When you say it "I don't think it took too much longer to get working code than it would've taken if I just coded up the entire thing by hand with no assistance. I'm going to guess that it took about twice as long, but for all I know it was a comparable amount of time." → I'm actually amazed that it performed as well as it did for native code generation.
I just tried "Write a summary of the content, followed by a list in bullet format of the most interesting points. Bold the bullet points, followed by a 100-character summary of each." Here's the output: https://s.drod.io/DOuPLxwP
Also interesting is "List the top 10 theories of why Sam Altman was fired by the OpenAI board in table format, with the theory title in the first column and a 100 word summary in the second column." Here's that output: https://s.drod.io/v1unG2vG
Helps to turn markdown mode on to see the list & table.
Don’t be. Almost all of it are speculations at this point and no one from inside is going to reveal the secret in a random HN comment.
And also, think about this, unless you’re a shareholder with openai, knowing the “opinion” of others isn’t going to impact your life in any form. I’m not saying you shouldn’t read it or participate in the discourse, but there’s no need to be overwhelmed by opinions let alone build a chat bot to get through it.
Yeah agree with all of the above but with one modification: We can spend a lot of time reading through 1800+ comments, and that's time when we're usually trying to just get a question answered like "what does everyone think happened?" etc.
Storytell's vision is to distill signal from noise → giving you a different way to navigate 1800+ comments where you're the one in control of how you spend your time doing it is our objective here.
Agree. There’s definitely use to get a general pulse of whats happening. This isnt directed towards storytell or such apps, sorry if that came off that way.
I was addressing the parent’s comment about being overwhelmed about not being able to go through the opinions. I’ve felt similar way before I realised they were just fomo. If anything, apps like storytell are actually better way to consume these things :)
Would love to know what you'd like to see us build to make it even better for you!
You can also get to the "ground truth" data by clicking on the [x] reference foot notes which will open up a 3rd panel with the Story Tiles that we pull from our vector DB to construct the LLM response.
Here's an example of how it works -- I asked for a summary of what happened in the voice of Dr. Seuss: https://s.drod.io/9ZuL6Xx8
"Given that the Bitcoin blockchain only produces new blocks once about every 10 minutes, Gamma’s data also changes at most every 10 minutes — making it an ideal use case for caching.
Thomas added Gamma’s GraphQL edge cache in front of Gamma’s Hasura GraphQL API, and immediately got an overall cache hit rate of 87%, which corresponds directly to a decrease in traffic to their Hasura instance.
Even better, their two most highly requested queries are sitting at a 97% and 92% cache hit rate, respectively"
Our GraphQL Edge Cache is based on Fastly's infrastructure, so we do use Varnish under the hood for the cache storage. In order to properly support GraphQL at the caching layer you have to understand GraphQL so you can do fine-grained invalidation.[0]
Essentially, our GraphQL Edge Cache is similar to a GraphQL client. It looks at the request body with the query & the response with the typenames and tags the cached query result with all the objects contained within. E.g. a getBlogPost query that fetches the blog post and its author will be tagged with BlogPost#asdf123 and User#gjkd489. Then we can invalidate whenever those specific objects change.
The number is actually greater than 1,000 right now.
No, we're not scared of Fastly or Cloudflare and are actually cooperating with them.
What makes you think that "they will just clone the moat"?
The question about how many people have the problem isn’t the same question as asking how many people you’ve managed to sign up, can you be clearer about what your number is referring to?
I’ve signed up for many more services than I actually need/use, and even amongst ones I use, there are certainly some solving problems I don’t really have.
I’ll be running Starlink off an Inergy solar generator at the Placerville adventure expo on June 3rd if anyone wants to see it in action. http://go.drod.io/InergyVanlife
Off grid = No need for a permit but more importantly, a proof of concept that with starlink, this small office “shedquarters” could be placed anywhere in the world — say a beach in Costa Rica for example!
Imagine if all the nerds that could afford it set up their shacks on the beach in Costa Rica. Instant slum. I prefer the beach without such structures.
Thanks! Quickly looking at the rules, it doesn't look to me like the issue is whether you're connected to the grid? If you have high-voltage wiring (50V+) it looks to me like you still need to get that permitted and inspected?
I found YC to be extremely valuable for all the reasons I shared.
A side story: After I sold my last company in 2013, I met a guy (who worked at the acquiring company) that had big startup ambitions. We were on a bike ride one day (probably around 2015) and I told him I was going to apply to YC the following year to do a startup. He scoffed, and said the same things as you above — “YC is for young founders w/o experience.” He said he, too, was going to do a startup, but not apply to YC.
Fast forward five years. Armory is growing very fast.
My buddy is unemployed after spending $350k of his own money attempting his startup.
Was YC the difference? Impossible to say. But I know my story is true, and I’m thankful for the intense focus YC provided — even for a startup veteran like me, with a successful exit, who didn’t necessarily think he needed it.
Thanks for trying Storytell for coding work, Dan!
When you say it "I don't think it took too much longer to get working code than it would've taken if I just coded up the entire thing by hand with no assistance. I'm going to guess that it took about twice as long, but for all I know it was a comparable amount of time." → I'm actually amazed that it performed as well as it did for native code generation.
We put code generation in the "Use with Caution" bucket as I describe in https://web.storytell.ai/blog/the-intersection-of-ai-curious...