The only difference I can see is that the controls for local/remote, branch, and worktrees disappear, and instead it shows office suite plugins. I would presume it affects the system prompt in the background?
Very confusing. But I do find it potentially interesting to treat general office work no differently from coding, which is something I had already been using Codex for in many ways before today
> That's assuming their flagship product remains relevant in an AI-powered world.
The big advantage Google has, in my opinion, is Android. I think there is a decent chance that people stop downloading the ChatGPT, Claude, etc. apps if they perceive that the phone just does the same out of the box for free. And I reckon the majority of people will prefer free, ad-ridden AI chat vs. paying subscriptions, at least for personal use. And on the B2B side, they have Workspace deeply embedded in a huge number of companies. So I wouldn't count Google out.
Not just harnesses, you can even use the subscription in CI/CD. That, plus the fact that web chat does not count toward the same limits, is why I think the Codex personal plan is easily 10x the value of Claude Code.
For work, I mostly use Codex and some Claude. For personal use, I’ve started using Chinese models directly through their respective providers, mostly for automation tasks and experiments so far, either via the API directly or through the Pi harness.
I do not trust any of them. Everything runs inside virtual machines, not just the sandboxes provided by the harnesses. I also do not run Claude or Codex directly on the host machine. Not just because of supply chain fears, but also because of how incredibly user hostile the VC funded companies are when it comes to installing random stuff on your machine.
I would say prompt engineering, in the sense of people claiming you need to include in every prompt magic incantations like "You are a senior engineer from a superintelligent alien species" and "take a deep breath and make no mistakes" doesn’t really do that much for everyday work I feel or they are all already included in the system prompt maybe. I reckon it can still edge out a few percentage points in automation.
What actually matters is the ability to communicate well in general, not anything LLM-specific. Being able to state what you want clearly and unambiguously, and having a sense for what additional information you need to dump, even when the other side claims they already have everything they need.
"I think we're going to start to see the title 'software engineer' go away. And I think it's just going to be maybe builder, maybe product manager, maybe we'll keep the title as a vestigial thing." — Boris Cherny
They been claiming more than just “coding is solved” for a while now.
> I'd love to see either Anthropic or OpenAI really step up their infrastructure game.
It's worth noting that OpenAI's official uptime numbers are significantly better than Anthropic's:
99.9x% for API/Codex, versus <99.5% for Claude API/Claude Code.
I'd obviously like OpenAI's numbers to be higher too, but this is one reason it really annoys me when the head of Claude Code goes on podcasts implying that software engineering as a whole, not just the act of writing out code, is basically solved.
One wonders why hasn't months of presumably near-unlimited internal Mythic solved the issues unrelated to hardware shortages yet.
I find it fascinating that every time this kind of discussion comes up, people talk about night and day experiences between Claude and Codex, in both directions. I’m really wondering what people are doing to get such different outcomes.
I’m currently working on two projects/clients one using Claude, one using Codex. I have a strong preference for the latter, but not because I think it is much more intelligent or writes much better code. It is simply because I find the way of interacting with it more pleasant: more literal, mechanical, makes fewer assumption and or double checks, and is less proactive in my experience. At least until some updates over the last few weeks.
I reckon right now the Enterprise concern is more FOMO around the AI wave and how to retrain or replace up to hundreds of thousands of employees. I don't think cost is the main concern right now.
But if AI doesn't lead quickly to vast large scale replacement of workers as promised, I could definitely see the C-suits and their gaggle of consultants starting to ask questions about token pricing.
In this case, handling all the edge cases and variants, and testing a codemod, would have taken significantly more of my time, which costs quite a bit more than the LLM.
Obviously, a deterministic tool is preferable in general, but it is not always worth bothering with for a one off task.
I had good experiences doing multi-hour refactoring/housekeeping tasks that basically consisted of applying the same steps and rules n times.
Worth noting, a significant chunk of those runs involved the agent waiting for the compiler, linters, type checks, and test suites, as well as updating journals.
It’s not the agent sputtering out code for eight hours straight.
And naturally I spend more time on manual verification in the end as much less of it is happening during the coding process.
> I do think that over the past few months, it feels like the hype around producing unmaintainable amounts of LoC has started dying down.
I wonder if a small part of this is more and more business and product people actually trying to incorporate AI into their daily workflows. I have seen this in both small companies I work for. People were very excited about getting Claude Cowork a couple of months ago, and while they use it daily, I would say they are rather underwhelmed compared to the magic they were expecting.
Complaints include the output being mediocre and verbose, it getting the most basic things wrong, hitting token limits all the time, and people going back to doing things themselves because it is faster.
Sure, there is some degree of holding it wrong in the beginning, but people are realizing that maybe, just maybe, there is still somewhat of a gap between what AI CEOs, LinkedIn grifters, and YouTube AI supplement peddlers claim and reality.
I don't know how effective the French protests are, since I haven't lived in Europe for a while. But even as a Swiss, at least judging from TV, protests in the U.S. generally seem very tame.
Not advocating punching the police as a default, but in my opinion, protests need to be disruptive if they're going to get anyone's attention at all. I don't really see what a few people standing on the sidewalk with cardboard signs are supposed to accomplish.
I think we too often treat other people’s jobs like spherical cows out of ignorance. Not just AI researchers.
Long before LLMs, programmers regularly and massively underestimated how hard it is to automate other people’s work. Knowledge workers often think carpenters just bang nails into wood, while blue collar workers think knowledge work as sitting in front of a screen copying values from Excel on the left into a form on the right while sipping a latte.
Only like 2.5 years ago, I thought programming would be one of the last knowledge worker jobs to be significantly affected by LLMs, not one of the first. I think AI models will continue to be very impactful. But for quite a while, they may mostly turn knowledge work into a last mile problem rather than eliminating it.
Earlier this week I started testing Chinese models on my codebase. I haven’t really looked at interactive coding yet, but more at issue triage, bug auto-fixing, log analytics, etc.
I used DeepSeek, Kimi, GLM, Qwen, and MiMO against GPT-5.5 high as reference, all running in Pi harness without anything installed.
So far, Kimi and MiMO look the most promising to me. I haven’t tested them rigorously enough to make a strong statement, but my first impression is that, in practice, all those models may be less behind on typical daily tasks than people think.
They are a bit “work hard, not smart". Getting to same-ish results more slowly and using more tokens, but at a fraction of the price
At least according to this, GPT-5.5 Cyber is on par with Mythic, as the only two models that were able to finish their 32-step corporate network attack simulation.
My worry is that, at least among the artists I know, many kept themselves afloat early career by doing commercial freelance jobs like illustrations for local events or companies. Those kinds of jobs might largely vanish.
On the other hand, with the internet inevitably becoming swamped by AI generated content, I can definitely see a de-digitalization of art moving into offline spaces. At least for independent work, you don’t necessarily need mass appeal or exposure, but rather access to individuals and small groups with an actual willingness to pay for art.
Having used Python on and off for 20 years, my experience with LLMs writing Python has been mixed. I don’t think that’s necessarily because of a low-quality dataset, but rather because Python’s applications are so broad and the language has gone through several paradigm shifts over time: sync vs. async, typed vs. untyped, scientific Python looking very different from web application code, some people really wishing it were an FP language, and others doing the clean-architecture OOP onion soup. It has gotten so fragmented.
Recently, I had a more pleasant experience using LLMs with Go. It reminds me a bit of Python 2.x, when the community seemed, in my view, more focused on embracing a stupid simple language, with everyone trying to write roughly similar "Pythonic" code.
Very confusing. But I do find it potentially interesting to treat general office work no differently from coding, which is something I had already been using Codex for in many ways before today