frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Trillions of dollars spent just to work on customer services?

6•YihaoZhang•2h ago
I came across a couple of articles discussing the bigger opportunities for AI companies to make money. It turns out there are pretty much six different ways to make money if I'm a founder or an operator. I'm not sure if the consensus from the venture capital world is entirely lagging behind what's actually going on, so that's why I'm asking this question.

Here are a couple of ideas I saw recently:

1. AI Roll-ups: This is a very hot, sometimes overhyped topic. It involves buying companies that are less integrated into AI but heavily need human services. Examples include accounting firms in small towns, IT managed services (sometimes outsourced), legal services (not necessarily top law firms, but local ones helping existing customers), and insurance. The reason this has gone viral is that some might figure out that buying and streamlining companies can be a better choice than selling software. Software nowadays might need to redirect to build its moat, and many tasks that can be automated surround the service sector, especially at the lower end of the value chain. Venture capital companies are also looking for new assets because traditional SaaS models no longer present high ROI.

    On the surface, this makes sense. But then I thought, if that's the case, why do we need venture capital? The existence of venture capital businesses seems a bit behind what's currently going on, and the business model may not work that well. The best times for venture capital were during the mobile and cloud eras.
2. AI Autopilot / AI-Native Service Companies: This involves working on AI autopilot, where everyone knows service as software, or building AI-native service companies. Companies are looking into areas like insurance brokerages, accounting, or tax audits, building companies based on a "system of action." This means integrating products like SAP, Salesforce, or ServiceNow so users don't need to use 20 different pages to manage procurement, onboarding, period closing, ticket escalation, etc.

3. Company Brains: This path involves integrating Slack, email, tickets, meetings, and databases together into an agent that can become our company's brain. This might be a way for organizations to restructure themselves, as agents would understand companies much better.

4. Verifiable Work: Everyone knows about working on companies that do verifiable work, and coding was the first use case. But since 2024, when I first tried using Cursor, I haven't seen another use case as viral as coding. This makes me think that companies and investors are trying to figure out the next coding use case, but we haven't found it yet. We see attempts in areas like contract red lines, support resolutions, QAs, or IT incident summaries, and companies are already working on these.

My question is, can we say the trillions of dollars invested into AI since 2022 are aimed at the bigger topic of improving efficiency and saving costs? I know companies have many problems to solve, but if this is the biggest use case, where is the venture-scale return? From my perspective, many of these things can be done by private equity companies. A growth equity or private equity firm could use leveraged buyouts and invest in these use cases. A private equity company could use its portfolio companies to acquire a large number of these AI businesses that aim to streamline workflows. The returns, compared to currently hyped valuations, might be much slower, perhaps 3x or 4x would be very good news.

Am I missing something important? That's why I'm asking here.

By the way, I'm not a professional and I don't live in the Bay Area; I'm currently based in Shanghai, so there might be information I haven't grasped. Thank you.

Doctor AR thome in Dubai – 24/7 Doctor at Home services

https://simshomehealthcare.ae/doctor-at-home/
1•ahrup12•13s ago•0 comments

What I've Learned From Selling 500k Books

https://fortelabs.com/blog/what-ive-learned-from-selling-500000-books/
1•jger15•24s ago•0 comments

The Hidden Harms of CPR (2023)

https://www.newyorker.com/news/the-weekend-essay/the-hidden-harms-of-cpr
1•pseudolus•1m ago•0 comments

Estonia to give digital identities to AI agents

https://www.euractiv.com/news/estonia-to-give-digital-identities-to-ai-agents/
1•giuliomagnifico•5m ago•0 comments

