Building army of agents: O-mega is the AI agent workforce for the autonomous enterprise | prev built one of world’s first AI agents in prod w 100k b2b users
Deep Tech Week in SF almost kicking off, here's my cal:
SUNDAY (tomorrow 23 June)
> noon: ai in motion: hardware showcase
@Studio45VR
> evening: SF Deep Tech Week Launch Party on the USS Hornet, organized by
@Andercot
MONDAY
> 5pm: TechTonic Robotics Happy Hour
@ngpcapital
>
Current stack for building multi-agent teams:
@crewAIInc
for defining agents, tasks, execution patterns and chain of command
@streamlit
for agent communication streaming to front end
@LangChainAI
for several chaining functionalities and quick agent tools
@serperapi
for
This is how I see the agent's universe, everything they are 'aware' of and what contributes to their autonomy and abilities to plan, act, reflect and collaborate
Introducing Project Sid: the first simulations of 1000+ truly autonomous agents collaborating in a virtual world, w/ emergent economy, culture, religion, and government
Humans are the only species to land the moon, because we can cooperate at a vast scale
Can AI do the same?
Young unemployed men allocate 75% of their time on leisure, mostly gaming, imagine the boom in gaming when mass exit from the workforce commences (after the AI productivity supplantation
No wonder why vc's like
@a16z
are putting their money into gaming
This will explode
This is how I explain agents and how they're different from (software) tools.
> Agents are autonomous (can act independently, without explicit instruction)
> They can be autonomous because they can reason and communicate
> Because they can reason, they can plan and act
> Because
Excited to announce O-mega, the platform to build your AI Workforce.
Agents promise to bring us something regular chatbots can’t: Action.
O-mega is built for exactly that: powering AI agents to act by knowing when and how to use tools and collaborate with each other to help
Non-technical founders don’t realise yet but they’ll be building apps from scratch themselves without their CTO soon, CTO will focus on infra, integration, managing dependencies, performance, efficiency, the more complex stuff, not as much functionality
@Dan_Jeffries1
It’s available in Cursor now, by default (no custom api key)
But for me not a huge difference yet with Claude 3.5
Curious if you have a different experience
AI with personality is one of the most exciting things, personality adds value, people think they want to talk to a computer and get exactly what they asked, but we don’t know what we want 99% of cases, personality fills in the blanks, inspires, and enrichment of our thoughts
This is an old illustration of mine but basically describes the current reasoning system of OpenAI's o1 as well, the iterative prompting mechanism is one element of what makes a system agentic (and uses it to plan, act, collaborate, reflect)
Inspired by
@karpathy
(legend!), started a new podcast called Men & Machine, didn't think two AIs chatting would now already be this interesting, already hooked
-
based on my own writing, every week new episodes
@levelsio
Lisbon/Portugal doesn't have the right governance to be the silicon valley of Europe and people who come there mostly have lifestyle businesses, not even close to SV level of ambition
The human future: A Case for Optimism
Intense, thoughtful and genius piece of work on technological and human progression.
by Melodysheep
@musicalscience
The AI ‘augmentation’ story is one of the biggest executive lies currently out there, many jobs are being lost due to AI and this will accelerate
Software engineering jobs have plummeted, early adoption of AI is high in this group, other professions will follow
A small elite
/6 Picture this, a world of abundance with so many resources no person on earth has to worry about food on the table, a roof over their heads and access to healthcare.
The current AI revolution is bringing this to us.
AI is taking more and more work out of our hands. The most
This is how I see the agent's universe, everything they are 'aware' of and what contributes to their autonomy and capabilities to plan, act, reflect and collaborate
I've been seeing some confusion in agent land on chain of command vs execution patterns, even the big ML researchers and agent platforms mix these two things up imo.
The chain of command is about the level of authorization on which agents act (almost all orgs are hierarchical).
How will AI agents establish dominance hierarchies? Will there be like an alpha agent who all other agents respect, will bottom chimp (agent) look up to top chimp (agent)?
@OpenAI
@rao2z
@OpenAI
's o1 thinks for seconds, but we aim for future versions to think for hours, days, even weeks. Inference costs will be higher, but what cost would you pay for a new cancer drug? For breakthrough batteries? For a proof of the Riemann Hypothesis? AI can be more than chatbots
/6 Picture this, a world of abundance with so many resources no person on earth has to worry about food on the table, a roof over their heads and access to healthcare.
The current AI revolution is bringing this to us.
AI is taking more and more work out of our hands. The most
Some people would worry that in a world of abundance we would degrade ourselves to lazy, tube-fed simpletons glued to our screens (remember WALL-E?), but I think people are simply too neophilic for that. We have a strong drive to stimulate our brains.
Why you shouldn’t kick robots:
- Robots learn based on video and vision
- Robots watch you kick them and their robot friends
- Robots might not like to be kicked
- Robots are registering this likely undesired behaviour
- Robots might not want to be friends with people who kick
Pedro Domingos says social media is overwhelmingly a huge force for good and if it could be improved with AI it could become the ideal engine to leverage humanity's collective intelligence
By the time people will have exited the workforce we'll on the one hand have stay at home human furniture and other the other hand space explorer hobbyists
6/ These are the main expected activities in a post-labor world:
1. Personal care and caregiving
2. General leisure time
3. Virtual worlds
4. Creative expression
5. Exploration
1. Personal care and caregiving
Personal care and caregiving will for a big part still be required in
@Replit
built an interesting developer agent with clearly all the ingredients of a truly agentic system; communication and reasoning capabilities driven by the ability to plan, act, reflect and collaborate
This is getting a lot closer to an AI employee
The AI ‘augmentation’ story is one of the biggest executive lies currently out there, many jobs are being lost due to AI and this will accelerate
Software engineering jobs have plummeted, early adoption of AI is high in this group, other professions will follow
A small elite
Three agent modi operandi:
1. On-demand - real-time chat with input / outcome, eg asking to write a blog
2. Triggered - agent actions triggered by an event, eg a payment > set of actions
3. Continuous: agents doing a task continuously, eg an SDR agent doing sales outreach
OpenAI: pip install openai and set OPENAI_API_KEY
Anthropic: yea same but s/openai/anthropic/g
Google: oh boy. ok so you have a GCP account? no? ok go set that up. and a payment method. now make a "project". SURVEY POPUP! k now gcloud auth. wait you have the gcloud CLI right–
OpenAI's Hunter Lightman says the new o1 AI model is already acting like a software engineer and authoring pull requests, and Noam Brown says everyone will know AGI has been achieved internally when they take down all their job listings
Replacing people with AI from the workforce is one of the most unpopular things to talk about, but I'm excited to replace boring jobs with AI as long as people who exit the workforce are supported