I'll second this. An external recruiter was under the (incorrect) impression that we are a 996 company. We found out because she said that no senior people she talked to were willing to work those hours.
Ultimately you can make a lot of short-term progress with 23-year-olds who are willing to live 5 minutes away from the office, have no life outside of work, and work 72 hour weeks. But you also end up with a product that was built by people who have no idea what they're doing.
I am only seeing that if the person writing the prompts knows what a quality solution looks like at a technical level and is reviewing the output as they go. Otherwise you end up with an absolute mess that may work at least for "happy path" cases but completely breaks down as the product needs change. I've described a case of this in some detail in another comment.
> the person writing the prompts knows what a quality solution looks like at a technical level and is reviewing the output as they go
That is exactly what I recommend, and it works like a charm. The person also has to have realistic expectations for the LLM, and be willing to work with a simulacrum that never learns (as frustrating as it seems at first glance).
I'm trying to work with vibe-coded applications and it's a nightmare. I am trying to make one application multi-tenant by moving a bunch of code that's custom to a single customer into config. There are 200+ line methods, dead code everywhere, tons of unnecessary complexity (for instance, extra mapping layers that were introduced to resolve discrepancies between keys, instead of just using the same key everywhere). No unit tests, of course, so it's very difficult to tell if anything broke. When the system requirements change, the LLM isn't removing old code, it's just adding new branches and keeping the dead code around.
I ask the developer the simplest questions, like "which of the multiple entry-points do you use to test this code locally", or "you have a 'mode' parameter here that determines which branch of the code executes, which of these modes are actually used? and I get a bunch of babble, because he has no idea how any of it works.
Of course, since everyone is expected to use Cursor for everything and move at warp speed, I have no time to actually untangle this crap.
The LLM is amazing at some things - I can get it to one-shot adding a page to a react app for instance. But if you don't know what good code looks like, you're not going to get a maintainable result.
One wonders if a German in 1600 would have cursed the invention of the printing press. The printing press accelerated the reformation, which led to over a century of bloody religious wars. Something like a third of the German population died as a result. From the perspective of 2025, the printing press was undoubtedly positive for humanity. But millions of people suffered.
I think this is something that a lot of supporters of the Gaza protests tell themselves, but I am not sure that it's actually true. The US and other Western countries sell weapons to Saudi Arabia and have extensive economic ties to that country. Saudi Arabia recently engaged in a bombing campaign in Yemen that looked very similar to Israel's campaign in Gaza. And yet there were no protests. Also, you can influence your country's policies towards another country whether or not the two countries are allies. Years ago, there was a mass protest movement in the west against the genocide in Darfur for example. Nobody said "we don't have a lot of economic or diplomatic ties to Sudan so there's no point in protesting".
I think the real reason has to do with 1) there was an existing, organized pro-Palestinian movement that had experience protesting; 2) many organizations on the left saw the Israel-Gaza conflict as fitting very nicely into their larger anti-imperialist ideology in a way that other conflicts don't; 3) everyone more-or-less knows where Israel is on the map and has some familiarity with it; 4) there were a lot of really shocking images and video from Gaza
Fair enough, I did not know that. Maybe add to your list of reasons that attention is divided over so many conflicts nowadays. Probably there have been conflicts all the time, but with Ukraine, Greenland, Minnesota, Gaza and Venezuela getting a lot of attention it feels like a lot. Note that I don't think the conflicts are remotely comparable with each other, but they each take up a lot of mindspace at least for me.
If you look at some of the most active groups in the pro-Palestinian left, like the PSL in the United States, you'll see that they have a very long history of praising horrific, oppressive regimes (even North Korea!) that oppose the United States, and dismissing accusations of crimes against humanity when perpetrated by those regimes. The PSL is a minuscule political party, but they're highly involved in organizing these protests.
If the US alliance with Israel is the reason why this conflict generated so much protest activity, then why didn't the pro-Palestinian left object to US ally Saudi Arabia's bombing campaign and blockade in Yemen? The US arms the Saudis. Much of what happened in Yemen is very similar to what happened in Gaza (airstrikes that hit civilians, hunger caused by blockading imports, etc)
And there have absolutely been examples of mass protest movements against regimes that are hostile to the US that are committing crimes against humanity. Years ago I went to a huge demonstration about the genocide in Darfur on the national mall in Washington. Raising awareness of what is happening and putting pressure on the Iranian regime (and on Western governments) can have an impact regardless of whether or not the West is hostile to Iran.
I'm assuming that the goal of the bloom filter is to prevent the model from producing output that infringes copyright rather than hide that the text is in the training data.
In that case the model would lose the ability to provide relatively brief quotes from copyrighted sources in its answers, which is a really helpful feature when doing research. A brief quote from a copyrighted text, particularly for a transformative purpose like commentary is perfectly fine under copyright law.
Ultimately you can make a lot of short-term progress with 23-year-olds who are willing to live 5 minutes away from the office, have no life outside of work, and work 72 hour weeks. But you also end up with a product that was built by people who have no idea what they're doing.
reply