Sasha why NOT
--
Codex from OpenAI: Not an assistant, but already a full—fledged programmer?
Listen, think about it - OpenAI has introduced a new version of Codex, and it doesn't look at all like the AI tools we are used to. This is no longer just an auto-completion of the code or the answer to the question "why doesn't it compile". This is a whole system of cloud-based AI agents that can actually do a bunch of things in parallel: fix bugs, refactor, write modules — and all this in the background, without your participation. Yes, while you're drinking coffee or, say, out for lunch.
Codex does not run on your computer, but on the OpenAI infrastructure. That is, it doesn't load your laptop, it doesn't slow down the IDE — it works as a remote employee with a separate machine. Only this "employee" can run multiple tasks at once, check himself, correct mistakes and suggest improvements. Just like an experienced engineer, whom you trust and do not stand over your soul.
Interestingly, they have already made a free local version (Codex CLI), but now we are talking about cloud-based Codex— it is more powerful, more autonomous and more scalable. And it works on their new Codex One model, which OpenAI calls the "best for today" for programming.
At the demo presentation, Codex fixed the code live in 30 seconds, which would have taken a developer half an hour. One of the team members said, "We trust Codex as much as we trust our colleagues." And this, you must admit, is not just a compliment — it's a whole cultural shift.
Interestingly, OpenAI is currently restricting access: only for ChatGPT Pro, Enterprise, and Teams, but they plan to open it later for Plus and Edu users as well. In the beginning, they promise "generous limits" and no surcharges, but it is clear that this is temporary — as soon as the data is collected, they will most likely introduce a fee.
Further — more: They want to integrate Codex into bug tracking systems, into the command line, and generally automate the development lifecycle. And at the same time, they continue to develop the local Codex CLI so that in the future both the local and the cloud will work together as a single team.
Sam Altman even joked on X (former Twitter) that two extremes were writing to him in his personal account.: "You created an AI programmer, why isn't it free?" and "Why do you charge only $20 per month for such a programmer?". And this really reflects the essence of the moment — people have mixed feelings: on the one hand, admiration, on the other, anxiety.
Now I'm sitting here thinking: if AI can already do almost all the engineering routine on its own, how soon will it start replacing entire teams instead of individual specialists? And most importantly, where will we stay in all this? Do you think Codex is a tool to strengthen programmers or a harbinger of their replacement?
#OpenAI #AI #AIAgents #Aİ
Codex does not run on your computer, but on the OpenAI infrastructure. That is, it doesn't load your laptop, it doesn't slow down the IDE — it works as a remote employee with a separate machine. Only this "employee" can run multiple tasks at once, check himself, correct mistakes and suggest improvements. Just like an experienced engineer, whom you trust and do not stand over your soul.
Interestingly, they have already made a free local version (Codex CLI), but now we are talking about cloud-based Codex— it is more powerful, more autonomous and more scalable. And it works on their new Codex One model, which OpenAI calls the "best for today" for programming.
At the demo presentation, Codex fixed the code live in 30 seconds, which would have taken a developer half an hour. One of the team members said, "We trust Codex as much as we trust our colleagues." And this, you must admit, is not just a compliment — it's a whole cultural shift.
Interestingly, OpenAI is currently restricting access: only for ChatGPT Pro, Enterprise, and Teams, but they plan to open it later for Plus and Edu users as well. In the beginning, they promise "generous limits" and no surcharges, but it is clear that this is temporary — as soon as the data is collected, they will most likely introduce a fee.
Further — more: They want to integrate Codex into bug tracking systems, into the command line, and generally automate the development lifecycle. And at the same time, they continue to develop the local Codex CLI so that in the future both the local and the cloud will work together as a single team.
Sam Altman even joked on X (former Twitter) that two extremes were writing to him in his personal account.: "You created an AI programmer, why isn't it free?" and "Why do you charge only $20 per month for such a programmer?". And this really reflects the essence of the moment — people have mixed feelings: on the one hand, admiration, on the other, anxiety.
Now I'm sitting here thinking: if AI can already do almost all the engineering routine on its own, how soon will it start replacing entire teams instead of individual specialists? And most importantly, where will we stay in all this? Do you think Codex is a tool to strengthen programmers or a harbinger of their replacement?
#OpenAI #AI #AIAgents #Aİ
Disclaimer: Includes third-party opinions. No financial advice. May include sponsored content. See T&Cs.
0
0