4 ms·

People who are saying they're not seeing productivity boost, can you please share where is it failing? Believe it or not, I still know many devs who do not use

People who are saying they're not seeing productivity boost, can you please share where is it failing? Believe it or not, I still know many devs who do not use any agents. They're still using free ChatGPT copy and paste. I'm going to guess that many people on HN are also on the "free ChatGPT isn't that good at programming" train.

dataflow 8mo ago
Which one would you recommend as the best right now? Claude Code?
- redhed 8mo ago
  I have been having a lot of success with Cursor. I like being able to switch between Anthropic and OpenAI models. Claude Code does gives way more tokens/$ than Cursor right now though.
throwaw12 8mo ago
> They're still using free ChatGPT copy and paste Probably that's the reason why some people are sure their job is still safe. Nature of job is changing rapidly
- aurareturn 8mo ago
  I totally get tech CEOs who threaten to fire their devs who do not embrace AI tools. I'm not a tech CEO but people who are anti-LLM for programming have no place on my team.
  - salawat 8mo ago
    And you are paying for their tokens on top of their salary, right? Right?
    - aurareturn 8mo ago
      You can do a lot with just $20 Codex CLI subscription. Tokens are cheap compared to the $20k we're paying for a dev each month.
      - baq 8mo ago
        Exactly, the $20 codex is so good value it’s irresponsible to not give it to everyone. Claude code $20 is otoh pointless, the limits are good enough for 10 mins of work twice per business day.
      - ido 8mo ago
        Even the $200 claude max monthly subscription is peanuts compared to salary cost.
        monksy 8mo ago
        Tell that to the company that I was just at that cut Intelij licenses as cost cutting measures.
        aurareturn 8mo ago
        If they really want to cut cost, fire the worst dev on the team and use that money to give everyone a Codex subscription.
        KronisLV 8mo ago
        Or better yet, fire the managers or bean counters that think decreasing everyone’s productivity is good for long term savings. I’m reminded of https://www.joelonsoftware.com/2000/08/09/the-joel-test-12-steps-to-better-code/ https://www.joelonsoftware.com/2000/08/09/the-joel-test-12-s...
        mikkupikku 8mo ago
        Fire the middle management, HR, and etc that have been enthusiastically using AI to do their jobs for the past two or three years already. 90% of them can be replaced by an agent with access to an email account.
  - cherrycherry98 8mo ago
    Anecdotally, the person on my team who is producing the most output and the highest quality happens to explicitly shun LLMs and is also relatively young. Other team members embracing AI spend their day being misled by chat bots, trying to get their agents to work properly by toying with contexts (just needs a little more knowledge!), and producing verbose code that's hard to review and with obvious bugs when you actually think about what it's doing. My favorite is how everything is excessively commented but more than once I've caught the code not matching what the comment said it would do! FWIW I find it useful if I know exactly what I want and it's quicker to prompt it than type it myself. Also for research and building understanding it's generally good. I still catch it being wrong on details of you're really paying attention or literally contradicting itself between prompts. That gives me a lot of pause about trusting things it told me that I just accepted as fact without having enough knowledge myself to question it.
salawat 8mo ago
Not everyone has the capability to rent out data center tier hardware to just do their job. These things require so much damn compute you need some serious heft to actually daisy chain enough stages either in parallel or deep to get enough tokens/sec for the experience to go ham. If you're making bags o' coke money, and deciding to fund Altman's, Zuckernut's or Amazon/Google's/Microsoft's datacenter build out, that's on you. Rest of us are just trying to get by on bits and bobs we've kept limping over the years. If opencode is anything to judge the vibecoded scene by, I'm fairly sure at some point the vibe crowd will learn the lesson of isolating the most expensive computation ever from the hot loop, then maybe find one day all they needed was maybe something to let the model build a context, and a text editor. Til then wtf_are_these_abstractions.jpg
- ijk 8mo ago
  This is my current problem: I can get work to pay for a Claude Max subscription, but for personal use or to learn how to use it that's a big price tag. I worry that we're returning to an era of renting core development tools. After the huge benefits from free and open source tools, that's a bitter pill to swallow.

