Claude becomes number one app on the U.S. App Store
apps.apple.com
apps.apple.com
What I'm getting at is that this could be, in part, because Claude is genuinely better at this point in time.
Sam squandered it.
Codex 5.3 Xhigh > Opus 4.6 in my work to this point.
Hoping for Opus 4.7 or whatever comes next to rectify this as I'm a bit annoyed over having to drop to a lower quality model.
But for the chat, I feel like ChatGPT got worse and worse.
Unless I specifically say "use git", it won't bother using git, apparently saying "configure AGENTS.md to us best practices" isn't enough for it to (at least in this case) use git. If this was isolated I might put that down to bad luck, given the nature of LLMs, but I have been finding Codex uses the wrong approaches all over the place, also stops in the middle of tasks, skips some tasks entirely (sometimes while marking them as done, other times it just doesn't get around to it).
I'd rank the output of Claude as similar to a junior with 1-3 years experience. It's not great, but it's certainly serviceable, a bit of tweaking even shippable. Codex… what I see is more like a student project. Or perhaps someone in the first month of their first job. Even the absolute worst human developers I've worked with after university weren't as bad as Codex, but several of them I'd rank worse than Claude.
Also, I noticed it skips instructions if I steer it with prompts while it is doing stuff instead of queueing my instructions.
OpenAI's 5.3 Codex model on xhigh still makes a huge number of mistakes, somewhere between 25-50% of commits, and it's still terrible at making its own plan, estimating how long tasks will take to complete, and recognising which tasks need to be subdivided*. Claude's model last November was better on both counts, even though it still wasn't IMO ready for true lights-off-no-code-check-needed-vibe-coding, it was making mistakes far less often and was scoping task complexity appropriately.
That said, given xhigh seems to be going through my token allowance far, far slower than on medium, I wouldn't be surprised if it turns out the Codex app itself is vibe coded and has mis-mapped that setting in some weird way. Either that or they've suddenly got a lot more spare capacity because of the boycotts.
* given the METR study, in the planning phase I ask all these models (Codex and Claude) to break down tasks into things that would take a junior developer 1-2 hours, but Codex will estimate 60 minutes for everything from "write 19 lines including comments to stub 3 empty methods in new class" to tasks I'd expect to take a senior 2 days.
I remember on the Opus 4.5 release data watching what it can do to my test app I wanted it to build and saying outloud to myself "oh shit" because of how much better it was at the conversation, planning, understanding, and building. Posts like this[0] say similar things, where Opus 4.5 release + Claude Code was the tipping point and the gap is widening and Anthropic has infinite more momentum and going in the better direction with useful models that aren't fully aligned with bad actors.
Then I remembered that the app store rankings [1] seem to be based on activity from just the past day or so.
And so a lot of "plugged-in" users switching to Claude all at once then would be enough to briefly send Claude to #1, since the migration would be sizeable in comparison to the normal daily download baseline.
But we can also expect that this would probably be just a blip for a couple of days, as it's unlikely to make much difference in the baseline ratio for the general population.
She said yesterday and Friday her instagram was full of people deleting ChatGPT and downloading/paying for Claude.
It’s definitely beyond the tech bubble.
Then again, I know nothing about ads these days. Last time I had any ad exposure was twenty years ago, when I lived at home with my parents.
Though, I happened to be watching the Superbowl, and Anthropic was all in taking pot shots at OpenAI's recent decision to start incorperating ads. Very much addressed towards a general audience. Here's one of them for example: https://youtu.be/FBSam25u8O4
When we created the account we shared essentially zero personal information and used a purpose-specific email, etc. No ID scans, no passports, no robust ID verification. It is a company account, after all, so it would make no sense to tie it to a particular person ...
All of this to ask:
In 2026, if I choose to boycott openAI on behalf of my firm, and perhaps sign up for Claude with the intention of using either the web interface or a terminal CLI ... what will be demanded in terms of identity and personal information ?
And if you want to use the API? OpenAI requires full identity documentation upload and potentially selfies sending to Persona (the exact thing Discord was just chastised over) if you want to use a remotely recent model. Anthropic doesn't ask for any of this - and neither does any other model provider.
Meanwhile I have a working desktop app but this magic email link nonsense doesnt seem to work on my phone.
I won't go back.
Anthropic currently is winning the public sentiment war, but OpenAI is being established as another to-big-to-fail capitalist regime institution.
american eagle sales are one thing, Sydney's image and PR are another...
