For context, I went to a top 5 college for undergrad. Later worked in Silicon Valley at a few startups, as well as big tech companies like Apple. I occasionally get asked by students for advice, so I thought I would write it down.

Don’t freak out about AI taking your job… yet

You’ve probably heard the predictions from Dario that AI will make software engineers obsolete within the next year.

You’ve maybe even seen charts like this:

but you probably haven’t seen the full context of long-term trends:

For now, just become really good at using AI in your own work, as Jensen recommends.

How to escape the catch 22 of getting your first job

Your first job may be the hardest to get, because everyone wants to hire people with experience. But how do you get that experience? Internships are probably the best way, but a lot of what you need on the job you may have to learn outside of college. So having a few side-projects in your portfolio that you’ve worked on is a good way to learn common technologies that employers may be looking for. Beyond that, here are some things you can try:

• go to your college career fairs

• go to talks by speakers of companies you’re interested in, and chat with the speakers afterwards

• e-mail alumni of your college who work at companies you’re interested in, and ask for a referral or a short call for advice

• try contributing small bug fixes or documentation to well-known open source projects, or projects published by companies you’d like to work at

Choose your own adventure

hat tip: Tim Urban

After college, there’s no longer a fixed path. You don’t just pass some tests, and move on to the next grade level. It’s not just about working hard, either: a furniture mover works plenty hard breaking his back 8-10 hours a day, but only gets paid minimum wage. A CEO may work the same hours and make >100X that. So pick carefully.

A useful framework is the Japanese concept of ikigai, which you can think of as the intersection of 3 or 4 circles in a Venn diagram:

The barbell strategy

Nicholas Nassim Taleb’s barbell strategy applies not just to investing, but your career as well once you have a spouse. If you have two working adults in your household, one of you can afford to take some risks, while the other has a stable job. I’ve seen this work in the past with startup founders, whose spouses often have stable jobs (a doctor, or a job at a large public company) while they pay themselves just a basic salary.

Optimize for growth

Given the opportunity to pick between similar roles at two companies, you will probably be better off at the company or team that’s growing faster. Growth provides more opportunities for you to be promoted, take on more responsibilities, and solves almost all problems.

Learn how to sell, because almost everything is sales

Convincing someone to join your company? That’s sales.

A doctor telling a patient that they really should get the surgery done, instead of waiting until it gets worse and risking complications? That’s sales.

Asking a girl out on a date? That’s sales.

Your boss’s boss probably didn’t go to the best college. But you know what they’re good at? Talking to people, connecting with them, managing them, and convincing them to do things.

Find a mentor, and stay in touch with them

It could be an old boss, a senior colleague, an alumnus of your college who you look up to; if there’s someone you’ve come across who you’d like to get advice from, reach out and have lunch with them. But don’t make the mistake of just doing it once - build a long-term relationship, send them regular updates, and get them invested in your success. Don’t just contact them when you need help, try to build a friendship.

Never stop learning

I don't feel that I understand cinema yet. I really don't feel that I have yet grasped the essence of cinema - Akira Kurosawa, Oscar acceptance speech

It’s unbelievable the number of colleagues I’ve met who don’t keep up with developments in the field. I knew a software engineer who was shocked that chatGPT could code, two years after it came out! You don’t necessarily need to read every new paper on arXiv, but you should go to 1-2 conferences a year, and follow a few blogs or newsletters in your field.

Share what you learn & build

The goal is not to necessarily build a massive following on social media, but to become known for your expertise in a particular area. For example, a former boss of mine got hired for a role because he had started posting technical deep dives about a topic he was interested in on his LinkedIn.

As you write, it’ll force you to think and learn about the topics you’re interested in.

Negotiate your offer, and Don't get screwed over

Use levels.fyi to know what you’re worth. Try to get competing offers. There’s no harm in asking for things - a better title, WFH benefits, a severance package.

In certain areas, employers are not allowed to ask what your current salary is.

Think like an investor; when you’re evaluating which companies to join, one hat you may want to wear is that of an investor. If you had $100k to invest in a company, would you pick the one that has the job opening you’re considering? This makes sense because that’s probably what your stock options or RSUs will be worth after a few years at the company, at a minimum. Lots of companies can give you a job, but very few will provide a good return on investment. Do your research.

Ask lots of questions:

• what are these stock options worth? what’s the strike price or liquidation preference?

• is the company profitable? how fast did revenue grow last year? (if it’s a public company, read their latest 10-Q report)

• is this role a backfill or new headcount?

• do you give annual equity refreshers?

• can I speak to some current/former members of the team?

Invest as soon as can

Once you have some disposable income, start investing it as soon as you can. A good guide is the Stanford course Personal Finance for Engineers. Stick to ETFs for the most part, unless you have really done your research on a particular company.

Forget networking; learn to host parties instead

Your count of contacts on LinkedIn doesn’t matter, as most of them are just acquaintances that you maybe met once, or are no longer in touch with. What matters is that you have a core group of friends that you can count on for support and friendship on a regular basis. Once you’re no longer in college, it becomes very hard to make new friends and keep in touch with existing ones. One thing I’ve seen that works well is to host regular (3-4 times a year) parties for your friends. Doesn’t even have to be any special occasion. Just say you’re having a get-together, have everyone bring their spouse or friend. If you don’t have much money just order some pizzas or just hold a potluck. If you don’t have a large space, rent out a room at a community center for a few hours.

Despite its documented shortcomings, MCP has now become the de-facto standard for connecting your AI to the outside world of apps, with every major developer IDE now supporting it. About two months ago, I was playing around with the Cline VS code extension and decided to try it’s “1-click” MCP server installations. I searched for “postgres” and clicked on the first result, confident that I’d be up and running within a few seconds.

