug: human-centered-instruction
published: 2004-12-02T00:00:00-08:00
description: When learning & development focuses on learner needs
changelog: ~
---
```

The way many companies approach Learning and Development (L&D) today is ineffective. In an attempt to bring development to the workplace, L&D and human resource teams frequently apply generic training in the form of offsite programs and seminars. These outdated methods of instruction are hurting companies. [Employees retain less than half of all instruction and training they receive](http://www.hci.org/blog/close-ld-effectiveness-gap-design-thinking), and in engineering, where I focus my training efforts, retention is even worse.

Over the years, I've found that there are three distinct reasons for low retention and failed training.

First, there is little attention paid to the relevance of materials. When materials are not rooted in the company's culture and norms, employees assume that the materials were not intended for them. Second, materials are rarely real or based on pragmatic examples. When content is abstract, it not only has lower relevance but it frequently contains no practical examples or situations to which learners can relate. Finally, instruction is not designed to be repeated. Many companies create training only to perform it once, failing to do additional training as new employees join the company.

It's [proven that when training isn't relevant, real, or repeatable, it is less effective](https://www.hrdive.com/news/effective-training-is-more-than-just-business-its-personal/514019/). Companies that have invested in modern Learning & Development strategies have discovered an alternative approach to building instruction.

That approach is called human-centered instruction, and it optimizes for a high signal-to-noise ratio, crafting content specifically to learner needs.

Human-centered instruction (HCN), developed in 2015, uses cognitive theory and adult learning principles in a three-phase approach to create high-relevance training materials. Through the three phases of **Exploration**, **Creation**, and **Integration**, HCN strives to make corporate training relevant, real, and repeatable. As shown in the following examples from LinkedIn, Nav Inc, and Google, the value HCN provides at every step of training development results in more effective instruction.

## Exploration: Creating Relevance with HCN at LinkedIn

Building relevant instruction begins with understanding both the desired outcomes for the company and the needs and goals of the learners. Conducting interviews, involving subject matter experts, and understanding company culture are necessary steps to creating relevant training that learners can relate to.
LinkedIn's Human Resources and Learning & Development teams discovered a consistent trait in their management team. In a survey of their engineers, employees pointed to their manager's inability to adapt to individuals in the team whose personalities didn't align with the manager's personal style. This behavior resulted in several instances where extroverted managers unintentionally silenced team members and employees felt less valued.

In interviewing these managers, LinkedIn discovered that leaders were unaware that their actions were problematic. LinkedIn developed a program called [ManageIn](https://learning.linkedin.com/blog/engaging-your-workforce/how-linkedin-works-to-make-its-managers-best-in-class) around this central behavior in order to educate first-time managers about their new role. As [Pat Wadors, former VP of Talent, describes the results](https://learning.linkedin.com/blog/engaging-your-workforce/how-linkedin-works-to-make-its-managers-best-in-class), managers who went into the training "were better prepared and eager to change. They recognized that diversity of thought and creating a high performing, engaged team was critical to all of our success." By asking the right questions at the outset, LinkedIn was able to craft their training around a single management behavior and create a program that was relevant to their specific context.

As the ManageIn training program grew, LinkedIn's L&D group identified another need. New engineering managers who had completed the training continued to struggle with the fundamentals of management. Focusing specifically on the engineering organization and employing the same strategies used to develop ManageIn, LinkedIn discovered that as the more general ManageIn program grew, the influx of non-engineering managers into the program was making it difficult for engineers to talk candidly about problems specific to managing engineers.

In response, LinkedIn collaborated with their engineering team and created the [Apprentice Management Program](https://engineering.linkedin.com/blog/2016/02/7-tips-for-aspiring-managers-in-engineering), a cohort-based model that focused uniquely on engineering management issues based on interviews conducted by L&D with engineering managers. Engineering managers who completed the program described it as "transformational," "the best training I've been in," and "all the stuff I wish ManageIn could have taught me."

By caring deeply about the learner's needs, LinkedIn was able to create two new training programs that really met those needs. The company saw more than an eight percent increase in how highly employees rated their engineering managers. Over time, this has directly resulted in higher retention and longer tenure among employees.

## Creation: Real Instruction with HCN at Nav

Human-centered instruction goes beyond the creation of a simple slide deck. Based on the learnings gained during the Exploration phase, a slide-based presentation (or even getting the team together for in-person training) may not be the right kind of instruction. Modern workforces are increasingly relying on "snackable" videos and putting skills into practice as a form of homework. Remote workforces make traditional instructional styles difficult. Building the right instruction requires an iterative approach.

