NYC clearly doesn’t have the level of activity in this area that the Bay does, but there’s a scene. LeCun and the NYU crowd and the big FAIR footprint create a certain gravity. There’s stuff going on :)

Me: On your first day of work, what would you say to me if I asked you to do the samething in your writeup but in production and at scale using the same third party services (chatgpt etc) you used in your writeup.

T: "You're an idiot. I quit."

Me: You're hired!

Clearly it's a fun exploratory excercise, kind of like using kafta instead of a db for the main store for a crud app just to see how it works. But if you asked a senior engineer who follows the sector he would probably guess all the answers correctly blind. Tamer himself says it was his most expensive sunday hobby night. Now scale that uselessness to enterprise level. And you're not even sure if some of the results are hallucinated.

Anyway, I have a tangential question, and this is the first time I see langchain, so may be a stupid one. The point is the vendor-API seems to be far less uniform than what I'd expect from a framework like this. I'm wondering, why cannot[0] this be done with Ollama? Isn't it ultimately just system prompt, user input and a few additional params like temperature all these APIs require as an input? I'm a bit lost in this chain of wrappers around other wrappers, especially when we are talking about services that host many models themselves (like together.xyz), and I don't even fully get the role langchain plays here. I mean, in the end, all that any of these models does is just repeatedly guessing the next token, isn't it? So there may be a difference on the very low-level, there my be some difference on a high level (considering different ways these models have been trained? I have no idea), but on some "mid-level" isn't all of this utlimately just the same thing? Why are these wrappers so diverse and so complicated then?

Is there some more novice-friendly tutorial explaining these concepts?

[0] https://python.langchain.com/v0.2/docs/integrations/chat/

The APIs between different providers are actually pretty similar, largely close to the OpenAI API.

The reason to use a paid service is because the models are superior to the open source ones and definitely a lot better than what you can run locally.

It depends on the task though. For this task, I think a really good small model like phi-3 could handle 90-95% of the entries well through ollama. It's just that the 5-10% of extra screw ups are usually not worth the privilege of using your own hardware or infrastructure.

For this particular task, I would definitely skip langchain (but I always skip it). You could use any of the top performing open or closed models, with ollama locally, together.ai, and multiple closed models.

It should be much less than 50 lines of code. Definitely under 100.

You can just use string interpolation for the prompts and request JSON output with the API calls. You don't need to get a PhD in langchain for that.

I'm totally new (and maybe somewhat late) to the domain, literally just tried right now to automate a fairly simple task (extracting/guessing book author + title in nice uniform format from a badly abbreviated/transliterated and incomplete filename) using plain ollama HTTP-API (with llama3 as a model), but didn't have much success with that (it tries to chat with me in its responses, instead of strictly following my instructions). I think, my prompts must be the problem, and I hoped to try the langchain, since it somehow seems to abstract the problem, but saw that it isn't supported for a workflow the OP used. But since this is a field where I'm really totally new, I suppose I also may be making some more general mistake, like using a model that cannot be used for this task at all. How would I know, they all look the same to me…

Ollama project itself is fairly stingy with explanations. Doubtfully there are many people out there trying to automate an answer to the "Why is the sky blue?" question.

So, I wonder, maybe somebody knows a more digestible tutorial somewhere, explaining this stuff from the ground up?

2. Use the best/largest model possible. Small models are generally stupid. phi-3 might work as an exception of a very well trained tiny model. Very large models are generally dramatically smarter and better at following directions.

3. Tell it to output JSON and give it examples of acceptable outputs.

4. The API for OpenAI and Anthropic is very very similar to ollama. The models are vastly better than llama3 7b. You can basically make some minor modifications and if you have the temp right I bet it will work.

Personally I think that langchain will just make it more complicated and has nothing to do with your problem, which is probably that you used a tiny rather dumb model with a higher than optimal temperature and didn't specify enough in your prompt. The biggest thing is the size and ability of the model. Most models that will run on your computer are MUCH MUCH stupider than ChatGPT (even 3.5).

In other words, if the text you want was not in the model training data, a higher than normal temperature may be required, depending on how frequently the term appears in the input data. If you provide a few samples in the input, then you may be able to use 0 again.

The problem with generating structured output like JSON is that temperature > 0 also increases the likelihood of a token belonging to the set of “wrong” answers being chosen. With prose that’s not the end of the world because subsequent tokens can change the meaning. But with JSON or code, the wrong token in the wrong place can make the output invalid: it’s no longer parseable json or compilable code. In the blog they were also generating bools in one spot, and temp > 0 would probably result in the “wrong” answer being chosen sometimes.

