zaphodq42

Using Claude to do exploration and research is good only for a while. Soon AI slop will take over the real human research. It will be a super expensive task to find out mistakes in AI research. I am worried about AI being trained on AI slop recursively.

Slow hand made path is required in programming because you are programming a machine to do the computation multiple times. Some programs might run for decades and do the same computation billions of times. So what it took a while to write the program! It is worth it to make sure it is right.

Sure, you can vibe code an MVP. But then let’s limit it at that.

Claude works best for context based pattern matching. I specifically use claude for tasks which are simply replicating the logic with minor contextual differences. I can’t even imagine Claude writing whole features by itself.

It can be like the youtube premium model. The search app is subscription based. So every time your content is served you will get paid. But you have to make your content available to the AI for crawling and mention your monetisation preferences.

one of best comment I have read.

Fastmail's spam filter is great if you are willing to put some minor effort. Read this: https://www.fastmail.help/hc/en-us/articles/1500000278142-Im...

Fastmail works great for me. I do not miss any emails as junk nor do i get any junk in my inbox because of using the methods explained in the link above.

Controlling your own email address is the way to go. It takes consistent effort to migrate but worth it when it is done.

Irritation: when the documentation is incorrect or missing and I have to figure out how the API works by guess work or hit and trial.

Anger: when there is unnecessary red tape to get access to an API while my competitors have the access.

Tired: when applying Nth money patch to use a broken API

PS: I work with external APIs a lot

Crypto influencers and companies use "inflation" incorrectly. "Inflation" has 2 means; the physical meaning and the economic meaning.

The physical meaning is the increase in size of an object.

The economic meaning is the decrease in purchasing power of a currency.

Bitcoin will become deflationary (physical term) one day, which means with time total accessible supply of bitcoin will decrease. But this is will be true only when;

number of lost bitcoins > number of mined bitcoins

We can assume this to be true after 24 years perhaps when mining reward will be too low. But at the current time bitcoin is probably inflationary (physical term).

Bitcoin's inflation (economic term) is a function of demand and supply. It is not a function of its reduced mining rewards because that does not say anything about demand. Whenever the price of bitcoin vs USD drops we can safely say the bitcoin has inflated (economic term). Over a period of last 5 years yes bitcoin has been highly deflationary (economic term). It might continue to be deflationary (economic term) for the next 100 years. Which means Bitcoin might continue to rise in value against USD.

I think no one should fool people to say that Bitcoin is deflationary (economic term) just because of limited supply. Limited supply tell us nothing about the demand. Someone's shitcoin is limited in supply too but they can't surely sell it for an ever increasing price. There has to be demand.

When I read PG’s essays I find they are subconsciously trying to pitch YC. How everyone PG knows including him are very special and deserving. This clouds my experience of reading his essays. I am unable to derive any value from them because of this.

The main reason for this is that I see a big divide between what he says about success people and what I have observed. Success has no set pattern. It can come from a huge variety of reasons.

I really hope this is not koolaid because a lot of young people are drinking it. It would suck if it was all just to suck them into applying for YC.

checkout this comment: https://issuetracker.google.com/issues/71840041#comment17

If GAE customers are not notified about issues in advance then GCP should announce that GAE is not production ready.

Hi, I just asked my question in the GAE slack channel. Please check what can I do.

I will try it now. The reason I am giving up is because GAE flexible is unreliable. twice it got the same issue and GCP had no answer why it is happening. Checkout these issues: https://issuetracker.google.com/issues/70365151 https://issuetracker.google.com/issues/71840041

PS: we are running a b2b app. The data stored is transactional. So any downtime is way too damaging for us and our clients.

guys please read my experience with GCP here: https://news.ycombinator.com/item?id=16658339

GCP is not an option for me. Please anyone who has used aws elastic beanstalk share your experience.