Does Iran have a moral obligation to not nuke the global population? What about Israel and Palestine?
But privacy? Yeah, no, if it benefits the safety or interests of US even slightly to know what the world is thinking, seeing, and planning, I have zero qualms using all tools at our disposal.
The prevalence of data and a general sense of powerlessness means this won't get talked about much sadly.
> The Pentagon has agreed to OpenAI's rules for deploying its technology safely in classified settings, though no contract has been signed, a source familiar with the talks tells Axios.
> Why it matters: The Pentagon has blasted OpenAI rival Anthropic for days, contending its red lines for AI use in the military -- mass surveillance and autonomous weapons -- are philosophical and "woke."
> Now, the department, which did not immediately respond to a request for comment, appears to have accepted OpenAI's similar conditions.
One big difference
OpenAI exec becomes top Trump donor with $25 million gift [1]
I guess that makes it "less woke". This is reprehensible political bullshit.
[0] https://archive.ph/9NcMf#selection-579.0-611.135
[1] https://www.sfgate.com/tech/article/brockman-openai-top-trum...
Where was the open letter then? No-one cared.
Recently, Claude was already used by the administration for the operation in Venezuela [1] alongside Palantir. Anthropic did nothing at the time and again...
No-one cared.
Now everyone cares when Anthropic finally said No? The decision for the contract was already predetermined for OpenAI, even with the open letter.
So the question is, why wasn't the open letter against OpenAI done last year when they signed that first military contract?
Either way, it seems that OpenAI and Anthropic were all OK with the US government using their models for warfare so really there is no point in defending both of them or even the employees who knew beforehand.
[0] https://www.theguardian.com/technology/2025/jun/17/openai-mi...
[1] https://www.theguardian.com/technology/2026/feb/14/us-milita...
DoD started asking for the ability to do more stuff. That's the issue here. "Two such use cases have never been included in our contracts with the Department of War, and we believe they should not be included now: Mass domestic surveillance... Fully autonomous weapons." https://www.anthropic.com/news/statement-department-of-war
>So the question is, why wasn't the open letter against OpenAI done last year when they signed that first military contract?
again, this story isn't about people that are against any military contract.
The DoD (NSA) has crossed those lines before with big tech when they allowed mass domestic surveillance on their own citizens (PRISIM) in the past. They are willing to break laws to get the job done which probably in 10 years time those actions would be found illegal later. By then it would be too late.
Companies should expect that governments may try to bend or even break rules or laws under veiled pretenses.
So why expect that they would be any different today? It is the nature of governments.
Scorpion (DoD) tells frog (Anthropic) to help them cross the river on the condition that they won't kill them. If you knew that the Scorpion always breaches the contract first, why work with them in the first place?
> again, this story isn't about people that are against any military contract.
Gen AI mass surveillance conducted by the adminstration (or any) is already done by Google, Microsoft, AWS, Oracle and Palantir and soon xAI (via X). So again, no point is being made here.
Given all the above, the clear indication was OpenAI already signed a military contract last year. Why didn't the employees and insiders make an open letter pledging that AI should not be used for mass surveillance back then?
some companies do it, so that means anthropic has to do it? this seems like nihilism
Then I would have expected open letters from anon OpenAI employees from the very beginning in 2025 when those military contracts were signed as a reassurance / pledge around those certain boundaries since they ultimately knew. But of course, only after Anthropic rejected the DoW in 2026. Very late for that.
Anthropic should've known about PRISIM and the nature of governments and should have not tested the benefit of the doubt in the first place since it's directly incompatible with their 'principles'. Otherwise none of this would have happened.
Their first mistake was trusting the government (especially this one) and like almost all of them, they are ready to test what they can get away with and breach the contract first.
Anthropic naively expected that the administration (and any other) of this time would change for the better today. Their lesson is that they should never trust governments to agree to their contracts.
So for all intents and purposes the DoD can simply ignore OAI's redlines.
Is doing a lot of work here
This a nothing burger
Most people just don't care.
Sam thought he was special. He is not.
The DoD office staff using it for information analysis, similar to Palantir providing data integration software, isn't quite the same as using it for weapons.
What do you think writes the code for that and trains the models?
They are not opposed to autonomous weapons in principle and surveillance on foreigners is fair game. The only moral stance is to use no AI at all (it’s not that hard)