David Arnoux

GenAI ⚫️ Circle

We are GenAI ⚫ Circle.  
An invite-only braintrust for AI users & builders.
Apply to join here: www.thegenaicircle.com

Founder

HeyArnoux

I deliver GTM / AI strategies that work for 2025 and beyond:
- Multi-agent GTM strategies 
- Micro-influencer-powered flywheels 
- Off-platform pushes 
- Community-first approaches
- Prosumer-first approaches
- Fully automated outbound + inbound
- As well as traditional playbooks (product-led growth, sales-led growth, growth loops, seo + pr pulls etc)

Get in touch.

Fractional CxO (GTM/AI Adoption)

Growth Tribe

Growth Tribe is here to empower adults to always be learning. Our solutions ensure people & organisations can acquire the digital skills they need to adapt, stay up-to-date and in the lead!

We’ve built a reputation for fast-paced, impactful & enjoyable learning, offering training & coaching in Data Science, Analytics, Artificial Intelligence, Growth & Marketing, Innovation, UX Design & Product Management. 

With campuses in 6 European countries, we have supported the learning journeys of over 16,000 adults in over 900 organisations. Our 80+ staff are dedicated to creating Europe’s best learning experiences, to empower humans to always be learning.

Chief Executive Officer

Co-Founder | Head of Growth

Twoodo - The Future Of Teamwork

Grew Twoodo Saas communication tool to 150-200 signups per day with a CAC lower than 2€. 

Twoodo: "Team collaboration for the #hashtag generation."

Twoodo helps your team organize itself using simple #hashtags and @mentions.
Save time. Communicate clearly. Execute fast.

Co-founder and Head of Growth

Blue Lemon Paris

Grew Blue Lemon from a small online e-commerce website to a €3 million+ revenue brand + stores + webshop.

★ Digital Marketing (Content MKT, SEM, SEO, PR)
★ HTML/CSS/Javascript
★ Lean Marketing
★ Lean Analytics
★ Data
★ User Acquisition, Activation, Retention, Referral, Revenue
★ Conversion Rate Optimization | A/B Testing
★ Growth, growth, growth
★ Mobile Growth

Co-Founder & Head of Growth

Maya Trading

★ Business Development
★ Sales
★ Marketing
★ International sourcing
★ Logistics

Ensure continual improvement of our clients' procurement strategies and supply chain flow.

Light consulting: I catered to a number of firms who desired consulting services on setting up a business and buying from China 

(in parallel) Planification, Creation and Implementation of the Blue Lemon Paris brand sales & marketing strategy

Owner & Marketing Director

Sanofi Aventis

★ Driving sales force excellence
★ Development and implementation of business tracking and management tools to help managers and sales teams optimize their time and efficiency
★ Ameliorate communicaiton between Sales and Marketing departments by developing a central communication tool based on sales performance

Business Development

building something autonomous
community-first @ www.thegenaicircle.com
sharing thoughts @ www.heyarnoux.com

Co-CEO

the 'consulting crash' is coming (i'd say it's already begun). i run ai transformations so i follow this space very closely. here's what's happening: consulting was always about information arbitrage. getting data others didn’t have, packaging it, and selling it back as strategy.

ai now does that better, faster, and cheaper than any analyst team ever could.

check out the full video here: https://lnkd.in/dTJttb8D

what's happening atm:

→ top firms are cutting recruiting cycles short, delaying offers, and hiring fewer grads.
→ growth forecasts have been slashed in half since last year.
→ mckinsey just did something almost unheard of: back-to-back layoffs across the firm. thousands of jobs gone in 2023 and 2024, and now another round in 2025 (not just analysts. not just consultants. support staff, design specialists, even mid-levels cut loose.)
→ firms are desperately rebranding with certifications and accreditations to prove “differentiation.”

it's not working...

for decades, consulting ran on a simple equation:
partners sell work → armies of analysts do the work → clients pay for leverage.

but clients are asking a different question now:
why pay $400k for a team of consultants when ai can do the same in hours for a fraction of the cost?

even in my practice, i make myself irrelevant within 6 months with a rag.

we all remember bcg’s own trials that showed that ai made junior analysts 30-40% more efficient right? ibm already has 160,000 consultants using ai agents daily. microsoft, google, and amazon now sell “ai transformation” as a product (openai as well)...essentially consulting without consultants.

end of the junior/medior/senior/partner pyramid? when ai replaces analyst work...like a food chain collapse... the pyramid collapses. for every partner, you don’t need 8-12 analysts anymore. you need maybe 2.

the leverage model that built the industry is now destroying it.

clients are reallocating budgets. here's the data:

→ ai job postings are up nearly 20% year-over-year and the hires are going in-house, not to consulting firms.
→ global consulting is projected to limp along at ~4% growth through 2033 = basically flat once you adjust for inflation.
→ corporate boards see ai as core strategy. they’re not outsourcing it anymore.

so what does this say? clients still want ai expertise. they just don’t want it wrapped in consulting decks.

so what happens next? here's my prediction (hope this doesn't age like milk :D )

2025: growth slows, more stealth layoffs.
2026: first major firm admits the model has to change.
2027: smaller firms start folding or consolidating.
2028: big 3 becomes big 2?
2030: traditional consulting is dead. only specialists survive.

my tip: if you’re in consulting right now, you’ve got 18 months to pivot. learn ai. join a startup. build something.

