"I'm Not Good at Using AI Yet" — That Person Made the Best Result

The person who said "I'm not good at using AI yet..." ended up solving problems the best and completing results most token-efficiently on the spot—that was SouthBridge's AI Workshop on August 1st..!

People from truly diverse fields—from university students to SW auditors, association staff, pet service operators, opticians, F&B, teachers, and developers—wanted to properly use AI. We met with everyone having thought about "what they wanted to create today" in their minds.

The AI workshop was divided into three main parts. First, I unfolded stories about the AI scene and communities I've seen while dragging my suitcase from place to place.

Countless moments including Zero100 where I met Daero Won, the absolutely amazing omocon, Hashed vivelabs (Nitro) led by the reliable Subin An, the beloved ulw event, AIMASTERS by Seung-oh Jung and Silvastone, the brilliantly blooming Bloom by Patrick Han, CF APMP, [ICML] Int'l Conference on Machine Learning, and Gajayeonguso (Pseudo Lab) truly led by Chanran Kim, and many more.

The journey that started with OpenAI Codex Meetup Busan with Ralphton, which opened the door to my 'planner Heejin Jung' clown(?) life, when I said 'let's do it together, yay!' with Jin Hyung Park (Sigrid Jin) and coordinated with Goobong Jeong at a gukbap restaurant in Busan's Jagalchi Market. From SouthBridge coffee chats, gukbap chats, events, to the AI sharing session and this workshop event in collaboration with Genspark, which recently invited us to their Korea debut event.! (Long, so long.. that's why it took an hour)

By reviewing honestly and unfiltered, mostly with posts I'd written on LinkedIn, we were able to discuss each and every 'perspective on encountering and facing AI until now and in the current generation.' Since I talked about the outputs I'd created or contributed to at hackathons and build events, people focused even through the somewhat lengthy session. (They said it helped their understanding.. thank you thank you thank you) (I was amazed and truly grateful to meet someone who consistently reads my posts)

Afterward, we had time for everyone to immediately create what they wanted to make on the spot, providing 8,000 credits from Genspark—nearly a month's worth of tokens—for practice.

The content I emphasized most before starting was 'introduce yourself well.'

1. I told them to try putting in all the materials they thought might be necessary. 'If you recite your situation thoroughly to someone whose mind works better than yours, your intent becomes clear and variables decrease, so desired results will come out well,' so I recommended leaving lots of materials and content and passing the baton to the superagent.

2. I told them to turn on cloud computer. Since it creates and runs a mini instance (easier to explain if you call it a small computer), it parses and reads well even documents it couldn't read before, and reads files and GitHub repositories well, so I told them that especially from the perspective of working with source code, quite interesting results could emerge.

The results afterward were definitely interesting.

When I asked before the practice session what AI tools they used, more than half were subscribing to AI plans from OpenAI, Anthropic, and Google AI, and there were people who used to spend $400 a month on Hermes agent but recently trimmed their workflow build well to about $30.

So I truly honestly expected this in this session: 'Genspark might.. be somewhat disappointing in terms of results, speed quality, and credit policy for users who have already tried many other AI tools, so maybe it would be seen as having a definite unique position here because the outputs that can be created in parallel based on cloud instances are decent..'

However, starting with someone easily whipping up a demo of the Southeast Regional Headquarters official website for an association with about 800 credits, with quite good design, evaluations continued that when they built an actual web game from a game proposal they'd done with another tool before, the quality came out with really great detail.

Particularly, even though they were using all other AI tools diligently, they said that Genspark's outputs were relatively more professional and polished compared to other tools.

Besides Genspark, diverse outputs came out as diverse as the fields—from financial seminar PPTs to easily making Instagram card news, Airbnb opening kits, pet-specialized business plans, Academy videos, business interior design, and SW management workflows.

With feedback that they could create much more diverse things at a satisfying quality level than expected and wanted to use it more diligently, most were satisfied with Genspark's output and gave feedback that they became more interested and wanted to learn more, or wanted to use credits more abundantly to complete what they were doing at a high level.

As we reviewed what each person created, we ended up running over 30 minutes past the expected time to finish the schedule. Nevertheless, thank you for the feedback that everyone enjoyed it and it was good to communicate by directly hearing and asking about various cases. We wrapped up the workshop in a friendly atmosphere.

I would like to thank Genspark once again for their great help at this event. I sincerely thank each and every person who showed interest in me and SouthBridge, supported us, and participated.

Going forward, SouthBridge will continue to meet you in various moments as a community that contemplates many attempts, problem-solving, sharing and development, and sustainability together. We look forward to your interest and support :)