AI Developer Assistants 2023

GitHub Copilot is so last year. 

Here are 5 cutting edge emerging AI Developer Tools in 2023 to supercharge your productivity. 

1) Cody from Sourcegraph

Chat and instruct your codebase, ask nicely to write Unit Tests 

https://about.sourcegraph.com/cody

2) CodeComplete

YCombinator backed, private beta CodeComplete AI

Self hosted, fine tuned to your codebase.

https://codecomplete.ai/

3) FauxPilot – an open-source alternative to GitHub Copilot server

https://github.com/fauxpilot/fauxpilot

4) Tabby

Open-source and Self-hosted AI coding assistant

https://github.com/TabbyML/tabby

5) Copilot X

Next level GitHub Copilot with code chat, terminal interfaces, Github pull request and workflow integration. Currently in Preview

https://github.com/features/preview/copilot-x

All these provide alternatives to the existing popular tools including: Tabnine, Replit Ghostwriter, Amazon CodeWhisper, Codeium. 

Ok thanks for reading. 

This is an exciting and rapidly changing space. 

Stay tuned!

AI impact on software developer jobs

AI is predicted to disrupt 300 million jobs across major economies.

And yes, that includes software developers too.

Your career is not safe. You need to prepare now to avoid obsolescence.

Here are 12 steps to adapt and protect your career for what’s coming 👇🏻

1) Supercharge your existing coding skills with the power of AI coding assistants such as Github Copilot, CodeGPT and Cody from Sourcegraph.

Spend less time creating boilerplate and repetitive code patterns, and more time on what matters – solving real customer problems.

2) Next learn to ‘learn faster’ with ChatGPT style LLM

Use conversational style prompts for rapid prototyping, learning and experimentation.

Ask ChatGPT to explain tricky code step-by-step. Or ask ChatGPT to explain code as if explaining to a ’10 year old child’.

3) Study the OpenAI code repos

Get up to speed on the OpenAI that’s turning the world upside down.

Check out the OpenAI code repo, examples and experiment.

Importantly understand the types of problems these models are solving for.

Add these tools to your tool box.

4) Study Hugging Face transformers repo. Hugging Face

Transformers provides thousands of pre-trained models to perform tasks on text, vision, and audio. The keyword here is ‘pretrained’. You can mostly use these out-of-the-box.

More tools for your toolbox.

5) Next learn the basics of ML

Consider courses such as the free Google Cloud Machine Learning Crash course Google Cloud

Venture deeper into popular tools and libraries like TensorFlow, PyTorch.

The more you work with AI, the more valuable you’ll be in the job market.

6) Stay up-to-date with the latest emerging AI innovations

We are living through an Explosion of Intelligence. The AI space is changing incredibly rapidly.  I follow trending Twitter accounts.

7) Now for some tough love

Your ability to write code has become less valuable. AI lowers the barrier of entry. So broaden your tech skills – cyber security, cloud, AI, data science.

8) Master debugging & problem-solving

AI may be able to write code, but it still struggles with complex problem-solving. Hone your debugging skills.

This expertise will set you apart.

9) People Skills

People aren’t going away any time soon! So develop your people skills.

Collaboration, teamwork, leadership, influencing, communication skills are more essential now than ever for developers.

Build your network, support and learn from each other.

The The Full Stack is here for this.

10) Develop business acumen

Understand how the business works and crucially how technology can drive growth.

11) Learn to Adapt

Be open to change and willing to learn new skills.

12) Creativity & innovation

Think outside the box and push boundaries. As AI automates repetitive tasks, human creativity and innovation will become increasingly valuable.

That’s a wrap!

If you enjoyed this post Like and follow me for more.

Thanks

#ai #career #jobs #software #developer #technology #innovation

Learn TypeScript by building a Reddit Api Client

TypeScript is seeing increasing adoption across the developer community. According to the Stack Overflow Developer Survey 2022 TypeScript ranks 3rd as the ‘Most Wanted’ programming language by developers.

