Hacker News new | past | comments | ask | show | jobs | submit | spacebanana7's comments login

Meta should actually go up from this. If deep seek is perfect they don't need to pay for an expensive Llama team. And even if deep seek isn't perfect, the low cost training strategies they've invented could be used by Meta to reduce the cost of Llama training.

Although Meta develops models they don't sell them. So a world where foundation models are free is fine for them.


But if they don't sell them, are they spending 50 billion a year on open source goodwill?

From Meta’s perspective, AI could be incredibly profitable in the context of generating adverts or interactive chat bots for businesses.

They just don’t want to use OpenAI/Google models because they fear being screwed over by them with anti-advert terms of service or price increases. Similar to what they suffered with Apple.


It's like everyone forgets that App Tracking Transparency (ATT) was supposed to put Meta out of business. By many accounts, Meta's ad targeting is even better now than before ATT. It's been reported that AI is what saved their ad targeting.

The OSS goodwill is just a side effect and a way to undermine companies who are not using AI to effectively make profits today.

Cheaper/more efficient is absolutely great for Meta. If they can lower their capex it would be an instant bump to their bottom line.


> By many accounts, Meta's ad targeting is even better now than before ATT.

Could you please provide any sources for this claim?


First link was soon after and they had already regained 80%+ back. The second link was much more recent and appears to be continuing to expand.

https://archive.is/8bRBH

https://medium.com/@omarkorim/is-meta-really-moving-beyond-t...


Thanks, but I don't really see how these articles support the claim that their ad network is more efficient. As you note, the first article has a single anecdata point about it actually being 10-15% worse, while the second one basically says 'trust me bro'. Also of note from the second article is the fact that the ad spend would actually increase.

Of course, if businesses are gullible enough to believe facebook when it fudges up some brand lift metrics without having a real impact on conversions, that's their choice. Trusting facebook to report any analytics is how you take your business behind the barn and help it pivot to video. https://en.wikipedia.org/wiki/Pivot_to_video


No, it seems to me like a stopgap measure against others, rather than anything for themselves. If they were out after "open source goodwill" they'd actually release the models as open source (like let people use it without signing a license/terms agreement, and use models for whatever). As it stands today, they're tricking people into "open source goodwill" but it will eventually catch up with them.

But now (presumably) they don't need to spend 50 billion? e.g. 5 billion or whatever might be enough which makes it even easier for them to justify this.

In the UK there's no significant regulatory advantage to big banks outside of the mortgage market, yet the same dynamics occur. The biggest issue for new digital banks is customer acquisition cost. Most consumers won't change bank accounts unless you spend hundreds of pounds on adverts and incentives.

I get a couple of cheques a year from family in the UK. It's an infrequent transaction but an important one, and cheque scanning is actually the only reason I maintain my legacy bank account.

> More than 2/3rds of my graduating physics class went straight into banking

This is why I advise people to never study physical sciences or non-software engineering. There just aren’t many jobs for it in the UK. And even fewer that pay well.


The usefulness of that depends on whether you think academic higher education is vocational ("I'm doing a chemistry degree so that I can be a chemist") or inquisitive ("I'm doing a chemistry degree because it's so heckin' interesting").

I'd tend to advise people to study stuff they find interesting. I'd wager the percentage of degree holders doing a job that's directly related to their degree is in the minority, and that's not a bad thing.


Inquisitive education is great for people who’re independently wealthy or academically exceptional.

But the average person takes a real risk of underemployment with that approach.

A friend of mine with a degree in marine biology works as a barista, another with a philosophy degree works as a retail assistant.


Those people don’t have those jobs because of their degree.

As a non-wealthy, not particularly bright creative writing major, the focus of your degree only holds you back if you let it, or if you are utterly unwilling to work outside of your focus area. Tech, especially, has an absurd number of on-ramps for anyone willing to do the tiniest modicum of extra-curricular work.


Unfortunately while it shouldn't be the case, education for education sake is regarded as a privilege for the wealthy.

Most are becoming educated and going into significant debt to get a specific job, or salary.


I don't live in the UK, but here in the USA I studied Physics and Music and then got a job programming. A lot of it was dumb luck, but I want to emphasize that school is about learning, not about vocational training. My experience in programming was due to being a giant nerd, not school.

I think another big difference is that in the UK higher education is not really valued. In Germany people get PhDs because then they can get into much higher paying jobs, even if the PhD is not in a topic related much to the job. For example because I have a physics PhD I was able to get an entry level software dev position with a wage of 73k where someone with a bachelors may get only 40-50k. I don’t think that dynamic exists in the UK

Unfortunately in 2025 you can probably add software engineering to the list of degrees to avoid.

At the moment on LinkedIn there are about 15k results for jobs containing the “software engineer” keyword in the UK, compared to just over 3k for the “biology” keyword.

Despite this, 56k people graduated in biological sciences but 24k people graduated in computer science.

(Graduation data is from 2019/20 so may have changed slightly, but unlikely enough to move the needle)

https://www.hesa.ac.uk/data-and-analysis/sb258/figure-17


The vast majority of jobs on LinkedIn are fake. Even so your figures confirm my claim; 24k graduates versus 15k jobs. The supply is greater than the demand.

