The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
AI in job ads: Indonesia vs the world
A research study of how employers rewrote their job ads after ChatGPT, comparing global tech hiring with Indonesian job boards.
What the four pages show:
Global engineering ads that ask for AI went from 3.4% to 23%. Indonesian ones went from 2.2% to 2.8%.
AI shows up as a requirement in every kind of job, not as a new job title. Marketing ads naming AI went from 0.7% to 25.2%.
Skills are shifting: large language models went from 0% to 10.4% of tech hiring posts while React fell from 30.9% to 22.3%. "Founding engineer" went from 2.0 to 21.2 per 1,000 posts.
Global ads ask for ownership of outcomes in 29.1% of cases; Indonesian ads in 5.1%, and 48.9% of them still require a degree (global: 10.3%).
Full report, method and free data (CC BY 4.0):
https://lokerdollar.com/en/research/ai-in-job-ads-indonesia-vs-global-2026?utm_source=contra&utm_medium=social&utm_campaign=labor-lens-2026-10
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
Has AI taken my job yet? 👀
As a designer, I feel like I constantly need to keep up with new tools and products, if only to keep my "has a computer replaced me already?" anxiety under control
At first, we expected AI to do everything with one click
Then we moved on to writing ridiculously long and detailed prompts
And now I feel like we're finally getting to a healthier place with our expectations: build a good system → get good results
I went to play with Recraft's new Logo Maker to see if AI can actually make decent logos now
Weeeeell, you can see the result for yourself! 🤩
Sure, it's not a magic button that reads your mind and does everything for you. But if you have a clear sense of your brand's style and personality, the results can be surprisingly good
Left - my usual aesthetic. Right - Recraft's take on me
So, what do you think of my one-click logo?
Does it feel like me? 🪩✨
“Build a good system, get good results” is exactly where I’ve landed too. The tool only goes as far as your sense of what you’re after. How many tries did it take to get there?
AI WhatsApp Support Assistant (n8n Automation Demo)
Problem: Small businesses lose customers when WhatsApp messages sit unanswered, leads get buried in chats, and complaints are noticed too late.
What I built: An AI assistant that reads every incoming WhatsApp message and prepares a reply, while the owner stays in control.
Writes a reply draft from the business’s own info sheet. Nothing is sent until the owner approves it by email.
Captures real leads (name, interest, budget) into a Google Sheet.
Saves orders and appointment requests with a draft confirmation.
Alerts the owner on Discord for complaints and “talk to a human” requests.
Transcribes voice notes and handles them like text messages.
Alerts the owner to reply manually when someone sends an image or sticker.
Posts a daily 9 PM summary of the day’s conversations to Discord.
Role: Nisbah Iqbal, AI Automation Builder
Good work .
Your trading bot isn't losing money because the strategy is bad. Here is the real reason.
I see the same post in trading forums every week: "My bot works in backtest, but it bleeds money live."
The developer assumes the strategy is wrong. They tweak the RSI. They add a moving average. It bleeds again.
The real problem is almost never the strategy. It is that the bot is trading on headlines that were never about oil in the first place—or it is trading on stale news because it fired on the first tweet it saw instead of the newest one.
When I built my autonomous oil trading system, I did not trust a keyword matcher to decide when real money moved. I put a reasoning model in front of every headline. It returns four things before a trade is allowed to fire:
Is this actually about oil? (Most aren't. They get filtered.)
Impact: BULLISH / BEARISH / NEUTRAL
Strength: 1–5, where 5 is "war, embargo, major supply shock"
Confidence: 0.0–1.0
A trade only executes if confidence passes a hard threshold. And if the reasoning API ever fails, a circuit breaker falls back to a deterministic engine—the system gets dumber for one cycle, it does not go dark.
The monitor polls every 20 seconds and primes on the first poll. It records the baseline tweet ID and does not trade on old news. That single design choice killed more false signals than any strategy tweak ever did.
This is not a script. It is a pipeline that reads the world, reasons through the noise, and executes with precision. All for $0 to run.
If your bot is bleeding in live markets and you cannot find the root cause, that is the exact problem I solve. I read the architecture, I find where the signal is leaking, and I ship a fix.
If you are stuck on an execution or signal problem, my inbox is open. Tell me the symptom, and I will tell you if it is fixable.
🔗 Full architecture: github.com/S-7-D-2023/OIL-NEWS-SYSTEM
#AITrading #Python #LLM #BuildInPublic #QuantDev
At a 20s poll the latency is fine, the cadence is what protects you. Two things I would add before a local model: a dedupe key on the headline, hash plus a short window, so a republished story does not fire twice, and a per-headline TTL so a two-minute-old item cannot trigger a...