Founder of GenAI ⚫️ Circle / Humanoidz (Autonomous Venture Studio) / Fractional CxO (AI/GTM)

this is my current day job, taking teams from traditional to augmented. and it's working (10 so far). exciting times. gtm teams are quietly being transformed.
with new roles. with new agents. with new ai-augmented-architectures. what i see = 

most orgs are still in the "people execute" model:
→ marketers writing posts and hoping for reach
→ bdrs sending cold emails by hand
→ ops teams in crms/cms
→ growth teams testing channels/campaigns manually

but then something shifts.

you go from:
→ doing the work yourself
→ to copiloting with agents
→ to orchestrating a largely automated gtm engine (wip)

and you do this step by step, high-impact use case by use case, with a human in the loop for eval

here’s what that evolution looks like:

phase 1: traditional
people run (almost) everything manually
→ fragmented roles, mainly manual execution, low leverage
→ content, campaigns, emails, ops, sales outreach
→ dozens of roles, fragmented tools, constant handoffs

phase 2: ai-augmented
small enablement teams + gtm agents
→ content ai co-handles creative
→ campaign ai helps build assets, optimize campaigns and run experiments
→ sales ai automates outbound
→ analytics ai shows you what’s working

phase 3: ai-automated
you don’t need a big team. just taste + ai orchestra
→ brand, creative, outreach, retention, insights .. all run through agents
→ humans steer and improve. agents execute.
→ at some point you're suprised with what is automated (and high quality) creative, campaign, sales, retention, insights 
gtm becomes software(ish)

and yes yes the reality: 

→ it’s not a flip-the-switch change
rolling out an ai-powered gtm stack takes time: realistically first big wins in 3 months (if we have urgency + right ai task force), then 12-24 months to scale to augmented and automated... for most orgs, depending on size, culture, and infra

→ bring value super super early. every use case has a compelling answer to "how does this improve the business?" if it doesn't make money, save money, save time, or improve quality don't do it.

→ you need a strong task force/guiding coalition
not too big, 4-5 max + others to support, someone has to own the transition and prove value early

→ cybersecurity and compliance matter
every tool you bring in needs to be vetted. shadow ai usage is already a risk. orgs need governance frameworks in place

→ not everyone’s ready for it (yet)
30% of team is ready/willing. inspire + upskill the remaining 70% your team on prompts, agents and workflows

→ adoption is your bottleneck
most ai tools fail not because of tech, but because teams never adopt them. the rollout needs to be supported, not just purchased

but if you get it right? everything changes.

the new playbook:
→ you steer
→ ai runs
→ growth compounds

i see this in action now. it's quite beautiful.

gtm is becoming highly augmented/automated. step by step. and humans stick around for what they do best. 


note: Adriana Pereira was the inspiration (and more) for this visual

feel like a kid again, same as 10yrs ago.. started a youtube channel. with the cringe thumbnails. yes, with the red arrows and yellow circles. yes. sharing:
→ all the startups and tools we build from scratch
→ how we're building our multi-agent workflows (content os, po os etc..)
→ macro trends, papers, research, insights
→ ai transformation programs/projects
→ how we run agents that create, learn, and grow accounts
→ how we use data to guide product, community, and growth

along with my partner Walid Boulanouar we'll be sharing weekly

so if you're into ai agents, future of work, ai transformation programs, content systems, niche growth loops, and building weird startups from scratch... this channel is going to be fun.

started last week. 40 subscribers 🐣

check out our "ultimate 2025 ai tech stack video"

search for "heyhumanoidz"

or find the link in the comments :)

llms are making us dumber, lazier and dependent (i'm guilty!) + how to fix.
it's called cognitive offloading and it's real. and it’s measurable.
→mit media lab tracked 54 people writing essays using ai...result =  the weakest brain activity on eeg. the ai users were barely firing neurons.
→ germany: students using llms had lower cognitive effort but worse reasoning
→ meta-analysis (51 studies): ai improves performance but harms understanding
→ a 2025 study of 666 people found: more ai = worse critical thinking.

and there's more

watch the full video here 🧠 https://lnkd.in/dDhVEcVG

→ that mit study? after 4 months: consistent declines in cognitive performance, language ability, and behavioral engagement. no plateau. worse over time.

wharton studied students using chatgpt for math help
→ ai helped them during practice (+48%)
→ but when ai was taken away, they scored 17% worse than students who never used ai at all.

creativity isn’t safe either.

