>Starting September 14, we're permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans. Until then, the current 50% increase will be in place.
> Compared to today, this works out to a 17% reduction in weekly limits on Claude Code. We’re working on exciting changes that will make it feel like you’re getting more from Claude, while having more visibility and control of your usage. Can’t wait to share them.
They are both raising the limits by 25% and apparently reducing them by 17%. I think they mean to say you can do more, but what a terrible press release.
I'm pretty sure they mean going from 150% to 125%, where 1.25/1.50 = 0.83, so they're calling that a 17% reduction. It's less than today, but also more than a limit they made up and then didn't apply.
Exactly this. You still have to understand the abstraction layer the LLM is working at in order to be able to form an opinion about the quality of the code. (If you DON’T care about the quality of the code, there’s always been ways to get an extra 1000% short-term speed improvement by just not giving a shit.)
AI isn’t “a code abstraction layer” any more than Jira is, or the engineer themselves is.
I think "the engineer themselves" is much closer to the abstraction layer I'm referring to.
Imagine for a moment that, rather than using a compiler to translate c++ into assembly, that it instead has to be done by a person on the team. If that were the case, the resulting assembly code would certainly have to be reviewed and assessed before it was accepted.
Why?
Because people can also be leaky, non-deterministic abstraction layers. The only reason that the output of a compiler isn't regularly reviewed (in 99% of cases) is because it's extraordinarily reliable and consistently correct, or correct enough for most cases.
It's not terribly dissimilar from delegating engineering tasks to other engineers. When I ask someone else to develop one component of a larger application, I'm not telling them exactly what lines to write, I'm giving them some kind of structure and they're filling in the rest. The communication to them is above the implementation layer.
Agents are increasingly letting us work at that same boundary. Just because the current state of LLMs requires engineering knowledge to review the result doesn't mean an abstraction hasn't occurred. It speaks more to the current quality of that abstraction than the absence of one.
I think that if LLMs were able to achieve the same consistency as a compiler, most people wouldn't ever bother to check the underlying code it produced. I also think it would be difficult to not acknowledge that LLMs have gotten better at converting natural language into functioning software. The abstraction is certainly not perfect, but it is clearly improving.
Damn, can you imagine a world where it would be so easy to convince one of these companies to stop poisoning the minds of an entire country like that?
Like, I kinda jest, but also kinda don’t. Would that really be so bad of an outcome? The net positives don’t really seem to bear up under all the negatives just… going away.
Yeah, but social media won’t be the only companies that notice. If Australia’s government will do it to them to prop up their local dying industry, they will do it to you, too. If the EU or US tries that, well, those are huge markets and maybe you bend (maybe you don’t!). But Australia is 27 million people, less than 10% the size of those. If they want to get heavy-handed, well, it’s their country, but as recent US actions have shown, other countries’ companies may pick up their ball and go home rather than play with you.
As an Aussie expat who uses Facebook to keep in contact with my friends and family: yes, it would be bad. Social networking tools have real positive effects on the world, let us not forget that.
It's exactly this - "coding was never the hard part" translates more directly to "coding was never was slowing me down". It's nailing down requirements, it's cross-functional team meetings, it's planning the testing and rollout, it's integrating with all the other parts of the product and systems, etc etc etc. Optimizing coding was just aiming at the fastest part of the job already.
So, people are using AI to answer questions about tribal knowledge? Is that really the assertion which is being made here? Because I find that hard to believe.
So, you have different roles and specialties, but you know for a fact that this person should know the answer to your question. I think we're speaking past each other, because this doesn't make sense to me.
I don't presume that someone will have an answer I'm looking for. I'm pleased when they do, but often I just get breadcrumbs, and I have to solve the puzzle for myself.
I'm not GP, but in research very diverse teams are not uncommon. I'm on a team now in which I'm the embedded software engineer / Android SME and my coworker is the 4G LTE/5G NR SME.
I ask him 4G LTE/5G NR questions all the time. I don't know ahead of time if he'll know the answer, but (1) it'd be a solid bet that he does and (2) it's his responsibility to know it given his role.
If you have a semi-recent MacBook with even 32GB, you can run 20GB models that are pretty damn smart, with room to spare for the rest of your toolchain.
If you’re reasonably connected to the code you’re writing and prompting the AI at the level of the code, not the level of the feature, you can get some fantastic results.
Sure, it’s not the completely automated dreamland that’s been sold, but it’s still a speed-up on par from going from assembler to a higher-level language, which is still immense.
And for effectively free, if you have a machine that would otherwise have been considered “developer-grade” for a lot of tasks anyway.
I tried this on my work M4 with 48 GB of memory and it's not as useful as I had hoped.
If you set a large context window you end up using all your memory. Prefill is slow. And the 10-20B models are somewhat poor at understanding intent and tend to think in circles, so if you set a small thinking budget to control context size the model ends up answering based on some half formed thoughts.
Or something like batch document summarization or image classification or whatever yeah it's great. But for "agentic" and coding workflows I think you would need a completely different harness setup and prompting style compared to what we are used to from the big proprietary models. I certainly have not yet figured out what that is.
What models have you found to work for which tasks? I find the local models very useful but not in a way that's replaced cloud models (yet, I remain hopeful).
I would say that the more-exhausting way to live would be to be unaware of this, to accept the “we’re a family here” lie at face value, and to be passed over and screwed over at frequent intervals without knowing or understanding why.
You don’t have to go full Patrick Bateman corp-bro, but keeping in mind the phrase “we don’t fire family” and adjusting your approach based on that counter-example can help you to keep things appropriately transactional and not be taken advantage of.
The least exhausting way to work a job is to find people to work with who are smart, well adjusted, and genuinely interested in cooperating to make money together.
reply