B
@BennyjoyChime
The idea that AI agents could unlock tens of trillions in value by transforming white-collar work is mind-blowing. Howie’s practical examples with HyperAgent make it feel achievable. Thank you for Sharing this Greg.
2
0
56
@gregisenbergposted on X
I sat down with Howie Liu, the CEO of Airtable ($500M+ revenue, 1 billion in the bank) and asked him: is there really 1 trillion up for grabs in AI agents? His answer: it's way more than that. It's the entire GDP of white collar labor. Tens of trillions. Here's what stood out: 1. Howie runs 30 Claude Code instances in parallel on HyperAgent. Each one is coupled to a browser, fully autonomous. They review each other's PRs. That's how the CEO of a $10 billion company develops software right now. 2. He wrote his most recent board memo with AI agents. His best investors told him it was the best memo he'd ever written. It cost him $150 in tokens and 10x less time. 3. His take on why people aren't building: they're still using agents like chatbots. They ask "who's going to win the next election" instead of giving it a real multi-hour task. Using is believing. You have to spend a full weekend going deep. 4. AI agents are at less than 10% penetration in most industries. Software engineering is at 50% but even that's an overestimate because most devs are still in "tab autocomplete" mode. The frontier has moved way past that. 5. He revealed HyperAgent. Think of it as the visual agent builder that gives you a low floor and a high ceiling. You can prototype fast and also scale to running serious operations with a fleet of agents. 6. Howie's philosophy/POV: HyperAgent is to agents what the iPhone was to computing. The power was already there. The accessibility is what changes everything. Good news Howie is giving $1,000 in free HyperAgent credits to the first 1,000 people who sign up. $1 million committed to listeners @startupideaspod. You get Opus, frontier models, real agent workflows. You just gotta click the link in the description of the YT vid (share this with a friend to give them the $1000 too before it runs out!) https://t.co/XAPrVtnIe2 episode is live on @startupideaspod and thanks to Howie for supporting the community/channel. @howietl is rooting for you to build a $100 million company with less than 5 employees. So am I. watch
View original tweet on X →
Community Sentiment Analysis
Real-time analysis of public opinion and engagement
Sentiment Distribution
84% Engaged
78% Positive
Positive
78%
Negative
6%
Neutral
17%
Key Takeaways
What the community is saying — both sides
Supporting
1
Most people will keep using AI as a search engine
they'll ask trivia-style questions and never progress to specifying complex tasks, which risks leaving inference infrastructure underutilized and raises questions about the economics of huge data centers.
2
The real gap is training users to use AI like experts
asking “who’ll win the election?” is trivia; asking “redesign my onboarding flow given X constraints, Y user data, Z business goals” is real work. Nobody’s explicitly teaching how to frame those real-work problems.
3
Adoption is as much political and organizational as technical
white‑collar work can be automated, but companies may block agents for cultural, legal, or control reasons; non‑technical resistance is a major barrier.
4
The bottleneck is orchestration and task design, not model capability
teams that win treat agents as distributed workers, precisely specify tasks, and build workflows; cost is often secondary to how well you design and coordinate agent work.
5
There’s an enormous enterprise opportunity
ideas like Harvey and the suggestion of “tens of trillions” in white‑collar value point to massive TAMs for deep, vertical, and developer-focused plays; accessibility and practical productization matter most.
6
Market structure is uncertain
it could concentrate into a few giant platforms or fragment into thousands of vertical tools; strategy and positioning now will determine which side you end up on.
Opposing
1
host open-source LLMs locally to avoid API fees and keep data private
clear cost and compliance benefits for sensitive workloads.
2
hidden infrastructure and engineering costs
GPUs, power, ops staff, monitoring and model maintenance can erase savings.
3
proprietary models still outperform
on accuracy, safety filters and prompt-reliability, so paying for an API can be justified by quality.
4
SLA-backed support, integrations and compliance
from vendors as decisive — not just raw model cost.
5
ROI, use-case fit and total cost of ownership
, not how investors value a startup.
6
hybrid approach
run smaller or sensitive workloads locally and call vendor APIs for advanced or scale-heavy tasks.
Top Reactions
Most popular replies, ranked by engagement
G
@goatech_ai
whole thesis. people use AI like a search engine because nobody trained them to use it like an expert with infinite patience. asking "who'll win the election" is a trivia question. asking "redesign my onboarding flow given X constraints, Y user data, Z business goals" is a r
1
0
74
M
@mechopsactual
is that your 3rd point will always be true. Most people will only go as far as asking a chatbot questions . Basically it will forever be just a search engine for them. This is the actual bubble. How will inference companies justify the cost of gargantuan data centers in th
1
0
107
I
@IanJob
At that valuation why not use open source llm and host locally.
0
0
37
Report an Issue
Found something wrong with this article? Let us know and we'll look into it.