It’s time for me to get up to speed and considering TypeScript is a superset of JavaScript the learning curve should be OK for me!

The best way to learn is to do right? With that in mind I’ve started a project to build a TypeScript Reddit Api Client. It’s pretty simple. You install the npm module and you can use it to interact with the Reddit API.

Now there is an official Reddit node module which last time I looked was written with JavaScript. So why not create a TypeScript version for learning?

Variety is the spice of life.

If you are interested in learning TypeScript feel free to catch up with me here.

Find out more here

Are you ready to adopt microservice architecture? 

10 reasons you are not ready for microservices👇🏻

1. When you consider microservices to be THE solution regardless of the problem.

2. When you think a microservice architecture is just what you need for your indie hacker side project currently with no users, pre-revenue and seeking product market fit.

3. You think microservices will ‘make things faster’.

4. If you think a mircoservice architecture is ‘simpler’.

5. When you have an IT operations team focused on preventing change from upsetting the smooth running of the system and a development team whose sole purpose is changing the system. HINT: combine them via devops

6. When the deployment of a new server into production is a painstaking and manual task of provisioning, updating and patching.

7. When you don’t consider how team topologies affects the architecture and architecture affects the team topology.

8. When you think a ‘monolith’ architecture is a ‘bad thing’. Why? Just because…

9. When your deployment process is painful involving multiple handoffs, approvals, signoffs and you don’t have a simple repeatable auditable deployment process. Microservices mean more deployments compounding the problem.

10. When you are manually handling server configuration, software deployment, scaling, and incident and log management. 

And this is coming from a fan of microservices architecture.

Hope you enjoyed the post. Interested to hear your thoughts on microservice architecture.

Image credit to https://microservices.io

We are joining the NDRC Startup Accelerator

Happy to announce we at terrabyte are joining the NDRC national startup accelerator, Ireland’s national startup accelerator programme for globally ambitious tech entrepreneurs in Ireland.

Our developer showcasing platform thefullstack.network was chosen along with 6 other companies out of a total of 300 applicants.

The accelerator is run by Dogpatch Labs Startup Hub and comprises of investment plus mentoring and support. I’m particularly impressed to date by the focus on mentoring and am excited to see how we will evolve over the next few months.

https://www.irishtimes.com/business/2022/08/04/six-start-ups-chosen-for-latest-ndrc-acclerator/

Cloud Architecture of thefullstack

Recently I shared a post on thefullstack giving an overview of our cloud architecture on Google Cloud Platform. I breakdown the GCP services with which we are building thefullstack platform.

Read Thefullstack Cloud Architecture

I have several further articles in mind regarding scaling and running costs. Should be interesting if you’re into that sort of thing! 🙂

If you have any thoughts or feedback let me know!

Book Summary: The Mom Test

Drop everything, drop whatever you’re doing now and read The Mom Test


We had been discussing product startup ideas and reached that most important question: 
How can you tell if your startup product is actually needed?


The answer: You’ve got to talk to people of course. But these conversations are rich in false positives and can send you down the wrong path.

The Mom Test by Rob Fitzpatrick offers are framework for how to handle these customer conversations. 

The Mom Test by Rob Fitzpatrick : How to talk to customers and learn if your business is a good idea when everyone is lying to you

Below are the The Mom Test Rules of Thumb helping me navigate these vital conversations.

