Alexy Joven

head of growth · marketing engineer | b2b + prosumer · prev. founder @modjo_me

[Alexy Joven @alexyjoven]

@xandervancil29 lets build it

[Alexy Joven @alexyjoven]

@allgarbled this is meritocratically meritocratical

[Alexy Joven @alexyjoven]

god gave growth marketers 2 things

- emotions

- https://t.co/XPOeTNjY3t websets

[Alexy Joven @alexyjoven]

@TalokCapital Im into anime right now

[Alexy Joven @alexyjoven]

@startupoppa well she is hating as much on ppl doing the opposite as her

[Alexy Joven @alexyjoven]

anthropic recently shipped memory, loops, agent orchestration

is what a lot of startups spent the last 6 months building as paid services on top of the api

they now got absorbed

the next 6 months of "harness optimization" startups will get absorbed too

i do growth. i watch where the money flows on this layer

from where 6 months from now is going to be, i wouldnt be building "harness optimization"

i'd be building on top of it, not next to it

[Linas Beliūnas @linasbeliunas]

Anthropic engineer nailed it:

“You’re not supposed to prompt Claude. You’re supposed to build a system that prompts itself.”

One of the clearest breakdowns of real agent systems I’ve seen.

In just 30 minutes, Anthropic’s Member of Technical Staff breaks down how to build agents that actually improve:

→ The evolution from basic CLAUDE .md files to agents writing their own memory
→ Why in-band memory hits hard limits, and splits focus at scale
→ How “dreaming” lets agents review sessions, learn from mistakes, and get smarter between runs
→ The production guardrails (versioning, concurrency) that make self-improving systems work

Coupled with the Loop Engineering Guide below, you will never prompt Claude the same again.

Instead of watching another Netflix show today, watch this talk.

[Alexy Joven @alexyjoven]

@tclarkmedia amen

[Alexy Joven @alexyjoven]

@evanbuhler peligro

[Alexy Joven @alexyjoven]

a client tried the "wait until they threaten" cash-flow play on an unpaid invoice

i fed claude the contract, the jurisdiction, and the timeline

out came a demand letter with the right code sections, the right terminology, the right leverage

checked it against primary sources. sent it

they paid same day

solo operators now have $500/hr counsel on tap

[Alexy Joven @alexyjoven]

@iFetis amen

[Alexy Joven @alexyjoven]

if you still confuse reach for influence
you will lose

the clickbaits, lies, optimizing for reach
you will lose

in 4 years when we'll all ask our AI to order food, find a growth person or a plane ticket

social media blockchain indexing will filter for relevance
not for your reach - i even think you'll be penalized

[Gary Vaynerchuk @garyvee]

https://t.co/f8pT7OVN0O

[Alexy Joven @alexyjoven]

@lukesophinos yep
curious ppl buy
most dont care yet

what im curious about is the value actually received/subscriptions being paid ratio

i have 20 subscription going on, i barely use them, when i need them

[Alexy Joven @alexyjoven]

growth/marketing engineers are so rare these days

how to become one:

work for a decade without shortcuts

content
SMM
ads
psychology
sales
api / mcp
tooling
hire / fire
bootstrap 0 to X (channel discovery mechanism)
optimize M$ budget at 0.0%
community
ops engineering
team building
design
ontology
ai
b2b/b2c
ugc
influencers
signals / intent
data + attribution
email + deliverability
CRM ops
brand + voice / positioning
SEO

took me 10 years. still learning

[Tommy @Shaughnessy119]

Palantir is really banging the drum on enterprises owning their AI stack

The main takeaways from their report is their view on models as commoditized intelligence and interchangeable inputs to the control layer above.

Palantir’s argument is that enterprises should not let OpenAI/Anthropic/etc own the compounding loop. They should own the ontology, workflows, permissions, audit logs, routing, evals, and context flywheel, while swapping models based on cost, latency, and quality

- Model liquidity: enterprises need the ability to switch between Claude/GPT/Llama/Mistral/etc with minimal friction

- Cheaper models slot in wherever quality holds: exactly the AI cost deflation point

- Control layer captures value: workflows, ontology, agents, permissions, auditability, context, and institutional knowledge persist while models rotate

- Misaligned incentives: model providers want tribal knowledge to flow into their weights/API surface; enterprises need a layer that keeps the flywheel internal

- Context flywheel > model flywheel: the durable asset is not the model, it is the structured record of actions, decisions, workflows, and permissions.

[Alexy Joven @alexyjoven]

nfts will be back

if i can track my best customers through on-chain collectibles and target/retarget through these datapoints + (onchain) lookalikes

i can share with them the right content at the right time

[TBPN @tbpn]

.@garyvee: "I've been in a 7-10 week cocoon around collecting becoming the next lifestyle genre."

"I think it's about to completely blow the roof off."

"I believe in 10 years, every Fortune 500 consumable product — apparel, CPG — will have a collectibles marketing strategy."

"You saw lines around the door at Popeye's for the One Piece collectible. People flooding to a Dodgers game to get the One Piece collectible."

"Differentiation in an AEO/GEO, agentic commerce world around collectibles and community to get people to stop their normal pattern of what they're just reordering — because they want to buy a deodorant that has a collectible of their favorite [entertainment thing]."

[Alexy Joven @alexyjoven]

reading "Taste for Makers" from Paul Graham in 2022

explaining taste, history, art, technology

i don't think he would bullish at all on any 'ai to replace taste'

nor content, design or art - nor anything related to it

[Alexy Joven @alexyjoven]

@DAcemogluMIT bc he does not have 1T in cash and that allows him to help more than 1 million families have food every day (his direct and indirect employees)

[Alexy Joven @alexyjoven]

- a 30% chance of making $3M is as good as a 3% chance of making $30M is as good as a 0.3% chance at making $300M.

- The distortion is exacerbated by the fact that people building profitable companies outside the sphere of the VC dominion have little systemic need to tell their story. VCs, on the other hand, needs the continuous PR campaign to meet their recruiting goals. They can’t just bag a single win and be content henceforth.

- There’s an incredible connection possible when you align your financial motivations with the service of your users. It’s an entirely different category of work than if you’re simply trying to capture eyeballs and sell their attention, privacy, and dignity in bulk to the highest bidder.

- It feels like honest work. Simple, honest work. I make a good product, you pay me good money for it.

- The most satisfying working relationships I’ve enjoyed in my close to two decades work in the internet business have been those that lasted the longest.

- Independence isn’t missed until its gone. And when it’s gone, in the sense of having money masters dictate YOUR INCREDIBLE JOURNEY, it’s gone in the vast majority of cases.

- It meant slowly growing an audience, rather than attempting to buy it, in order to have someone to sell to."

Reconsider - https://t.co/WhtwtqwcwX

[Alexy Joven @alexyjoven]

@dara_venture free/cheaper work is never taken seriously into account to get enough accountability from the team - therefore you dont get the time, agency nor ressources to leverage your work - its psychologic