'science' paper: ai users wrote more “creative” stories, but across the group?
→ everyone started using the same metaphors, structures, and ideas.
→ 9 students independently named their invention the same thing: “build-a-breeze castle.”
→ wharton: 94% of ai-generated ideas shared overlapping concepts.

creativity is converging. thinking is declining. and your brain is quietly atrophying.

why? because ai rewards speed over struggle. and your brain needs the struggle.

idea for a fix" it’s called the centaur method.

pretty name to basically say " keep struggling with tasks first before throwing them at an llm"

→ you think, ai executes.
→ you plan, ai formats.
→ you solve, ai checks.

and it only works if you do the hard part first.

before prompting:
→ write your expected answer
→ challenge ai’s response
→ identify what might be wrong
→ argue with it
→ revise based on that

it’s called thinking with ai vs letting ai think for you.

the difference is massive > and visible in brain scans.

use ai at the end of a task, not the beginning.
structure work into alternating blocks: human → ai → human.
never outsource the hard part. ever.

your brain is your only real edge. same for your teams, same for your kids.

not prompt writing. not speed. not subscriptions.
actual thought. actual strategy. actual creativity.

long story short...use it or lose it.

if you’ve been sitting on an idea, waiting for “the right time,” this is the right time. we partner up and build it in 30 days.
once a quarter, we open up our venture studio to just two ideas. and this is how it works:

we're looking for insights, insider secrets, untapped business ideas. two opportunities. once selected, it's then thirty days to go from idea/concept → full live functional product.

⭐️ comment below and i’ll dm you all the details on how this works + link to apply

we’ve already done this with five partners and are opening up for a new round of partners (will do this quarterly). idea is simple...you have an insider secret / business idea and a network/audience? we build for you. we then help grow.

here's how it works:

→ you bring the idea, the insights, the "secret", the network and/or audience = the business idea
→ we score, validate and refine it with our internal engine. 
if selected...
→ we handle the build. in 30 days, you’re not staring at a prototype with flaws, you’re running a product that can actually generate revenue.

- authentication/roles/admin
- payments
- database
- api layer
- cms & blog (seo-ready)
- emails
- seo pages
- highly converting ui
- more

why?

sure we've entered the "idea era"

most ppl are still sitting on ideas far too long. tinker with cursor prototypes, wireframes, or decks. it feels safe, but nothing moves forward.

we're offering to fix that. 

we're taking a bet/risk together. so the cost is a small fraction of what you’d spend hiring an agency or trying to assemble your own dev team.

the two paths

1. build & grow
you lead the vision, we build and support.
you keep large majority share, we take a minority stake.
30-day v1 build with everything you need to launch.
we handle the heavy lifting: dashboards, auth, payments, email, ui.
you focus on acquiring customers and shaping the roadmap.

2. full venture partner
we co-found from day one.
we take a majority stake but you keep a large piece of the business.
30-day v1 build, but deeper involvement: strategy, ops, growth.
every system pre-configured: analytics, ai-enhanced ui, live integrations.
ongoing partnership beyond launch = your success is our success.

both paths get you:
- a fully-fledged product (not a prototype).
- revenue potential from day one.
- speed to market that most startups can’t match.
- agencies will quote you six figures. building a team takes 12–18 months. we - compress all of that into 30 days.

we only take two projects per quarter. every one of them becomes part of our 'family'.

if you’ve been sitting on an idea, waiting for “the right time,” this is the right time.

⭐️ comment below and i’ll dm you all the details on how this works + link to apply

you bring the spark. we bring the build.
together, we make it real.

the ai job collapse? the split economy? already happening? entry-level jobs are taking a beating. like any tech revlution, the transition period hurts,... my thoughts + research on what's happening + how to approach this moment in time

everyone’s debating whether ai will kill jobs or create more of them. the answer is both. but it's the transition period, the 'messy passage'

so both can be true at the same time
→ whole generations can get crushed in the transition.
→ capital captures the value first, workers much later.
→ new jobs don’t look like the old jobs, and they don’t appear in the same places.

atm: 
→ junior devs aren't getting hired.
→ law firms paused first-year recruiting
→ goldman replaced 30% of analysts with ai systems
→ customer support agents got 14% more productive... but wages dropped for top performers.
→ chatgpt reduced worker inequality by 14% in months... but also compressed wages.
→ ai doesn’t fire you. it just ensures you never get the job in the first place.

that’s the collapse. not layoffs. hiring freeze at the bottom.

but there’s also a weird boom.