Rules of thumb

  • Customer conversations are bad by default. It’s your job to fix them. 
  • Opinions are worthless. 
  • Anything involving the future is an overly-optimistic lie. 
  • People will lie to you if it’s what they think you want to hear
  • People know what their problems, but they don’t know how to solve those problems 
  • You shooting blind until you understand their goals. 
  • Some problems don’t actually matter. 
  • Watching people do a task will show you what the problems and inefficiencies really are, not what the customer things they are. 
  • If they haven’t looked for ways for solving it already, they are not going to look for (or buy) yours. 
  • People stop lying when you ask them about money 
  • While it’s rare for someone to tell you precisely what they’ll pay, they’ll often show you what it’s worth to them. 
  • People want to help you.  Give them an excuse to do so. 
  • Compliments are the fool’s gold of customer learning: shiny, distracting and worthless. 
  • Ideas and feature requests should be understood, but not obeyed. 
  • If you’ve mentioned your idea, people will try to protect your feelings.
  • Anyone will say your idea is great if you’re annoying enough about it. 
  • The more you’re talking, the worse you’re doing. 
  • you should be terrified of at least one questions you are asking in every conversation. 
  • There’s more reliable information in a “meh” than a “Wow!” You can’t build a business on a lukewarm response. 
  • Start broad and don’t zoom in until you’ve found a strong signal, both with your whole business and with every conversation. 
  • You always need a list of your 3 big questions. 
  • Learning about your customer and their problems works better as a quick and casual chat than a long, formal meeting. 
  • If it feels like they are doing you a favour by talking to you, it’s probably too formal. 
  • Give as little information as possible about your idea while still nudging the discussion in a useful direction. 
  • “Customers” who keep being friendly but aren’t ever going to buy are a particularly dangerous source of mixed signals. 
  • If you don’t know what happens next after a product or sales meeting, the meeting was pointless. 
  • The more they’re giving up, the more serious you can take what they are saying. 
  • It’s not a real lead until you’ve given them a concrete chance to reject you. 
  • In early stage sales, the real goal is learning. Revenue is a side-effect. 
  • If it’s not a formal meeting, you don’t need to make excuse about why you’re there or even mention that you are starting a business. Just ask about their life. 
  • If it’s a topic you both care about, find an excuse to talk about it. Your idea never needs to enter the equation and you’ll both enjoy the chat. 
  • Kevin Bacon’s 7 degrees of separation applies to customer conversations. You can find anyone you need to if you ask a couple of times.  
  • Keep having conversations until you stop hearing new stuff. 
  • If you aren’t finding consistent problems and goals, you don’t have a specific enough customer segment. 
  • Good customer segments are a who-where pair. If you don’t know where to go to find your customers, keep slicing your segment into smaller pieces until you do. 
  • If you don’t know what you are trying to learn, you shouldn’t bother having the conversation. 
  • Notes are useless if you don’t look them up. 
  • Go build your dang company already. 
  • It’s going to be okay.  

Book Review: The Cold Start Problem

Recently I finished listening to the book The Cold Start Problem.

This couldn’t have come at a better time for us as we continue to build to our pro developer network thefullstack

The Cold Start Problem details the principles and tactics that have helped massively successful products such as Uber, AirBnB and Dropbox to grow via network effects.

2 key takeaways for me are the importance of building ‘Atomic Networks’ and also measuring product network effects via the ‘Growth Accounting Equation’.

Meet The Full Stack

I am excited to share a new project I am involved with called The Full Stack and I would really appreciate your valuable input and feedback.

Meet The Full Stack – The professional network built for Software Engineers. https://thefullstack.network

The Full Stack is a platform designed for software engineering that I wished had existed throughout my career and to this very day. We’re hoping this platform helps engineers, managers and teams on their journey:

  • Showcase you and your projects on a professional network specifically for Software Engineering.
  • Connect with like minded engineers and build a more meaningful network (no spam!).
  • Don’t let your work get overlooked, instead let it connect you to opportunities from Teams and Engineering Managers looking for engineers just like you.
  • Unlike other networks liked LinkedIn, when “real” engineering hiring managers reach out to you on The Full Stack, you get paid into your wallet for reading and responding to opportunities, putting you in control.

The platform is built in Next.js, Tailwind, Spring, NodeJS, MongoDB running on Google Cloud Platform.

I’d really appreciate your feedback and thoughts. Reach out!

Check out https://thefullstack.network/.