mogn
66 subscribers
170 photos
15 videos
4 files
244 links
Building useless stuff and learning useful things along the way. Being a fool.
Download Telegram
OpenAI DevDay πŸ”₯
πŸ”₯1
I've surely forgotten the names of the graph search algorithms I learned in my AI class, but there are some insights I gained that I believe I either wouldn't have learned or would've taken me much longer to learn if it wasn't for that class. Here are some of those insights:

1. Judge decisions by the information available at the time, not in hindsight
When evaluating past decisions, you can only fairly assess whether you made the best choice given the information you had at that moment. Hindsight criticism using knowledge you gained later is invalid. The real questions to ask are: "Did I make the best decision with what I knew then?" and "Did I fail to gather information that was actually available to me at the time?"

2. Optimizing for immediate wins can lead to worse long-term outcomes
Always choosing the option that looks best right now: "winning" at every decision point, doesn't guarantee the best overall result. Sometimes the optimal path requires accepting short-term setbacks, costs, or seemingly worse positions in order to reach a better destination. Greedily maximizing at every step can trap you in local optima and prevent you from finding the global optimum.
Dunning-Kruger Effect at scale πŸ˜‚
😁2πŸ€”1
mogn
The CEO’s Wife Is In Your Marketing Meeting
An interesting read. I wouldn't say engineering fields are totally immune to the problem marketing faces.
There are often numerous situations were there are multiple approaches to solving a problem. And more often than not, we (without being fully aware) fight/argue for the approach we come up with regardless of what is best for the product being built.
Forwarded from LinkUp Premium
Voice assistants, which can be woven into everyday tasks, are increasingly common in places. Yet, remain largely inaccessible because they support only a limited number of languages.

Eyob’s observation of this gap led him to create Rewina, an Amharic voice assistant that listens, understands, and acts. Can Rewina’s early stages grow to Ethiopians using their phones with ease?

Read more: https://linkupaddis.com/app/articles/388

@linkupbusiness
❀2
Ever wondered about how Nano Banana Pro is so damn good?
It was trained on Google Photos πŸ’€πŸ’€
😁3
Just do projects for yourself for fun. You'll get better and better. If you try to make such projects, unseen by others, as perfect as any human could, you'll develop skills that other professionals don't have.

– Steve Wozniak
πŸ”₯4
መልካም αŒˆαŠ“ everyone!
πŸŽ‰3
mogn
If you try to make such projects, unseen by others, as perfect as any human could, you'll develop skills that other professionals don't have.
This is a great advice if you're building for the sake of learning, but the worst advice if you're building to ship. The perfect product never gets released.

Perfectionsim is a bug not a feature.
Developing for/with apple products is the most frustrating experience ever. It's like they go out of their way to make the developer experience/life miserable.
😁3