Hype, A Fake, and The Future

The Weekly Variable

The Weekly Variable

Way more podcasts this week and a good amount of coding!

A surprisingly busy week!

Here’s the list:

Hype Bid Continues

Last week, Hype Bid officially came to life.

And the feature development continued this week, slowly but surely, as I continue mastering Go.

There are 10 days until the winners of the hackathon are announced, so I figured it doesn’t hurt to sneak in some extra functionality before the final decisions are made.

This week I added signup and login pages with a few pages using session management after logging in, including the Dashboard and the League Details pages.

The app is close to having test screens for all endpoints I whipped up a few weeks ago, but I still need to add members to a league, create auctions and enable trading between members.

Not a ton of pages, but those are getting more complex so it might take some time to get them working properly.

It will be nice to see all of the features finally working in a testable capacity, though.

And it’s been fun to see them come to life on hacked together HTMX pages.

More development to come, but in the meantime, you can check out the progress at hypebid.tv and let me know what you think!

 

The Solo Dev

Due to a miscalculation on my part (still adjusting to the time change and a different timezone) I messed up the coordination of another Dev Sync episode.

We’ll hopefully get another makeup recoding in soon.

But it did bring me back to the idea of doing episodes on my own.

I’ve been debating a Solo Dev episode where I would cover similar topics that Eric, Will and I would normally cover, except it would be more of a lecture type presentation, similar to a Huberman Lab podcast.

Huberman coincidentally released his first episode in his “Essentials” series this week, featuring more condensed 30 minute episodes versus his usual 2-hour-plus run times.

30 minute tech lectures feel much more approachable.

If this is something you’d be interested or have any topic ideas, let me know!

A Fake Friend

I’ve become good coding buddies with the o1 AI models recently while teaming up with them in Cursor.

I actually switch between gpt-4o-mini, Sonnet 3.5, o1-mini and occasionally o1-preview depending on the situation.

For quick debug answers or an error explanation, I’ll use gpt-4o-mini.

Then for building out pages, or if I’m not feeling like debugging an issue but I have an idea of what it is, I tend to switch to Sonnet 3.5.

And for more complicated logic, I’ll defer to o1-mini or preview typically since it handles more complex ideas by reasoning out the problem first rather than just blurting out the most probabilistic answer.

But I heard recently on Y Combinator’s Lightcone podcast that o1-preview may be a fake!

OpenAI might not actually be giving people access to the real o1 model, but instead have replaced it with a fake model that mimics the “thinking” process.

I’ve still been impressed by the results whether it’s just a modified 4o model or not, but it’s makes me more all the more excited to see what the full o1 model is capable of, which, according to the podcast, the full version should be made public in the near future.

I’m not mad about the fake o1 if that is the case, but I do find it interesting.

The theory is OpenAI doesn’t want to give access to the real version because people will immediately figure out the magic behind it’s thinking, but ultimately it does seem inevitable.

The magic will be revealed eventually.

It makes sense to keep an edge while you can but I’m not sure that edge can be maintained forever.

For now, I’ll happily become coding buddies with whatever the top model may be as long as it keeps being as productive for me as it has been the last few months.

More details about o1 and what it’s capable of in the full episode of Lightcone:

Building the Future

I can’t help but buy into the hype of AI as I personally experience what it’s capable of.

I haven’t done a deep dive into the power of iterating with AI but I think there is still something there that hasn’t full be unlocked.

But I could see how it would be easy to brush off the hype of AI in the real world since it’s still primarily a software tool, and it’s hard to appreciate what it can do if you don’t understand it from a software context, but it will slowly creep into everyday life over the next few years.

Sam Altman talked with Garry Tan about what do to in light of this massive change in software and how to think about the future.

With that in mind, Sam insists this is one of the best times to start a business, and I have to agree.

There’s an overwhelming amount of ways to apply this technology and that’s even from the limit scope of the last few years.

As it continues to improve, as Sam mentions, there may be even better opportunities to start new businesses and build cool things.

This episode is the first in a reboot of their original series “How to Build the Future” and I’ll be anxious to hear more about where these startup founders see things headed as AI continues to improve and apply to more aspects of everyday life.

Also can’t help but be excited at hearing the phrase “solving physics".

More about that in the full episode:

The New Book Tour

I believe this is the first time that Matthew McConaughey has been on Chris Williamson’s podcast, which seems like a big deal for Chris.

Modern Wisdom has grown a huge audience so it’s not surprising, but it’s always interesting to see real world celebrities cross over into a much smaller internet realm.

I know McConaughey has been on Lex Fridman’s podcast so it’s not the first time he’s done a YouTube podcast before, but still surprising to see when it happens.

There does seem to be a trend that podcasts are a powerful marketing tool these days.

If you have a new book to promote, then you make the podcast rounds.

Coincidentally or not, McConaughey just so happened to have a new book coming out which was part of the reason he appeared on the pod.

It’s a great conversation nonetheless, McConaughey is naturally, a very entertaining story teller, giving some back to story to how he got started in Hollywood and how he came back after taking a break.

It’s a fun talk worth listening to and it’s not even a full 2 hours!

And that’s it for this week! More development, AI and podcasts. We’re narrowing down into a nice niche after 50 plus issues!

Those are the links that stuck with me throughout the week and a glimpse into what I personally worked on.

If you want to start a newsletter like this on beehiiv and support me in the process, here’s my referral link: https://www.beehiiv.com/?via=jay-peters. Otherwise, let me know what you think at @jaypetersdotdev or email [email protected], I’d love to hear your feedback. Thanks for reading!