Cooper Haas

CS & Business Major at Skidmore || Lacrosse || Interested in AI

[Cooper Haas @CooperHaas6]

Exited to try this out!

[Cooper Haas @CooperHaas6]

I used 5.6 sol high and 5.5 pro on virtually the same task: doing some M&A research in a specific domain and giving me a document and presentation formatted output. Here is what I found.

5.6 worked twice as long - 28 min vs. 14 min

Both spun up 3 subagents with research tasks

5.6 looked at 30 different sources while 5.5 50

Both outputs were quite good but 5.6 more thorough and looked better

One issue I have had is when giving a model an example it fixates on that specific example instead of looking at others. In this case I asked 5.6 to look at licensing the product out to different legal jurisdictions like the EU and Japan and it only looked at those cases. It might be an issue with prompting but I have found that mentioning an example can introduce unwanted bias.

[Cooper Haas @CooperHaas6]

Idea for helping GenZ with budgeting.

What features are good/what would you drop?

link to promo vid: https://t.co/4l3ybjuAyQ