running ai-powered agencies | weeklyaiops.com
- Tweets 6,533
- Following 737
- Followers 146,848
- Likes 4,513
how to give your agents the best research in the world...
the whole setup is ONE simple skill... here's how to build it:
- web engine: Parallel for search, full-page extract, and deep research runs that return every claim with its source
- last30days: what real people said in the last 30 days across Reddit, X, YouTube, TikTok, Instagram, Hacker News and Polymarket
- YouTube: searches the videos and pulls the full transcripts, so an hour-long walkthrough becomes text my agents can quote
- x-search: live X through Grok, the fastest read on what's moving today
- Firecrawl: pulls the full text when a page fights back, anti-bot walls and X posts included
but the tools are the easy half... the rules are the research:
> a tool's summary is a lead, never proof... evidence starts when the source page is fetched as full text
> a finding needs 3 independent sources or 1 primary source, single-account hype stays marked as unconfirmed
> 3 skeptic agents attack every claim (correctness, recency, source quality) and the majority vote wins
> one sweep hunts only for people saying it's wrong
> everything lands in the vault with an expiry date, so the next run starts from what i already own
your agent is only as good as the worst source it believed
6
3
67
4,438
join 15k founders using my system, plus one email per week (it's free) weeklyaiops.com
672
Opus 5.5 is the best model launch of the year for me... it couldn't have been better
the timing was impeccable, they basically made the GPT-6 releases pointless
because this is what we're getting:
- better, faster and cheaper than Fable 5.1
- with the same feeling Fable had, the one that made it so good to work with
- cheap enough to actually run in prod
- limits that aren't that bad anymore
so a lot more people can afford it now, which genuinely opens up use cases that were out of reach before
they put OpenAI on silent mode for a few days with this one
37
6
270
16,439
hiring in 2024:
>start up raises $2M
>hires a bunch of people in different US states
>HR doesn’t know how to setup tax and payroll by state
>HR cries
hiring in 2026:
>finds top 0.1% talent from any US state
>tells Warp to setup state registration, payroll, benefits, and send employee a laptop with one prompt
>HR spends the afternoon sipping iced coffee
We’ve raised $85M for this moment.
Introducing Warp 2.0: The first AI Head of HR.
Every company is building AI to replace jobs. Warp is building AI to do the jobs no human should have to:
If you work in HR, I want you to spend time with the manager who needs help or building company culture people actually want to work at.
If you’re a founder, I want you to focus on signing clients or spending time with your family.
You shouldn’t have to figure out how to register state tax in California. You shouldn’t have to pay outrageous penalties because you don't know what a DE 9C is.
I want to make HR human again. Today, this is finally possible with the Warp Agent.
I’d love for you to see it in action: warp.co/agent
33
6,902
Jev can't write a single word, which is exactly why it should pick what my agents read
my Obsidian vaults hold 5k+ notes, one idea per note, each with a one-line summary in an index
but Claude can't read all of that on every turn, and a keyword search misses the note that makes the same point in different words
so Jev does the picking:
> step 1: Jev reads the 28 vault descriptions and picks where the task belongs
> step 2: it checks every note summary in that vault in one call, each judged alone
> step 3: the top 5 come back with full text - Jev drops any that miss the task
> step 4: my code loads the survivors into Claude's context (and nothing else)
every pick comes from a list i define, so Jev can't invent a note
the worst it can do is pick the wrong one... and each one carries a confidence score that tells my code when to fall back to Claude
14
2
115
7,567
join 15k founders using my system, plus one email per week (it's free) weeklyaiops.com
1
892
Machina retweeted
Today we’re launching micro1’s PII transformation model, flow-transform 1.0, delivering frontier-level performance across detection, identity synthesis, and transformation of personally identifiable information.
On PrivacyBench, our model reaches 96.0% F1, outperforming every detection baseline we tested, including Tonic Textual, Claude Opus 4.8, Sonnet 4.6, Microsoft Presidio, Haiku 4.5 and GLiNER2.
Some of the most valuable training data for frontier AI models lives inside fully functioning companies. It captures years of real work across decisions, communications, tools, handoffs, exceptions and the relationships connecting them.
The problem is that this data is also full of PII.
Traditional redaction makes the data safe, but it also destroys the very workflows and relationships frontier models need to learn from.
flow-transform 1.0 solves this by turning enterprise operational data into high-fidelity training data for frontier models by replacing real-world identities without flattening the reality the data captures.
85
106
686
204,181
be Alex Mashrabov
> start programming at 10
> get a Meta internship offer at 18... turn it down
> build an AI startup called AI Factory instead
> Snap buys it for a reported $166 million in 2019
> leave and co-found Higgsfield to launch the AI video platform in 2025
> get 32 million users and nearly 1 million paying subscribers
> on track for $1 BILLION+ in annualized revenue
> built one of the fastest-scaling gen AI companies in the US
and he says they're barely scratching the surface
Today, 18 months after launch, our annualized revenue crossed $1 billion.