As participants begin learning and new understandings emerge, training must evolve to meet those emerging needs. By adapting the material and the delivery methods in concert with these changes, training can remain relevant and grounded in the current company culture and values.

At Nav Inc, the company was migrating from Angular to React. To prepare junior engineers for the transition, the company created time for employees to learn the new framework through blogs, video presentations, and the online learning site Lynda.com. As the migration progressed, junior engineers continued to struggle with migrating and managing code in transition.

When I [created custom technical training modules for Nav](https://web.archive.org/web/20200108233815/https://www.leadsv.com/stories/nav), we believed from interviews and code samples that the learner problem was rooted in a lack of understanding about why the organization was moving from Angular to React. To test this theory, the first instructional module of React Ready focused on the history of React and the unique approach that the framework took to solving UI problems.

By the end of the first session of React Ready, it was clear that the problem was more complex than what was discovered in the original Exploration phase. During workshops, junior engineers struggled to create working code specifically within the Nav environment. A series of conversations and evaluations after the first training pointed to the next hurdle: proprietary code built by senior engineers to ease the transition from Angular to React.

Without an iterative approach to instruction, the second training module would have focused exclusively on extending the knowledge of the React ecosystem, and it would not have met with much success. Instead, we adapted the training.

In place of workshop instruction, individual one-to-one training was used to reinforce React fundamentals and ensure that all engineers had a shared understanding of the original framework before modification.

Once a baseline understanding was reached, the second workshop focused on the reality of the Nav codebase. Checking out code from git and applying it to their branches, junior engineers followed the same integration steps that ultimately led to the proprietary code that the senior engineers had created.

The final result, a training course called [React Ready](https://web.archive.org/web/20200108233815/https://www.leadsv.com/stories/nav), taught the theory of React and Redux, provided hands-on workshops for applying React to the existing codebase, and prepared the organization to complete their conversion to React with confidence. While slides were valuable for context, the majority of learning and education occurred in workshop pods where developers paired up to implement code and follow the path taken by senior engineering.

## Integration: Repeatable Training with HCN at Google

Instruction that ends when the training session is over holds little value for most organizations. The process of learning is evergreen, and the educator must not only work to train the organization but must also "train the trainers", increasing the number of in-house instructors available to the organization. This is the only way instruction can be sustainable, adaptable, and scalable. Scaling requires subsequent iterations of training programs, allowing them to both grow and more deeply address learners' needs.

There is perhaps no better example of scaling instruction than Google's [re:Work](https://rework.withgoogle.com/) project. Started in 2008 under the name [Project Oxygen](http://www.nytimes.com/2011/03/13/business/13hire.html), Google undertook one of the most extensive Exploration phases imaginable. The goal was to understand what engineers wanted from their managers. Using data from over 10,000 employees, Google arrived at a series of key behaviors shared by all strong Google managers.

In the Creation phase, Google followed a path similar to LinkedIn by creating their own [New Manager Program](https://www.fastcompany.com/40405205/five-things-ive-learned-as-a-new-manager-at-google). When the program launched, Google saw similar improvements to employee satisfaction, retention, and growth. Unlike LinkedIn's approach, when scaling up presented a relevancy problem to engineers, Google did something unique to Google's culture.

Instead of creating a separate training custom-tailored to the needs of the engineering organization, Google opted to equip senior employees with the skills they needed to conduct the training themselves. Provided through the company's unique [Googler to Googler](https://eleaptrainingsoftware.com/how-to-incorporate-elements-of-googles-training-and-development-into-your-business/) sessions, instructors were given latitude to adapt and evolve the company's training materials to their learners.

By combining formal and informal courses, as well as formal and informal teachers, Google was able to train many more engineering managers at a quicker rate than they could have otherwise. To scale the training further, Google formalized the tools and materials for teaching Google's unique style of management and leadership. Released as ["re:Work"](https://rework.withgoogle.com/), Google made it possible for other companies to leverage the work of Google and create their own customized training experiences.

More than a [half dozen companies](https://rework.withgoogle.com/case-studies/) have shared their re:Work stories, applying their own unique Exploration and Creation steps and leveraging Google's fantastic Integration work.

## Bringing HCN to Your Company

While the stories of human-centered instruction shared here focus on larger organizations, technology behemoths like LinkedIn and Google do not hold a monopoly on excellent training. Nav Technologies, the company I highlighted above, has 25 folks in engineering. At all sizes, getting engineering teams aligned helps companies move quickly and deliver value.

Size isn't a barrier to great training and developing talent. Any company can create training that is relevant, real, and repeatable. When training is "relevant", retention is higher. Training that is "real" will adapt with the needs of the learners and changes in the company. "Repeatable" training ensures that anything built can continue to grow, adapt, and scale to the company.

