Can @TPracticalAI

Testing AI tools and prompts in real work. What actually saves time. What doesn’t. No hype. Joined October 2026

Tweets

249
Followers

683
Following

974
Likes

49

Can @TPracticalAI

15 hours ago

Yesterday: same task, 2 AI tools, 1 try each. Tomorrow: same task, third tool. What saves time and what doesn't.

0 0 0 5 0

View Details

Can @TPracticalAI

a day ago

@builtwithrubel E) for me. Misuse for bad purposes

0 0 0 3 0

View Details

Can @TPracticalAI

2 days ago

Prompt: Summarize the email below in 3 sentences. Then list the one action I need to take. Keep a neutral tone. "sample email, no real names"

0 0 0 11 0

View Details

Same task. 2 AI tools. Summarize a sample email in 3 sentences + 1 action. Both followed the rules. Both picked the same action. Grok kept more detail (the Wednesday fix, the IT dependency). ChatGPT was cleaner, but added a small guess about why the date moves. Tries: 1 each. Which one would you want ?

1 0 1 26 0

View Details

Can @TPracticalAI

2 days ago

@sandwithin Impressive

0 0 0 7 0

View Details

Can @TPracticalAI

2 days ago

@BenBusker That’s right

0 0 0 6 0

View Details

Can @TPracticalAI

2 days ago

How many AI apps do you actually use every day? Which one are you happiest with right now? Why that one? Let’s talk in the replies.

0 0 1 23 0

View Details

Can @TPracticalAI

3 days ago

@ardent__dev Hard to pick without the same task on all four. I'd compare how many retries it takes to get something that actually runs. That number tells you more than any feature list

0 0 0 5 0

View Details

Can @TPracticalAI

3 days ago

@sssirxn The glowing star and the cloud texture make this work. Did you get it in one prompt, or did it take a few tries ?

0 0 0 9 0

View Details

Can @TPracticalAI

3 days ago

Ultra-detailed cosmic wallpaper, spiral galaxy with glowing blue and violet nebula clouds, bright star clusters, deep black space, subtle dust lanes, cinematic lighting, high contrast, vertical 9:16, small subtle text "@TPracticalAI" in the bottom right corner, clean readable font, no other text

0 0 1 33 0

View Details

Can @TPracticalAI

3 days ago

Made this galaxy wallpaper with Grok. 3 tries. The first one was the best. Re-rolling didn't help this time. Prompt in the replies.

1 0 1 56 0

View Details

Can @TPracticalAI

3 days ago

@YouArtStudio For me the real test isn't the best output, it's how many retries it takes to get a usable one. Did either model need a re-roll here ?

0 0 0 24 0

View Details

Can @TPracticalAI

3 days ago

Same task. 3 AI models. 6 outputs. We tested ChatGPT, Claude, and Grok with basic and detailed prompts. The detailed prompts produced more focused answers. But the basic prompts had strong ideas, too. The takeaway: Better prompts give you more control, not guaranteed quality .

0 0 1 53 0

View Details

Can @TPracticalAI

4 days ago

@drewxeth @termix_ai Hiring agents is easy. Trusting the output without looking is the hard part.

0 0 0 2 0

View Details

Can @TPracticalAI

4 days ago

@diegocabezas01 Walking while agents work sounds nice until you come back and have to check what they did.

0 0 0 2 0

View Details

Can @TPracticalAI

4 days ago

@Tajudeen @AnthropicAI @claudeai Claude asking before it does something sensitive is the part that actually matters.

0 0 0 3 0

View Details

Can @TPracticalAI

4 days ago

Agents are handling more daily tasks now. Still feels like most of the real value is in catching the small drifts. Anyone else finding the review step hasn’t disappeared?