Elie Hamouche
968 posts
Elie Hamouche
@thehamouche
Building https://t.co/MBdE4ej0gA
London, England Katılım Ekim 2010
1.7K Takip Edilen526 Takipçiler
@jamonholmgren This is exactly what I'm doing.
Would love to share some skills that have made my setup easier.
English
My current agentic workflow is about 5x faster, better quality, I understand the system better, and I’m having fun again.
My previous workflows have left me exhausted, overwhelmed, and feeling out of touch with the systems I was building. They also degraded quality too much.
This is way better.
I’m not ready to describe in detail. It’s still evolving a bit. But I’ll give you a high level here.
I call this the Night Shift workflow.
English
pay attention to what just happened here...
this guy built a full ad launching system using Claude Code
> a single tool that launches bulk ads across Meta, TikTok, and Snapchat
> different formats per placement... 4:5 for feed, 9:16 for stories
> one campaign config screen for everything
> auto UTM tagging on every single URL
> Meta's auto-enhancements disabled by default (no more clicking through every ad to turn them off)
it is so over for media buyers
J.B.@VibeMarketer_
ok so i finally got this working bulk ad creation > configure once > push to meta, tiktok, snapchat > all live in minutes here's the setup: told claude code to build a bulk ads launcher using the meta, tiktok, and snapchat APIs what it does: -> add as many ads as you want. each with its own creative, copy, CTA. -> set shared defaults across all ads. override individual ones where needed. -> upload 4:5 for feed, 9:16 for stories/reels. per-ad placement optimization. -> launch into a new campaign, new ad set, or existing ad set. your choice. -> push to one platform or all three at the same time. -> all meta auto-enhancements turned OFF by default. no more clicking through every single ad to disable advantage+ creative, music, text optimizations. it's just off. (this one was for all my media buyers out there) -> auto UTM tagging on every URL. source, medium, campaign, ad set, ad name. no manual work. -> structured naming conventions. brand, product, concept, creator. ad accounts stay organized across platforms. -> media flexibility. upload files, paste URLs, or drop google drive / dropbox links. it handles all of it. -> campaign config in one screen. objective, budget, optimization goal, bid strategy, targeting. -> launch as active or paused. preview everything before pushing. -> per-ad error handling. if one fails, the rest still launch. you see exactly what happened. what used to take 2+ hours clicking through three different ad managers now takes 10 minutes. drag in creatives. set config. select platforms. hit launch. done.
English
@theo @kentcdodds Would love to see how you use this with convex based projects - I work against a cloud development instance and it limits how much I can do in parallel
English
T3 Code is now available for everyone to use.
Fully open source. Built on top of the Codex CLI, so you can bring your existing Codex subscription.
English
@samlambert @vchennai2 @ArdentAI Is the idea that you’re giving Claude access to a read replica only?
English
@vchennai2 @ArdentAI i don’t know why you are saying it will eventually happen on PlanetScale. What exactly have you built that any of the providers listed can’t just build themselves?
English
If you’re working with RDS, Supabase, Planetscale, etc. this will eventually happen if Claude code has access to prod
Exactly why I built @ArdentAI so you can work on prod but zero risk of deleting your data at the infra level
Alexey Grigorev@Al_Grigor
Claude Code wiped our production database with a Terraform command. It took down the DataTalksClub course platform and 2.5 years of submissions: homework, projects, and leaderboards. Automated snapshots were gone too. In the newsletter, I wrote the full timeline + what I changed so this doesn't happen again. If you use Terraform (or let agents touch infra), this is a good story for you to read. alexeyondata.substack.com/p/how-i-droppe…
English
Introducing Oz: the platform to orchestrate agents in the cloud.
Spin up hundreds of agents from your terminal, browser, the API, or your phone. Each agent gets a @docker environment to build, test, and write PRs.
Come back from your lunch break to code that’s ready to merge.
English
@EthanWoo I think the reason it’s taking off is because you can self host, otherwise there’s plenty of ai assistants that have connectors and mobile apps people can use
English
Someone’s going to just script this and charge $99 setup fee and $50/ a month + token overage. They will become a millionaire.
nader dabit@dabit3
Running @openclaw on Digital Ocean is pretty simple. And for $7/month it might be the best option for beginners as the UX is much simpler than AWS or even Hetzner and the cost is still negligible. Also laid out the instructions here: gist.github.com/dabit3/42cce74…
English
@hellorob Can’t really recover product details with an upscaler either
English
Great prompt, but fine details won't survive the grid.
Nano Banana Pro generates each in ~1k resolution.
Product text gets lost.
This workflow fixes that:
→ Automatically extracts all 9 grid image
→ You select your favorite
→ Upscales to 4K using your input as reference
Firat Bilal@firatbilal
This prompt works great, even with local mineral water brand! :)
English
@alexhillman The Mac mini has enough memory to run models locally
English
Genuine question:
Why are people buying expensive Mac minis to run an agent that rips on a $25/mo VPS you can access from anywhere?
I must be missing something. Please let me be missing something.
Nat Eliason@nateliason
Yeah this was 1,000% worth it. Separate Claude subscription + Clawd, managing Claude Code / Codex sessions I can kick off anywhere, autonomously running tests on my app and capturing errors through a sentry webhook then resolving them and opening PRs... The future is here.
English
Lovable is running ~30k sandboxes concurrently across Modal and GKE, creating and churning ~10 every second.
When one has issues, traffic shifts across regions & providers automatically in a exponential smoothing pattern thanks to a custom scheduler.
Currently working on running hundreds of agents in parallel per project.
We're hiring platform engineers if you want to help.
English
@mattshumer_ @GroqInc @rork Not building this but interested in the motivation behind this. Is the idea that agents from different companies/ecosystems have a way to chat?
English
We did a little bit of tuning to create a virtual photographer.
Reve@reve
2/ Dynamic lighting “Show interesting lighting variations”
English
Static prerendering OOTB with Start on workers now :)
developers.cloudflare.com/changelog/2025…
English
@DmytroKrasun 100% - I think the feedback we’re all giving Claude is way more valuable than the code it sees. Anthropic is going to win by making a very good decision making machine.
English
Nobody is safe anymore.
People who think they are safe if they are maxing Claude Opus 4.5 don't understand they are training the next version of the model and will be replaced, too.
English
English