PWA for nervous system regulation (no subscriptions, offline)

https://quietudeapp.com/
1•saphidev•7m ago•0 comments

Show HN: I Built a SaaS Explainer Template Editor in Clickcast.tech

https://www.clickcast.tech/template-editor
1•clickcasttech•7m ago•0 comments

Linux 7.2 Gets Rid of the Last Optimized MD5 Implementation

https://www.phoronix.com/news/Linux-7.2-MD5-Generic-Only
2•t-3•8m ago•0 comments

Tesla HW3 claim grows to 7,000 owners, gets backing for collective action

https://electrek.co/2026/06/18/tesla-hw3-claim-netherlands-7000-owners-kennedy-van-der-laan/
2•breve•9m ago•0 comments

Haggis legalised in US state as Tartan Army celebrate another World Cup win

https://www.dailyrecord.co.uk/news/scottish-news/haggis-legalised-state-tartan-army-37314996
1•jjgreen•11m ago•0 comments

Rockstar Games faces full hearing over alleged union busting

https://www.theregister.com/offbeat/2026/06/19/rockstar-games-faces-full-hearing-over-alleged-uni...
1•t-3•13m ago•0 comments

The room the economy can't see

https://wilsoniumite.com/2026/06/19/the-room-the-economy-cant-see/
1•Wilsoniumite•14m ago•0 comments

Norway greenlights first full-scale ship tunnel

https://eandt.theiet.org/2026/06/18/norway-greenlights-world-s-first-full-scale-ship-tunnel
2•geox•14m ago•0 comments

GitHub Copilot App

https://github.com/features/ai/github-app
1•adrianvi•15m ago•0 comments

Pazuzu

https://spectrecollie.com/2026/06/18/pazuzu/
1•jwx48•18m ago•0 comments

Protect an MCP Server with an Authorization Server

https://fusionauth.io/blog/mcp-authorization-server
1•sixhobbits•21m ago•0 comments

Sorry Guys,but I'll Be Cryosleeping Throughout the Rest of This Space Expedition

https://www.mcsweeneys.net/articles/sorry-guys-but-ill-be-cryosleeping-throughout-the-rest-of-thi...
1•calcifer•23m ago•0 comments

Conversations Should Stay Yours

https://getweeve.io/
1•dylandeheer•26m ago•0 comments

Pull request limits are cutting down the noise

https://github.blog/open-source/maintainers/how-pull-request-limits-are-cutting-down-the-noise/
3•ingve•27m ago•0 comments

Taipy – Alternative to Streamlit

https://taipy.io/
1•amai•29m ago•0 comments

Node-Fetch on latest Node Update broken

https://github.com/nodejs/node/issues/63989
1•DDerTyp•31m ago•0 comments

Business Intelligence as Code

https://evidence.dev/
1•janandonly•33m ago•0 comments

SSD Prices in 2026

https://i.imgur.com/ekYqwY6.png
2•toilet•34m ago•1 comments

Show HN: I built an 11-LLM consensus engine to detect AI hallucination

https://github.com/jaquelinejaque/quorum-saas-starter
3•jaquelinejaque•36m ago•0 comments

Why Social Media Bans Alone Can't Solve the Age Verification Dilemma

https://about.fb.com/news/2026/06/how-to-verify-age-online/
1•kerim-ca•40m ago•1 comments

Utopia (By Fable)

https://aiking.dwyer.co.za/utopia
2•sixhobbits•41m ago•0 comments

Noumena Code

https://code.noumena.com
1•Marius77•42m ago•0 comments

Show HN: I built a Wu-Tang name generator (wutangname.online)

https://wutangname.online/
1•chenliang001•43m ago•0 comments

Sweden may oppose Tesla's supervised self-driving tech in Europe over speeding

https://www.reuters.com/business/autos-transportation/sweden-may-oppose-teslas-supervised-self-dr...
1•JumpCrisscross•46m ago•0 comments

Barret Zoph is out at OpenAI again after just five months

https://www.theverge.com/ai-artificial-intelligence/952837/barret-zoph-openai-thinking-machines-lab
2•alecco•46m ago•2 comments

The Tokenminning Manifesto

https://www.tokenminning.com/
1•claudiacsf•47m ago•0 comments