AI. Before it's obvious. → Tools, prompts and workflows for people who build with AI | DM for Collab | dexxonai@gmail.com
New York, USA
- Tweets 305
- Following 112
- Followers 22,325
- Likes 550
10. GLG (Gerson Lehrman Group)
→ Pay: $150 to $1,000 per hour.
→ Users can earn by sharing expert knowledge with the world's top investment firms, consulting companies, and Fortune 500 executives. No long-term commitment. Work on your schedule. One hour of your expertise can earn more than a full week at a traditional job. The most powerful remote income platform nobody talks about.
1
47
Dex retweeted
Stop telling ChatGPT “Make a graphic.”
Stop telling ChatGPT “Design a flyer.”
Stop telling ChatGPT “Create a poster.”
Weak prompts = Poor designs and AI slop.
Try these 7 powerful commands to create scroll-stopping visuals with ChatGPT:
11
53
175
14,013
Dex retweeted
10 Ways Claude AI Will Upgrade Your Stock Market Trading Strategy 📈
Stop analyzing stock charts manually. Let AI do the heavy lifting! 🤫
[🔖 Bookmark this thread for later]
43
62
93
3,436
(1/4)
Stop wasting premium API tokens on coding tasks that don't need frontier models.
Not every refactor requires an expensive cloud model. For routine boilerplate, code transformations, and smaller debugging tasks, local inference can be a practical alternative.
My approach: run coding models through Ollama, keep sensitive workloads local where possible, and reserve premium APIs for complex architectural reasoning.
Here's how to think about the stack.
9
48
194
1,934
(3/4)
IDE INTEGRATION & DATA PRIVACY
A local model is only one part of a private coding workflow.
You also need to understand how your IDE handles prompts, repository indexing, extensions, and external requests.
My checklist:
Connect through a supported local-model integration or compatible client.
Verify where prompts and source files are processed.
Disable unnecessary external integrations for sensitive projects.
Test the workflow on a real repository before relying on it.
Running Ollama locally doesn't automatically make your entire IDE workflow private.
1
1
144
(4/4)
LOCAL VS. CLOUD: THE PRACTICAL TRADE-OFF
Local inference makes the most sense when predictable costs, privacy, and control matter more than raw model capability.
Cloud models remain useful when you need stronger reasoning, difficult debugging, or complex multi-file changes.
My rule:
Local models for routine work.
Cloud models for problems that justify the cost.
The goal isn't to eliminate cloud AI. It's to stop paying premium prices for tasks that don't need it.
Are you running coding models locally, or do you still rely entirely on cloud APIs?
1
150
Dex retweeted
This AI short film is next-level cinematic 🔥
Hollywood might be in trouble fr 👀
Made with Seedance 2.5
35
51
128
5,737
Dex retweeted
iRobot and its competitors have sold 40+ million robot vacuums worldwide.
99% of owners use exactly one feature: pressing "Start" and praying it finds the dock.
I spoke to a robotics engineer who builds autonomous navigation systems. He revealed that your vacuum isn't just a cleaning tool—it’s a self-driving car shrunk down to fit under your couch.
You are using maybe 20% of what the hardware can actually do.
Here are 9 hidden features your robot vacuum has that you're entirely ignoring: 🧵
21
40
59
963