Trending topics
#
Bonk Eco continues to show strength amid $USELESS rally
#
Pump.fun to raise $1B token sale, traders speculating on airdrop
#
Boop.Fun leading the way with a new launchpad on Solana.

Hari
@cantinaxyz @spearbit whole // prev @solidity_lang
I'm starting to come around GPT-5, if you can get it to think hard, it is quite good!

HariAug 8, 23:24
My initial impression of GPT-5 is that it was designed to compete with Gemini.
It seems to be a workhorse model rather than a racehorse.
2.53K
I tried 5 different math problems and it failed on all 5 of them. One of the problem was even in the training set, and it still failed.
It tries its best to find an 'inductive pattern', but doesn't reason quite well.

HariAug 19, 12:13
I had high hopes for Gemini Deep Think. I gave it an Olympiad-style math problem today, and it performed poorly.
It made an incorrect assumption early in the process and continued on the wrong path.
1.91K
Ironically, we're going to see a return to tradition for CS skills.
I believe the most important CS skills now will be compilers, infrastructure, and security.
There's a meme about taking a CS class and not being able to build a full-stack app at the end of it. Not anymore!

HariAug 10, 06:51
On LLMs Writing Code:
- Invest in your CI/CD systems. Think about how you'd set up your pipelines if you had a massive engineering team. That's the level you want to aim for.
- Over invest in automated tools, especially for security. Big teams do this but often ignore the output of these tools. LLMs can help sort through the output.
- Go for a popular tech stack with tons of public training data. These will get better model support. Like React, Tailwind, Postgres, Typescript, AWS, etc.
4.5K
Ironically, we're going to see a return to traditional for CS skills.
I believe the most important CS skills now will be compilers, infrastructure, and security.
There's a meme about taking a CS class and not being able to build a full-stack app at the end of it. Not anymore!

HariAug 10, 06:51
On LLMs Writing Code:
- Invest in your CI/CD systems. Think about how you'd set up your pipelines if you had a massive engineering team. That's the level you want to aim for.
- Over invest in automated tools, especially for security. Big teams do this but often ignore the output of these tools. LLMs can help sort through the output.
- Go for a popular tech stack with tons of public training data. These will get better model support. Like React, Tailwind, Postgres, Typescript, AWS, etc.
80
On LLMs Writing Code:
- Invest in your CI/CD systems. Think about how you'd set up your pipelines if you had a massive engineering team. That's the level you want to aim for.
- Over invest in automated tools, especially for security. Big teams do this but often ignore the output of these tools. LLMs can help sort through the output.
- Go for a popular tech stack with tons of public training data. These will get better model support. Like React, Tailwind, Postgres, Typescript, AWS, etc.
7.31K
Top
Ranking
Favorites
Trending onchain
Trending on X
Recent top fundings
Most notable