The future belongs to idea guys who can just do things

There, I said it. I seriously can't see a path forward where the majority of software engineers are doing artisanal hand-crafted commits by as soon as the end of 2026. If you are a software engineer and were considering taking a gap year/holiday this year it would be an incredibly bad decision/time to do it.

I'm no longer hiring junior or even mid-level software engineers.

Our tokens per codebase: Gumroad: 2M Flexile: 800K Helper: 500K Iffy: 200K Shortest: 100K Both Claude 3.5 Sonnet and o3-mini have context windows of 200K tokens, meaning they can now write 100% of our Iffy and Shortest code if prompted well.

It won’t be long until AI will be writing all the code for Helper, Flexile, and Gumroad.

Our new process: 1. Sit and chat about what we need to build, doing research with Deep Research as we go. 2. Have AI record everything and turn it into a spec. 3. Clean up the spec, adding any design requirements / other nuances. 4. Have Devin code it up. 5. QA, merge, (auto-)deploy to prod.
- Sahil Lavingia (Founder of Gumroad)

It's been a good 43 years of software development where everything has remained roughly the same but it's time to go up another layer of abstraction as we have in the past - from hand rolling assembler to higher level compilers to programming LLMs. It's now critical for engineers to embrace these new tools and for companies to accelerate their employees "time to oh-fuck" moment.

Companies, look at the roadmap you have carefully developed and consider throwing out parts that no longer make sense. Start motions towards up-skilling how your employees think. There's going to be a-lot of scared people out there right now within the ranks. It's critical to steer people past the emotional "deer in the headlights" phase of "oh shit, do I have a job in the future?".

the people stages of AI adoption

• detraction/cope/disbelief - "it's not good enough/provide me with proof that AI isn't hype"
• experimental usage with LLMs
• deer in headlights/worry after discovering more and more things that it IS good at - "will I have a job? AI is going to take my job. The company is going to replace me with AI"
• concern/alarm/we need to bin our planning - "everything else we are doing right now feels just so inconsequential"
• engaged, consuming AI and starting to build using LLMs (ie. using Cursor) and evolving their thinking, trying new approaches. Realising the areas where it is not currently great at and learning how to get the right outcomes.
• engaged, realization that you can program the LLMs itself and doing it.

what companies can do

People are cut from two different cloths, those whom have failed, failed, failed and grinded their way to becoming an entrepreneur and that those whom have chosen the life of a steady predictable pay cheque. Employees are at an established company because they made a decision in life for stability. Embrace that, continue to provide stability (currency for employees) in these uncertain times as it will be key to steadying the ship and retaining talent.

Do anything you can to accelerate accelerate your employees "time to oh-fuck" moment. Consider adjusting personal expenses policies and enabling employees to personally expense LLM/AI tools - even if they are used at home.

Develop internal training that quickly shifts minds from "using LLMs as a Google replacement" to "here's how you can program LLMs to automate tasks with accuracy"

Noise, fuckery and politics. It has no place. Understand that we are now in a strange new world with these LLMs and it's going to create a brand new category of cultural problems within the hyper-competitive software profession:

• When someone figures out an 'edge' will co-workers in corporate share it freely or to they keep it to themselves so that they can game systems for "impact" (ie. perf reviews)
• Will software development become like business development (sales) where edges are closely guarded secrets?
• Will software development become less of a "team sport/activity"?

Create space, time and slack for people to experiment at work with these LLMs. I'm fortunate enough that my kiddos are now much older and thus the pace that I can put into personal learning/development is much higher than it would have been if they were younger. If your company employee age is trending younger (ie. folks have only just started pushing out babies) then you'll need create space for them to be able to grow on the job.

Ultimately, personal growth is the responsibility of an individual and a company can only (or should) do so much but if a company can support employees to increase it's amount of lotto tickets, it should.

It's hackathon (during business hours) once a month, every month time.

on the tech front here's what I've learned so far

The current generation of autonomous agents work by brute force. It's imperative that investments are made into anything that can be used as a "re-enforcement" technique:

• The stronger the type system of the language and the better the compiler errors the harder you can drive these LLMs. If software doesn't compile and the compiler provides strong primitives that compiling = success, the harder it goes.
• Tests. A failing test is a reinforcement to the LLM when they make changes that they have potentially taken a wrong path. Teams and code-bases where test coverage is high are the best positioned to capitalize on LLMs.
• Cycle times.
  • Compilation times matter and your test suite needs to run fast. The faster the inner-loop, the faster you can provide reinforcement to the LLMs and reward it with a pass or fail.
  • Your developer environment time-to-onboard for humans should be as close to zero and be completely automated.

what companies should be aware of

There is a well known successful company in Australia and as the tale goes about 10 years ago an idea was born but the two founders were turned down for funding in the early days because they didn't have a good technical co-founder. They found one and the rest is history. In 2025/2026 it's different. People with ideas+unique-insight can get concepts to market in rapid time and be less dependent on needing others expertise as the worlds knowledge is now in the palms of everyone's hands.

Technologists are still required, perhaps it's the ideas guys/gals who should be concerned as software engineers now have a path to bootstrap a concept in every white collar industry (recruiting, law, finance, finance, accounting, et al) at breakneck speed without having to find co-founders.

The huge thing that software engineers don't realize right now is – they can program the LLMs and build a "stdlib" that manufactures successful LLM outcomes. Folks are stuck thinking at a primitive level of "what if I had an AI coworker" and haven't, yet, got to the thought of..

no fam, what if you had *1000* AI coworkers that went ham on your entire issue backlog at once
- Anni Betts (Anthropic)

Why write code directly into an IDE when you can program the LLM to replace the need to find co-founders? Currently there is unbounded opportunity available right now for any proficient software engineer to become the next Garry Brewer if they capitalize in this moment of time as people still haven't had their oh fuck moment or are stuck in the "deer in the headlights" phase of doom.

If you’re a high agency person, there’s never been a better time to be alive...

Ya know that old saying ideas are cheap and execution is everything? Well it's being flipped on it's head by AI. Execution is now cheap. All that matters now is brand, distribution, ideas and retaining people who get it. The entire concept of time and delivery pace is different now.

ps. socials @ https://x.com/GeoffreyHuntley/status/1887575083008598265