Jeff Lunt Games

Indie game developer.

[Jeff Lunt Games @jeffluntgames]

Working on the CPU player, as we would have called it in the 80s.

https://t.co/XKa8WndUBc

#indiegamedev #gamedev #rts #devlog

[Jeff Lunt Games @jeffluntgames]

New King's Claim devlog!

https://t.co/rc6FT9tsEj

#devlog #gamedev #indiegamedev

[Jeff Lunt Games @jeffluntgames]

I've been doing a devlog on a browser-based RTS. Episode 12 just posted - check it out:

https://t.co/iW7qdgH0eZ

#kingsclaim #devlog #gamedev #indiegamedev

[Jeff Lunt Games @jeffluntgames]

Latest king's claim devlog

https://t.co/oXfgas16uW

#indiedev #indiegamedev

[Jeff Lunt Games @jeffluntgames]

I've started a weekly video series for my work-in-progress game, "King's Claim".

https://t.co/347AAa8dep

[Jeff Lunt Games @jeffluntgames]

Agentic coding works for games as well. Follow my new series where I live code a game from scratch in close collaboration with an agentic AI tool.

#indiedev #IndieGameDev #AgenticAI

https://t.co/q7lhfInTYW

[Jeff Lunt Games @jeffluntgames]

@thdxr Good reason to believe that things will shake out to a stable state eventually, even if bumpy along the way.

[Jeff Lunt Games @jeffluntgames]

Here's how it happened for me:

- Coding by hand for 30+ years
- Found ways to automate a lot last year
- October: last hand-written line of code
- Found the course below to take me every futher

Not sponsored. Paid for this myself. Glad I did.

https://t.co/7JKjgmPIDl

[Jeff Lunt Games @jeffluntgames]

I agree with this. If you know what you want upfront, the agent turns into a big code completion exercise, but that also means that if you were wrong about any of your design decisions, then the agent is going to do what you asked anyway, and then you're stuck with something that turned out to be a design flaw. You can walk this back, but it may be more work to undo a bad design, rather than work with the agent to iterate together towards a design.

I try to think of coding with an agent as a pair programming exercise - if the pair programmer has a good idea along the way that improves on my original idea, then I'd like to integrate that idea, not discard it. But if I'm overly prescriptive with what I ask the agent to do, then it will dutifully do what I ask, even if it turns out I was wrong in the first place.

[Jeff Lunt Games @jeffluntgames]

I haven't seen the stream - I'll look for a VOD - I've noticed that the more detailed of an upfront "ask" I give at the beginning of a planning session, the more the AI takes that ask as non-negotiable rather than questioning or offering alternatives.

In other words, I start my planning sessions with something more general, such as, "I'd like to build a simple load balancer, what are some techniques for doing this, and what are the pros and cons of each?" - I specify nothing about the outcome, really. Instead we interactively go from high level goal to picking an outcome, to deciding the scope, and finally writing an implementation plan. I then feed that specific plan into a new session and let it cook, but I intervene at every major implementation point and check in with the agent to see how things are shaping up. I treat the whole situation like a pair programming exercise.

[Jeff Lunt Games @jeffluntgames]

Ep. 23 of "The Wilderness and the Wellspring" - CompositionKat joins me for the 2nd time to explain how her creative journey and launch to independence is going.

https://t.co/ADYMDVQJlm

#indiebook #indiegame #Entrepreneurship

[Jeff Lunt Games @jeffluntgames]

@easyman789 I'll announce a demo as soon as it's ready. Looking forward to it!

[Jeff Lunt Games @jeffluntgames]

It's nearly the end of the year, and it's a time of change for me. On the plus side I have a new game project launching around summer 2026 (more to come). On the sad side, Stranger Sips will be shutting down soon.

Thank you all for following along, commenting, participating, and playing!

[Jeff Lunt Games @jeffluntgames]

@TheBlindLynx Same to you! May the new year bring you all the pixels your heart desires.

[Jeff Lunt Games @jeffluntgames]

@muscarian_games Always a pleasure!

[Jeff Lunt Games @jeffluntgames]

Ep. 21 of The Wilderness and the Wellspring: reflecting on one year of Stranger Sips with Sean.O of @muscarian_games

https://t.co/Dh2PdwCxQu

#strangersips #indiedev #podcast

[Sam Castree, III: Ace Attorney @IndieGameLawyer]

@ThrillaRilla369 Donkey Kong
Donkey Kong Jr.
Zelda
Zelda 2
Excitebike
Punch-Out!!
Duck Hunt
Popeye
Ducktales
Chip 'n Dale Rescue Rangers
Darkwing Duck
Mega Man 1-4
Final Fantasy
Castlevania 1-3
Tetris
Dr. Mario
Battletoads
Felix the Cat
A Boy and His Blob
Double Dragon 1-3
Paperboy
Batman
Contra

[Jeff Lunt Games @jeffluntgames]

This is an interesting take from a Meta developer who builds systems that need to scale, and also leverages AI coding to speed up some parts of the engineering process (assuming I'm reading this and the conversation context correctly).

If I'm following correctly, the story we sometimes hear that code good enough for high scale systems _cannot_ be produced in partnership with AI is false, at least in some cases.

[vas @vasuman]

@maceskridge Billions of users.

Meta arms developers with AI because it speeds up coding, and humans like me review and edit every step. That’s a massive efficiency gain. Why are we acting like AI only writes bad code / operates fully autonomously?