Their point is that Claude is now engineering its responses to specifically conform to a detectable pattern, something that can of course be easily removed via a seperate LLM.

Zedstrian
Moved over from Reddit after the API debacle. Primary account history:
@Zedstrian@kbin.social (2023) @Zedstrian@lemmy.dbzer0.com (2023–26) @Zedstrian@sopuli.xyz (2026–)
AI agents have no agency of their own; their operators are the ones with that control. Negligence in giving an AI agent access to the internet does not absolve the operator of liability.
Replying to @mereo@piefed.ca
More OEMs should ship devices with Linux as an option, discounted by the Windows license cost.
Replying to @NoneOfUrBusiness@fedia.io
The entire AI industry is built on IP theft; even if you don’t believe in IP ownership, companies like OpenAI and Anthropic rely on their ability to scrape the internet for content, with only largest of publishers having the resources to sue for anything in return.
So long as copyright is being enforced at the individual level for consumers, large corporations shouldn’t be allowed to get away with infringing on it for profit.
Replying to @NoneOfUrBusiness@fedia.io
It’s one thing for individuals (particularly those without the means of paying) to pirate content that its original creators have already been compensated for and now exists to milk money for an IP holder for a century; it’s another matter entirely when AI companies with multi-billion dollar valuations scour the internet for content to include in their models to then profit from without permission from or compensation to the millions of people whose work they depend on.
Replying to @FoxtrotDeltaTango@sh.itjust.works
Whoever you are, don’t outsource your understanding, judgement, empathy and taste to AI. Don’t abdicate your responsibility. Don’t be a meat proxy.
By expanding the ability to program to people without the knowledge of how that code works in the first place, it’s naive to expect that most uses of AI in programming will be informed uses by senior developers able to recognize errors and security flaws.
Such a pro-AI position also ignores the enormous amount of theft that these tools are built on.
Replying to @beep@piefed.world
AI bots cannot be and should not be licensed mental health care practitioners, so existing licensing laws should be enforced.
Replying to @nucleative@lemmy.world
Maximizing shareholder value also includes maximizing the non-tangible value of the Roblox brand (something that could be harmed by short-term cash grabs), but unfortunately many investors only think one quarter at a time.
Replying to @beep@piefed.world
Reddit’s main strategy now seems to be injecting AI into everything and hoping investors go “wow!!!”.
Replying to @y0kai@anarchist.nexus
The problem with many of the articles that discuss how the system is being ‘misused’ act as if the mass surveillance involved is otherwise entirely acceptable, so long as it’s used for its intended purpose.
Replying to @Nobody_Special@piefed.social
Companies won’t stop doing this; they just won’t be admitting to it as openly.
Replying to @sanitation@lemmy.today
Protectionists complain about Chinese market dominance, yet Chinese companies are the only ones entering the DRAM market in response to the dramatic increase in demand.
Replying to @6_Electrons@lemmy.world
Given that the fossil fuel industry receives an annual subsidy of $35 billion (not including the negative externalities that the industry is allowed to get away with), denying comparable subsidies to renewable energy projects encourages investors to invest in fossil fuels instead, stunting renewable energy growth and harming the environment in the process.
Replying to @6_Electrons@lemmy.world
Solar installs are increasing because they’re the most economical means for power generation when substantial cheap land is available, but major wind projects have been cancelled at the behest of the Trump administration.
If Trump had not been elected, renewables would be growing even faster.
Replying to @6_Electrons@lemmy.world
Their pace has been slowed down greatly under the Trump administration.
Replying to @FoxtrotDeltaTango@sh.itjust.works
The fact that an organization with the budget and security requirements of the U.S. Military wouldn’t develop on-premise AI demonstrates just how much grift exists in the military industrial complex.
Replying to @beep@piefed.world
Combined with the fee increase to 10%, this seems like an effort to boost profitability for an IPO.