As a sibling comment has pointed out, there are too many software engineers in the UK.


Only because there are too many software engineers, not the fact that AI will replace those jobs. Experienced software engineers are still required for successful businesses.

I do wonder how long this will hold up. It's true (though cliche) that software is "eating the world", but as the low hanging fruit gets automated away you do need people that understand more complex underlying processes to work on the software. I feel the right combination at this point is to do a little of both.

Them going into banking shows it is still a valuable degree,.

It’s slightly more nuanced than that. Investment banks and consultancy companies are really interested in graduates who are smart and articulate. The nature of their degree is not that relevant.

I know an extremely clever young woman who graduated in performance arts. She had no difficulty in getting a job at a top tier consultancy company. This company was far more interested in her than some mediocre plodder who hacked his way through a comp. Sci degree.


It’s not that these degrees are totally without value. Most banks and consultancies accept any quantitative degree as long as it’s from a prestigious university.

But if you know you’re going into banking anyway you might as well do something vaguely relevant like accounting, economics or statistics.


The very success of Northern England in the nineteenth century may have also precipitated a great deal of its subsequent failures.

The rail, coal and manufacturing industries crowded out other forms of enterprise economically. The related union / capitalist disputes crowded out other political activity.

Similarly the great agricultural “successes” of the Scottish Highlands in the 17th & 18th century crowded out everything else, leaving it something of a backwater ever since.

Perhaps something similar will happen to London when the time of finance comes to an end? Then finally the South West will take its turn being the heart of the British economy.


We'll have the same thing for the whole country when the time of housing investment comes to an end. Why start a business when you could just buy another house?

I’ve found LLMs extremely useful for Bible study.

If I’m struggling with a verse they can quickly show alternative translations and compare the interpretations of historical commentators.

Of course all this was possible with Google, but it could take an unreasonable amount of effort.


Doesn't the section where the author goes through the LLM completely fabricating or confounding different parts of scripture give you pause about the truthfulness of the responses?

How do we know the truthfulness of the original text that was written hundreds of years after the words were originally spoken?

It's asked about what a specific version of a text says and gets that wrong. It doesn't particularly matter how correct or carefully copied that scrap is, the machine can't faithfully reproduce or quote a particular version. In fact it made up a quote that doesn't exist in any version of the Bible the author could find at all.

a machine could, this machine cannot

I was more trying to add an interesting philosophical perspective than to comment on this particular instance


As we build and understand them now there's pretty good structural reasons to believe that LLMs cannot be tweaked or tuned to be more than incidentally truthful.

If you use an LLM only solution, sure, but there are many more opportunities if you expand the system to include more.

We could, for example, use the existing tools for copyright detection, on the output and refuse to send it to the client side. It's just another moderation check in that perspective


Copyright protection software is notably bad because it requires human context and decisions to decide infringement from fair use. Say we ignore that part that's the lowest hanging fruit on the improving generative models tree. Deciding if something is true to have the bot not say output it if it isn't is waaaay harder and the thing needed to fix the problem we've actually been talking about.

If you are asking that question seriously check out the recent Rogan podcast with Wesley Huff. Has the usual Rogan conspiracy theory pontificating, but also some good discussion about the history of how we got the Bible we have today.

First off, the claim that the original text was first written 100s of years after the words were spoken is completely false and isn’t taken seriously by anyone in the field.


I saw this post a while ago: https://benkaiser.dev/can-llms-accurately-recall-the-bible/

While LLMs are in generally fairly good at recalling bible verses, they can't do it perfectly. If the Bible truly is the infallible word of God that we believe, then shouldn't we use more caution than just "welp, sometimes it makes mistakes"?

You could counter this by saying a person can't remember Bible verses, and this is true, but a person usually recognizes when they can't remember something instead of making something up. If you asked me to recall any random Bible verses, chances are I wouldn't be able to do so. However, unlike an LLM, I would admit I don't know for sure and I would pull out a Bible or look online for an authoritative source rather than adlibbing something on the spot.


I’ve found verse recall to be a problem, but this can usually by mitigated by adding the verse to the prompt or pushing the LLM to search the web.

The main use case for me is interpretation, which LLMs are excellent at.


There is good venerable software that does exactly this. A lot of options including several open source ones.

Twitter was such a big deal because it digitised the social networks of analogue media. Pretty much anybody on TV did well on Twitter whether they made their name in politics, sports, news or reality TV.

Nobody else will ever be able to do this because those analogue social networks don't exist anymore in a way that's separate from Twitter.


Also that it's in the correct folder


I recommend trying alcohol free gin & tonic. It’s very competitive with ‘normal’ soft drinks like Pepsi & Fanta, especially if you’re looking for variety.


Amazon is excellent at selling physical books. I can order pretty much any vaguely popular book and have it delivered the next day at a price rarely higher than anywhere else.

That’s Amazon’s core business philosophically, everything else is an add on or side project that happened to be profitable.

I think that just like the original sin of web development is trying to run apps in a document browser, the original sin of Amazon is trying to sell everything in a bookstore.


Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: