Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
_pdp_
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
31 ms
·
541.
▲
by
_pdp_
1y ago
also, we are one prompt away from achieving AGI...
542.
▲
by
_pdp_
1y ago
What problem does it solve?
543.
▲
by
_pdp_
1y ago
As someone who is deeply involved in this space, I fail to understand what the product does. This is genuine feedback. I signed up but not sure what I am looking at.
544.
▲
by
_pdp_
1y ago
The app router is certainly challenging. This is why we stick to the page router and it works really well. There are some cases where the app router comes handy though. Honestly, nextjs is not that bad and I would go as far as saying that i
545.
▲
by
_pdp_
1y ago
That is a real problem indeed. It is the same problem though. Obviously the developers that will get hired to do the 10% need to be developers with a lot of experience but as you say where you are we going to find these people if the overal
546.
▲
by
_pdp_
1y ago
I agree and I think short term companies will most likely reduce their developer count because AI seems to be cheaper. However, as more code is written by AI the demand for more software will increase therefore the demand for developers wil
547.
▲
by
_pdp_
1y ago
Assuming that the demand for software remains at the same level as before the infliction point. However, this is not likely to be the case ( https://en.wikipedia.org/wiki/Jevons_paradox ). What I am speculating that we w
548.
▲
by
_pdp_
1y ago
In both cases productivity increases though but I agree it will increase the valuation of certain companies tremendously. That being said I cannot see this as being a zero-sum game. We multiple on both ends of the equation.
549.
▲
by
_pdp_
1y ago
BTW, I also posted a longer version of this realisation here for what's worth: https://go.cbk.ai/more-work
550.
▲
Ask HN: Should we stop worrying that AI will replace developer jobs?
10 points
by
_pdp_
1y ago
|
20 comments
551.
▲
by
_pdp_
1y ago
Very simplistic view on the problem domain IMHO. Yah sure we can add a bunch of functions... ok. But how about snapshotting (or at least work with git), sandboxing both process and network level, prompt engineering, detect when stuck, model
552.
▲
by
_pdp_
1y ago
I mean these are two completely different approaches for completely different purposes. So saying that git replaces vector database is like saying that that a filing cabinet replaces a search engines.
553.
▲
by
_pdp_
1y ago
Amateur level of security - what more is there to say?
554.
▲
by
_pdp_
1y ago
You are joking, but there was actually a very popular enterprise SAST tool that used to offer a "cloud" version of their software. It worked by having someone from their team manually download the zip file of your code, run it thr
555.
▲
by
_pdp_
1y ago
The reason text works is because it has higher bit rate then speech. This is way many believe that CLI tools are still considered supreme in terms of getting things done quick. While fun this game-like interface is too casual and it certain
556.
▲
by
_pdp_
1y ago
LLMs cannot build software on their own yet. They are can sure build software with some help.
557.
▲
by
_pdp_
1y ago
I've asked copilot (Claude Sonnet 4) to edit some specific parts of a project. It removed the lines that specifically have comments that say "do not remove" with long explanation why. Then it went ahead and modified the unit
558.
▲
by
_pdp_
1y ago
The way the function is described to the LLM matter. Even when the parameters are the same and the effect is the same the title and description can fundamentally influence how the task is performed. The other issue is that you cannot think
559.
▲
Before AI Slop We Had Spaghetti Code
(chatbotkit.com)
2 points
by
_pdp_
1y ago
|
0 comments
560.
▲
by
_pdp_
1y ago
On a somewhat related note, I am the founder of a company that relies on integrations with 3rd-party systems. Good luck building for Google, Shopify, Slack, etc without going though absurd number of requirements and sacrificing hard earned
561.
▲
by
_pdp_
1y ago
The technology is useful but not in the way it is currently presented.
562.
▲
by
_pdp_
1y ago
It is worth pointing out that both of them are workflow tools first. The AI agent part is expressed though steps where a model is invoked. This opens the door to a number of issues including vulnerabilities. My personal, and my I say biased
563.
▲
by
_pdp_
1y ago
As a side note, Kiro is a diminutive of Kiril (Кирил) in Slavic-speaking countries and in Greek. A quick web search also reveals that it can mean "a curse".
564.
▲
by
_pdp_
1y ago
The reason Typeform free tier limits are so strict is likely because they have run the numbers based on real usage data. I am sure those limits are designed to capture just enough free users who are likely to convert, while minimizing the r
565.
▲
by
_pdp_
1y ago
This is a problem for new generations that should be mindful of but then again how many people can whip some assembly? Certainly not the majority of developers and it is certainly not required for most programming tasks. We might end up in
566.
▲
by
_pdp_
1y ago
> I've spent the weekend trying my hand at making an AI assistant SaaS (I can't believe this doesn't exist yet!), and the biggest lesson I learnt is that I need to pay attention to what Claude Code does. It is part of the
567.
▲
by
_pdp_
1y ago
thanks, we are building chatbotkit.com / cbk.ai (not opensource)
568.
▲
by
_pdp_
1y ago
The classic playbook: copy an open-source project (or just vibe-code something similar), slap an open-source label on it, and toss in an unproven design system / framework (like Liquid Glass) to give it a shiny veneer. Less about build
569.
▲
by
_pdp_
1y ago
Mildly interesting article - I mean, you can already run a ton of libraries that talk to an inference backend. The only difference here is that the client-side code is in Python, which by itself doesn't make creating agents any simpler
570.
▲
by
_pdp_
1y ago
Considering that most meeting software have built-in call record and transcription features, it is less wasteful and less distracting to automatically send the artefacts to all participants.
More ›