The three facets of human-centered instruction, Exploration, Creation, and Integration, map directly to the goal of creating relevant, real, and repeatable instruction. The result is a more effective way to approach Learning & Development at any company. Whether you do this with your in-house L&D team or with a consultancy that specializes in human-centered instruction, applying these principles will improve your training programs and transform your employees for the better.


When learning & development focuses on learner needs

Human-Centered Instruction

```yaml
---
slug: introducing-lead-sv
published: 2017-10-16T00:00:00.000Z
description: Helping engineering teams help themselves
changelog: ~
---
```

Over my career, I've had the fortune of meeting hundreds of unique companies in the technology space. Consistently, I've found that engineering organizations struggle with the same three problems regardless of size:

1. onboarding and training new engineers and leaders
2. performance and talent management
3. cultural "norms" that didn't align with the company

Every single engineering group wanted to invest more in the above, but fell back on the same pair of reasons for why they hadn't addressed these problems.

**There's no time:** The pressures of funding and execution has left the organization no bandwidth to develop the team or culture.

**What exists is not specialized:** Even when a company has a Learning and Development or Education team, that group lacks an understanding of the unique problems Engineers and Engineering Managers face.

My new journey began when I offered to help one company build training materials for their core in-house technology. I worked with another to do a management training cohort for their 4 newly minted managers. By the third engagement, it was clear engineering organizations wanted to solve these problems, but didn't know where to begin.

I decided to become a force of change for these companies, and founded Lead SV.

![Lead SV Logo](https://user-images.githubusercontent.com/1795/146832043-1687c945-0e3b-41b7-8de6-d10ae7db0e52.png)

If you put people first, grow them, and give them the space to flourish, engineers will build something amazing. If you do the same with your engineering leaders, you'll create something truly unique in our industry.


Helping engineering teams help themselves

Introducing Lead SV

```yaml
---
slug: pinspiration
published: 2015-05-18T00:00:00.000Z
description: Helping people accomplish something through software
changelog: ~
---
```

Inspiration is a raw material from which great things are made. I will be joining a San Francisco company that is out to make the world more inspired.

Search for ["red chair" on Google](https://www.google.com/search?tbm=isch&amp;as_q=red+chair&amp;as_epq=&amp;as_oq=&amp;as_eq=&amp;imgsz=&amp;imgar=&amp;imgc=&amp;imgcolor=&amp;imgtype=&amp;cr=&amp;as_sitesearch=&amp;safe=images&amp;as_filetype=&amp;as_rights=). Search for ["red chair" on Pinterest](https://www.pinterest.com/search/pins/?q=red%20chair). While Google through its algorithms is able to give me things which are chairs in name and undoubtedly red, something is missing. The results are hollow. By contrast, the Pinterest results are rich and full of life. There is a human quality to every single image you view on the site. Like Google, there are [red chairs](https://www.pinterest.com/pin/242490761158377365/). But unlike Google, there are also [guides to making my chairs red](https://www.pinterest.com/pin/8796161744806580/), settings in which [red chairs look stunning](https://www.pinterest.com/pin/125186064615892091/), and [chairs crafted](https://www.pinterest.com/pin/12173861462220825/) from other red things. You walk away from a Pinterest search full of ideas.

That's the magic of Pinterest.

As a web site, Pinterest's `<title>` tag says "Discover and save creative ideas." This is Pinterest as we know it today. "Discover" neatly captures the searching and browsing facets of the site, while "save" encapsulates the entire pinning ecosystem. The story is not completely told. Every Pin has an end state where a user clicks through the image and is taken off of Pinterest into a world that is significantly less inspiring.

To answer "what do I do with this Pin" is to focus the raw inspiration that Pinterest creates and channel it into action. Organically, this is already emerging through the community as [DIY Pins](https://www.pinterest.com/search/pins/?q=diy) and [tutorials](https://www.pinterest.com/search/pins/?q=tutorial). Through these, people are creating amazing things. The unspoken opportunity is that Pinterest can help facilitate this inspiration into action.

Ultimately, to "put pinners first" is to make them successful. Pinners pin because they want to do something. The ability to help them accomplish their goals rests solely with the company that inspired them to begin with.

That's a mission I feel particularly inspired by and I'm proud to be joining the team at Pinterest.


General

2024

A Eulogy for a Pet Where LLMs Persist

2017

Blind References

Introducing Lead SV

2016

SCD Diversity Fund (concluded)

2015

Too Many Males

(P)inspiration

2014

In Memoriam

2004

Human-Centered Instruction