dataflow 8mo ago

Which one would you recommend as the best right now? Claude Code?

redhed 8mo ago
I have been having a lot of success with Cursor. I like being able to switch between Anthropic and OpenAI models. Claude Code does gives way more tokens/$ than Cursor right now though.

throwaw12 8mo ago

> They're still using free ChatGPT copy and paste Probably that's the reason why some people are sure their job is still safe. Nature of job is changing rapidly

aurareturn 8mo ago
I totally get tech CEOs who threaten to fire their devs who do not embrace AI tools. I'm not a tech CEO but people who are anti-LLM for programming have no place on my team.
- salawat 8mo ago
  And you are paying for their tokens on top of their salary, right? Right?
  - aurareturn 8mo ago
    You can do a lot with just $20 Codex CLI subscription. Tokens are cheap compared to the $20k we're paying for a dev each month.
    - baq 8mo ago
      Exactly, the $20 codex is so good value it’s irresponsible to not give it to everyone. Claude code $20 is otoh pointless, the limits are good enough for 10 mins of work twice per business day.
    - ido 8mo ago
      Even the $200 claude max monthly subscription is peanuts compared to salary cost.
      - monksy 8mo ago
        Tell that to the company that I was just at that cut Intelij licenses as cost cutting measures.
        aurareturn 8mo ago
        If they really want to cut cost, fire the worst dev on the team and use that money to give everyone a Codex subscription.
        KronisLV 8mo ago
        Or better yet, fire the managers or bean counters that think decreasing everyone’s productivity is good for long term savings. I’m reminded of https://www.joelonsoftware.com/2000/08/09/the-joel-test-12-steps-to-better-code/ https://www.joelonsoftware.com/2000/08/09/the-joel-test-12-s...
        mikkupikku 8mo ago
        Fire the middle management, HR, and etc that have been enthusiastically using AI to do their jobs for the past two or three years already. 90% of them can be replaced by an agent with access to an email account.
- cherrycherry98 8mo ago
  Anecdotally, the person on my team who is producing the most output and the highest quality happens to explicitly shun LLMs and is also relatively young. Other team members embracing AI spend their day being misled by chat bots, trying to get their agents to work properly by toying with contexts (just needs a little more knowledge!), and producing verbose code that's hard to review and with obvious bugs when you actually think about what it's doing. My favorite is how everything is excessively commented but more than once I've caught the code not matching what the comment said it would do! FWIW I find it useful if I know exactly what I want and it's quicker to prompt it than type it myself. Also for research and building understanding it's generally good. I still catch it being wrong on details of you're really paying attention or literally contradicting itself between prompts. That gives me a lot of pause about trusting things it told me that I just accepted as fact without having enough knowledge myself to question it.

salawat 8mo ago

Not everyone has the capability to rent out data center tier hardware to just do their job. These things require so much damn compute you need some serious heft to actually daisy chain enough stages either in parallel or deep to get enough tokens/sec for the experience to go ham. If you're making bags o' coke money, and deciding to fund Altman's, Zuckernut's or Amazon/Google's/Microsoft's datacenter build out, that's on you. Rest of us are just trying to get by on bits and bobs we've kept limping over the years. If opencode is anything to judge the vibecoded scene by, I'm fairly sure at some point the vibe crowd will learn the lesson of isolating the most expensive computation ever from the hot loop, then maybe find one day all they needed was maybe something to let the model build a context, and a text editor. Til then wtf_are_these_abstractions.jpg

ijk 8mo ago
This is my current problem: I can get work to pay for a Claude Max subscription, but for personal use or to learn how to use it that's a big price tag. I worry that we're returning to an era of renting core development tools. After the huge benefits from free and open source tools, that's a bitter pill to swallow.