→ mckinsey says ai adds $4.4t annually to global gdp.
→ david autor (mit) flipped his stance...now says ai will create brand new job types.
→ historical data? every major tech shift (telephones, internet, robotics) increased total employment.
→ automation from 1980–2020 killed jobs...but created 2.6x more in new categories.

so what’s really going on?

ai is replacing tasks, not jobs. but early-career workers are mostly tasks. they’re the ones getting wiped out first.

and the transition gap? 

→ history says recovery takes 15-20 years.
→ 40% of displaced workers never regain income.
→ their kids do better. but that doesn’t help you now.

and here's what i focus on in the vid:

→ ai replaces cognition, not just labor.
→ top consultants using gpt-4 got worse...they trusted it too much.
→ capital gets the upside first. labor lags behind by decades.

so what’s the move in my opinion?

→ 5.4m new businesses launched in 2023.
→ $170b deployed into 15,000 startups.
→ robots create more jobs in the long run...just different ones.

unlike the industrial revolution, you can adapt in real-time. the people who keep learning, keep re-skilling, and use ai as a multiplier will keep their jobs, and often get promoted faster. the ones who wait for the old jobs to come back won’t.

the future belongs to two groups:
→ builders, who create the new roles and companies.
→ learners, who adapt fast enough to stay valuable.

everyone else gets stuck in the transition i guess?

openai may be the riskiest company in tech right now
a common pushback is that “future tech always gets funded” and that openai will keep finding money. but there are limits. raising 40-50 billion every single year is not the same as a startup raising another round. that scale of funding has never been done before, not even close. softbank is already borrowing heavily to cover its first 10 billion. ratings agencies say its financial health is deteriorating. even giant sovereign wealth funds or big tech players can’t plug a 40 billion hole every year for the next decade. at some point, investors stop writing checks.

tried to pain the picture in simple terms:

→ openai announced a 40 billion raise. only 10 billion has actually landed. the rest depends on it turning into a for-profit by the end of 2025
→) chatgpt subscriptions bring in about 5 billion a year. the company says it will make 12.7 billion this year
→ but it will spend closer to 28 billion. that’s a 14 billion loss in a single year
→) compute costs with microsoft are 13 billion this year alone. on top of that, openai owes billions to coreweave and billions more to its own data center project, stargate
→) stargate’s first phase is estimated at 5 billion. the full build could cost 30-50 billion. openai’s share is about 19 billion over several years

how's the money flowing?
users → openai → microsoft (gpu time)
openai → coreweave → nvidia (chips) → banks (loans)
openai → oracle → crusoe (data centers)
softbank → openai → everyone above
if openai slips → the whole chain cracks

some simple numbers

• microsoft cloud bill: 13 billion in 2025
• coreweave contract: 12.9 billion over five years, starting late 2025
• stargate texas site: 5 billion for the first phase, 30 billion if fully built
• nvidia’s sales: coreweave alone bought about 8 billion in gpus last year, more than 6% of nvidia’s total revenue
• users: 20 million paying, 500 million weekly active, but openai loses money on every single one

why this feels like a pyramid?

• every company in the chain is betting on openai’s future revenue
• loans are being issued against gpus and contracts, not profits
• new data centers are built for one customer, not many
• if openai can’t pay, coreweave, oracle, and others are left holding the bill

also, capacity is already tight: altman has said several times this year that gpus are running out. free image generation overloaded systems. rollouts have been delayed. stargate is behind schedule. the gap is yuuuuge.

what could break first?? openai could run out of money or gpus. if it runs out of money, prices go up and free usage shrinks. if gpus run out, service slows and reliability drops. either way, users leave. if users leave, funding dries up. that weakens the whole chain.

the simple takeaway

the models work. but the economics don’t. the entire system depends on openai raising 40-50 billion every year, forever, and finding more gpus than anyone else 😅

not sure it beats my "do my taxes, but faster"

got an idea? we'll build it in 30 days. 
join our humanoidz partnership program. only 2 partners per quarter. more details in this video. already got a bunch of applications so pls be quick! go to our website humanoidz[.]ai/partners or dm me for more info...

Taplio

David Arnoux's Linkedin Analytics

What is David talking about?

's Best Posts (last 30 days)

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}David Arnoux's Linkedin Analytics

What is David talking about?

's Best Posts .css-qqfgvy{font-size:var(--chakra-fontSizes-md);}(last 30 days)

Want to drive more opportunities from LinkedIn?

David Arnoux's Linkedin Analytics

's Best Posts (last 30 days)