A minute later, I was greeted with a $0.13 usage bill of an LLM model, a plethora of messages and plans from the installer script, and was no closer to actually installing the MCP server!

• It seems like it:

  • 1. Decided to build a Docker image.

       2. This failed less than halfway through, apparently because I didn’t have BuildKit installed on my Mac.

       3. The above error in the terminal was was not detected by the installer script, which blithely continued on and reported that “the Docker image…. has been successfully built”.

The whole experience makes the Apple App Store look like Valhalla. But OK that’s just one example, maybe that was just an exception? Let’s look at another server, one maintained by an established company. Algolia has an official MCP server. And the installation process is to…. unzip a file, remove the quarantine flag, and run the binary (that’s only supported on Mac OS)?

As an individual developer, dealing with issues like this, even for a half-dozen servers you may want to use regularly, is a bit of an inconvenience, but it’s nothing insurmountable; within a half hour or so you should be able to sort out your environment, all the dependencies, and have your servers setup. As more SaaS companies shift to hosting remote MCP servers, this installation headache becomes less an of issue, but for business with hundreds or thousands of engineers, all the enterprise complexities remain - especially for security and IT teams, who struggle with questions like:

• How do we setup a centralized, hosted MCP server for all our engineers to connect to, to make onboarding easy, and use single-sign on for auth? Oh, and it has to be deployed on-premise so we can connect to our databases, warehouses, data pipeline orchestrator, etc. which are all behind the VPN/firewall.

• How do we control which apps we allow our developers’ AIs to connect to? And restrict high-risk tools and resources of specific MCP servers?

• How do we keep an audit log of every interaction between the AI and an MCP server?

• We need an integration with (random legacy tool) that doesn’t have an MCP server, how do we do that?

Integration challenges are nothing new, and there are some standard ways of approaching things like this.

Let’s consider them one at a time:

Approach 1: Write a single MCP server with connections to every API you need

• We know of Fortune 500 companies that have tried this. It starts out well enough, with a small team of 1-2 engineers. They read the MCP spec. Add a layer of abstraction over different APIs, and start off with a few wrappers of APIs that they commonly use. It works great. But soon, the requests start coming in from the engineers: can you support Slack? What about the wiki? And the CI/CD tool? The team eventually gives up trying to support everything and just asks the end users to write their own connections following their example code, which of course no one does, because who has time for that?

• But you say, no need to reinvent the wheel. Every company is going to need this, so there should just be a few providers that setup all the MCP servers, and then the rest of us can just use their platform. This would be analogous to how ETL apps like Fivetran and Airbyte have connectors for the most common apps.

• There are a few problems with this approach:

  • Good MCP servers shouldn’t just be API wrappers, but opinionated functions that make it easier for the LLMs to reliably do their job. For example, the CircleCI API requires multiple API calls to be strung together to do something like get the logs for a recent build failure in a CI/CD pipeline. A simple API wrapper is not going to make that easy for an LLM to figure out in a reliable manner.

  • A third-party MCP server is unlikely to be ever as good, or as up-to-date as the official MCP servers that some companies will offer themselves. Sure, you can convert the Github openAPI spec into an MCP server, and filter it down to just the APIs you need. But is your company going to ship updates to it almost every day, like Github does? Are you going to fix its bugs and security vulnerabilities faster than Github itself? Highly unlikely.

  • In summary, even if there were vendors who dedicated their existence to just building and maintaining MCP servers for the most commonly used developer tools, you’d have to tradeoff the quality and reliability of those integrations for scalability.

Approach 2: Just setup a gateway and run a Docker container for each existing MCP server and employee that needs it

• If you compiled a registry of available MCP servers and their Docker images, you could setup something akin to a gateway that could spin up, on the fly, a container for each employee that needs it, for each app they’d like to connect to.

• This scales with O(m x n), where m is the number of apps per employee and n is the number of employees, which is not great.

• Furthermore, managing potentially hundreds of unique Docker images, and thousands of containers is a logistical nightmare. I’ve worked at large companies where it took weeks just to deploy a single Docker image (Kubernetes, firewall rules, Terraform configs for load balancers, deployment tickets and approvals). Not to mention ongoing security tickets and upgrade requests.

• The downsides to this approach are that it quickly becomes very expensive to scale, and doesn’t offer a complete solution, as not every tool will offer an MCP server.

Is there an alternative to the options above, that avoids each one’s disadvantages, but maintains scalability, quality, and ease of use?

• We’ve developed a federated MCP server that allows you to run and proxy any supported MCP server in a single Docker container. We support servers written in Python, Node, and Go, which accounts for the vast majority of all the servers on Github, along with any remote servers that are hosted by SaaS vendors themselves. You just specify which servers you want to enable, and we launch them on the fly when the container starts. It can run locally on a developer’s laptop, in the cloud, or on-premise for an entire organization. For enterprise customers, we also bake the requirements of each server into the Docker image itself for greater security and speed.

• For apps that don’t have their own MCP servers, we have an open source library that let’s us connect to every major database/data warehouse, and any API with an openAPI spec that can be customized to your needs. We’ll be adding a thousand+ apps within the next two months for customers who are able to leverage the public cloud.

• This offers you the best of both worlds:

  • the low cost and ease of deployment of a single Docker image for all your employees

  • support for all official MCP servers, so you’ve got the best quality, reliability, and security straight from the source

  • a scalable framework for the long tail of apps that you want to connect to