My bets for the next 5:
1. Single or small 3-5 person teams that will takeover huge market shares of different industries due to the complete removal of bureaucracy and with the power of indefatigable agents at use.
2. Rise of service agencies with done for you services providing Results-As-A-Service type offers
3. Niche knowledge workers will rise to be super valuable either themselves or with the help of those who can find a way to extract that knowledge into usable skills.
4. Attention and Agency are the two words that will drive investors the most since those will rather become differentiators when labour (via agents) is commoditised at scale.
5. Real social proof will become super critical since mass social media scale, numbers, AI generated UGC etc boom… people will tend to ask their colleagues, a founder friend for recommendations than fall for ads. Essentially word of mouth will become the most important thing to optimise for. P.S: this has always been the best thing to optimise for in business except it takes years.
This post was hand written.
Manually writing code
Tab based code completion
Sidebar agents with small context
Full blown parallel agents with a swiss army-ish toolkit to do anything
Modelling an entire humans behaviour via skills, memory and self learning along with parallel agents.
We’ve come a long way in 5 years.
Please try @itsorcadotdev
I have built this for the exact same reason. Every new agent runs on a container and not just a worktree. So there is no port conflicts, no config setup to transfer and no clutter.
It supports all agent harnesses like CC, Codex and Cursor too. Would love your feedback.