karmasimida

7 ms·

91.
by karmasimida 8mo ago
They do have value, because they are much much cheaper. But no, 3.0 flash is not as good as 2.5 pro, I use both of them extensively, especially in translation. 3.0 flash will confidently mistranslate some certain things, while 2.5 pro will
92.
by karmasimida 8mo ago
Flash models are nowhere near Pro models in daily use. Much higher hallucinations, and easy to get into a death sprawl of failed tool uses and never come out You should always take those claim that smaller models are as capable as larger mo
93.
by karmasimida 8mo ago
I am fine with the founder joining OpenAI, he gets to get paid regardless. I am not confident that the open source version will get the maintenance it deserves though, now the founder has already exited. There is no incentive for OpenAI to
94.
by karmasimida 8mo ago
Why? You can literally ask codex to build a slim version for you overnight. I love OpenClaw, but I really don't think there is anything that can't be cloned.
95.
by karmasimida 8mo ago
It is over
96.
by karmasimida 8mo ago
LLMs are new runtimes.
97.
by karmasimida 8mo ago
> If you would like to grieve, I invite you to grieve with me. I think we should move past this quickly. Coding itself is fun but is also labour , building something is the what is rewarding.
98.
by karmasimida 8mo ago
Then you are using it the wrong way Driving is a skill that needs to be learnt, same with working with agents.
99.
by karmasimida 8mo ago
I give a year, the realization would be brutal.
100.
by karmasimida 8mo ago
> With Codex (5.3), the framing is an interactive collaborator: you steer it mid-execution, stay in the loop, course-correct as it works. > With Opus 4.6, the emphasis is the opposite: a more autonomous, agentic, thoughtful system tha
101.
by karmasimida 8mo ago
Your feeling is not my feeling, codex is unambiguously smarter model for me
102.
by karmasimida 8mo ago
For those who cared: GPT-5.3-Codex dominates terminal coding with a roughly 12% lead (Terminal-Bench 2.0), while Opus 4.6 retains the edge in general computer use by 8% (OSWorld). Anyone knows the difference between OSWorld vs OSWorld Verif
103.
by karmasimida 8mo ago
I think people here need to accept that software is becoming electricity, you get charged when you use it and by how much. You don't pay for a box shaped electricity or purple color electricity, it is just electricity.
104.
by karmasimida 8mo ago
It is obvious. A middle 100-500 heads firm don't need enterprise level SaaS, a vibe coded website will suit them better. Fundamentally, those workflow/orchestration SaaS needs to answer the question why people should pay you premi
105.
by karmasimida 8mo ago
This is true. The font and animation feels basic to me, even as a programmer focused app
106.
by karmasimida 8mo ago
Not to rain on the parade, but this app feels to me ... unpolished. Some of the options in the demo feels less thought out and just put together. I will try it out, but is this just me, or product/UX side of recent OpenAI products are
107.
by karmasimida 9mo ago
Regardless, knowing syntax of programming language or remember some library API, is a dead business.
108.
by karmasimida 9mo ago
The one that has high security barrier to enter. Payment is probably safe. All workflow orchestration related are totally f*cked
109.
by karmasimida 9mo ago
You might as well just have planner and workers, or your architecture essentially echos to such structure. It is difficult to discern how semantics can drive to different behavior amongst those roles, and why planner can't create those
110.
by karmasimida 9mo ago
Codex’s only caveat is too slow. This is the biggest UX killer，unfortunately
111.
by karmasimida 9mo ago
TBH, the nano banana ones are closer to pixel art than Qwen Image ones. Much closer. This looks like early 2000 2.5D art, like Diablo style.
112.
by karmasimida 9mo ago
This is fantastic, can't overstate how revolutionary this would be for gaming development.
113.
by karmasimida 9mo ago
Same just make it a survival game
114.
by karmasimida 9mo ago
I am able to beat this 1487 benchmark by switching between LLMs, doesn't seem that hard lol. Albeit, I do not fully understand what the solution is, loll
115.
by karmasimida 9mo ago
As impressive as Opus 4.5 is, it still fails in one situation that it assumes 0-index while the component it supposes to work with assume 1-index. It has access to the said information on disk, but just forgets to look into. Opus 4.5 is inc
116.
by karmasimida 10mo ago
Can't really fault him for having this feeling. The value proposition of software engineering is completely different past later half of 2025, I guess it is fair for pioneers of the past to feel little left behind.
117.
by karmasimida 10mo ago
With RAM/memory price this high, open source is not going to catch up with closed source. The open source economy relies on the wisdom of crowds. But that implies and equal access to experimentation platforms. The democratization of PC
118.
by karmasimida 10mo ago
Hangul is its own alphabet and has its own keyboard, so the letters typed don't have correlation with typical romanization scheme at all. It is probably more like bopomofo keyboard for Chinese
119.
by karmasimida 10mo ago
Saas is going to chipping away. Agents offer a very attractive level of abstraction, and its customers, aren't necessarily human: it could be other agents. Many saas we have today would simply be unnecessary in the future. If 90% of th
120.
by karmasimida 11mo ago
That is the problem. AI is optimized to solve a problem no matter what it takes. It will try to solve one problem by creating 10 more. I think long time/term agentic AI is just snake oil at this point. AI works best if you can segment

More ›