The platform now powers organizations across the Fortune 500. Enterprise adoption has grown 10x since June.
More than 30 million people worldwide now use Higgsfield.
Thank you to the creators and teams building with us.
24
188
706
73,811
Machina retweeted
Introducing Agora-2, our next-generation multi-agent world model.
Agora-2 supports up to 20 humans and agents interacting inside a shared environment, all simulated in real time.
Our multiplayer research preview is available to try right now!
233
2,499
2,715
1,142,543
We live in a new era of “Entertainment Ads”
Majority of the top brands are doing one of these formats
>Song ads
>Drama ads
>Educational animation
They all share the same reason why they work
They are entertaining, they don’t feel like an ad, they blend into the scroll
People aren’t on their phone to buy other wise they’d be on Amazon, they’re there to pass time
Making your ad entertaining is the number one way right now to capture that initial attention
Then you need to use the copy to sell them
I made this AI Pixar movie trailer ad for a brand and I haven’t seen anyone create in this way before
14
16
87
10,450
an ex Meta guy who literally built Meta Ads just launched an agent that will run entire campaigns on autopilot…
the AI agent makes static, UGC and VSL ads with b-roll from ONE prompt
then it runs the whole campaign by itself:
- scans competitor ads from Meta ads library, pulling the winning ads in the niche before making any ads
- determines angles they are using that you aren’t
- builds and improves on the winning ads
- pushes the ads it made into Meta ad account
- collects and reviews data by the minute
- suggests of ads based on that data
- 24/7, no human input required
entire marketing teams being outperformed
try this and lemme know the results
AI can run a brand better than humans.
We ran a brand with ONLY AI agents to prove it.
Introducing Notch: the AI behind it. usenotch.ai
Paid partnership (ad)
14
43
725
175,186
Andrej Karpathy said something that really stuck with me:
"you have to ask yourself: how do i get MAXIMUM output from AI with minimum input from me, and have it run for a long time without me"
this is how everyone should be working with AI... and these guys pretty much built the GTM version of it:
- one prompt turns your browser into a GTM agent
- it goes and finds leads, researches companies, sources creators for your campaigns
- it checks inboxes and updates your spreadsheets while you do literally anything else
hours of manual GTM work, gone from one prompt
30 seconds of typing can now turn into 100 hours of GTM work.
Introducing Polar for GTM.
Give Polar your ICP and tell it what needs to get done.
It can find accounts, research companies, find decision makers, enrich leads, update sheets, check replies and prep follow-ups.
Then save the workflow and run it again. 👇
Paid partnership (ad)
11
26
395
35,677
how to make money building a clone for Higgsfield:
the easiest way to market an ai video SaaS is handing people the tool to easily make them... so don't build the tech, sell it
> plug the Higgsfield api into an agentic system
> get stupid pricing up to 50% off while the deal is live (it ends today)
> stack your offering with everything they built over the years... Genjutsu, Cinema Studio, Soul, cheap Seedance 2.5 with face inputs in the US
> add prebuilt templates for the formats already going viral, charge extra for one-click
people will buy from you if you completely delete the part where they have to engineer stuff and understand how it works
1 day left to lock in up to 50% OFF Higgsfield API.
Build your own AI app with our product endpoints, including:
• Higgsfield Genjutsu
• Cinema Studio 4.0
• Higgsfield Soul
Plus all frontier video and image models through the same API.
Paid partnership (ad)
10
5
70
14,336
Machina retweeted
DigitalOcean Managed Agents is now in public preview.
Run Claude Code, Codex, or your own LangGraph agent in a runtime environment that pauses when idle. Put its tools behind one governed endpoint, and pick from 75+ open and proprietary models. One cloud, one bill.
Prompts to get started available in the blog: do.co/4ysh3it
263
843
2,783
3,813,798
most of the Jev use cases i'm seeing on X are absolutely useless
no denying AI agents will lean on system one models hard in the upcoming months... but from my (short) experience it overcomplicates systems most of the time
i got it working in some SEO and content pipelines i run, and honestly that's about it... everywhere else a cheap flash model gets the job done
because it can't technically write text, you still need an LLM to process its outputs, and i hate stacking layers over layers for simple tasks
that being said, it's incredibly fast at computer & browser use, i think this is where it will shine most
excited to see where this goes, i genuinely think it ends up in every serious agent stack... but like everything you gotta let it breathe a little
46
2
161
14,226
join 15k founders using my system, plus one email per week (it's free) weeklyaiops.com
1
2
2,005