For that reason I’d suggest generating JSON fields independently and then create the full JSON object from those outputs the old fashioned way. That way different fields can use different temperature settings. You’d probably want temperature=0 for generating bools/enums/very short answers like “New York”, and temperature > 0 for prose text like summaries or descriptions.

And, as stated, I also don't know the answer to that question, so this is kinda one of the primary concerns here. I mean, one possible answer seems pretty obvious to me: it would be better to keep your app vendor-agnostic (to be able to switch from OpenAI to Anthropic with 1 ENV var) if at all possible. Neither of articles and doc-pages I've read in the past few hours tries to answer to what extent this is possible and why, and if it even is supposed to be a real selling-point of langchain. TBH, I still have no idea what the main selling point even is.

https://github.com/m3at/hn_jobs_gpt_etl

Only using the plain OpenAI api. This was on GPT-3.5, but it should be easy to move to 4o and make use of the json mode. I might try a quick update this weekend

It gets picked by people with more of a top-down approach maybe, who feel like adding abstraction layers (that don't abstract pretty much anything) is better. It isn't

What the author could have done, and what I should have (but didn't) also, is add a bunch of possible values (enums) for each possible field value. This should solve it from coming up with variations e.g. node, nodejs

In zod/tooling it would look like this; remote: z.enum(['none', 'hybrid', 'full']), framework: z.enum(['nodejs', 'rails']),

But this just shifts the problem further down, which is now you need a good standard set of possible values. Which I am yet to find, but I'm sure it is out there.

On top of that, I am working on publishing a JobDescription.schema.json such that the next time the models train, they will internalize an already predefined schema which should make it a lot easier to get consistent values from job descriptions.

- Also I tend to forget to do it a lot recently in LLM days but there are plenty of good NER (Named Entity Recognition) tools out there these days, that you should run first before making robust prompts

The training data or some kind of enrichment of the data would have to make the systems understand node.js and nodejs are the same just like on the new notes calculator Apple-sauce = $2.50 * 8 makes the first statement a variable.

LLMs aren't reliable for actual number crunching though (at least from what I've seen). Which is why I really appreciate seeing this blend!

Join the club :) I'd say that's a pretty mainstream idea right now.

Exciting nonetheless

I personally would like AI to help humans be better humans, not try to replace humans. Instead of using AI to create less understanding with black box processes, I'd rather it help us with introspection and search (I think embeddings[1] is still a pretty killer feature that I haven't heard much noise about yet).

[0] https://news.ycombinator.com/item?id=21786547 [1] https://terminusdb.com/blog/vector-database-and-vector-embed...

As an added detail regarding the "remote" v "in-person", another interesting statistic, to me, is to know how many of those in-person job-seeking companies are repeats! It could absolutely mean they're growing rapidly, OR it could mean they're having trouble finding candidates. Equally, missing remotes could mean either they're getting who they need OR they're going out of business.

All interesting plots on the graph!

That is, until December last year. Then it was like a flood gate had opened up. Suddenly we had tons of candidates to try to select from.

I knew things were bad over in bay area and such, but I didn't expect it to hit the market over the pond here so quickly and abruptly.

Roughly, BLS accurately reflects government employment and then extrapolates to the private sector using a census model that can be several years old. ADP is private sector data and doesn’t reflect government employment. They are essentially sampling disjoint employment pools which leads to some known artifacts.

Sometimes there are large spasms of government job growth through expansion of spending and jobs programs when there is no matching growth in the private sector and the private sector may actually be declining (as seen in ADP data), yet BLS will extrapolate that the private sector must have similarly grown based on the government growth data. Similarly, ADP is generally considered a reasonably accurate model of private sector job growth but doesn’t account for government employment even when it is substantial. ADP can significantly exceed BLS in boom markets.

A heuristic that has been around forever for tracking “productive employment” is that the job market is good when (ADP - BLS) is a positive number and bad when it is negative. Currently, this heuristic is tracking at something like -100k to -150k, which indicates a poor job market.

> there’s an argument that they may be a bit overstated

Not one, not two, but three qualifiers[1] that you left out of your retelling.

[1] “there’s an argument,” “may be,” and “a bit”

The Fed chair’s statement you cite doesn’t support it. He’s just saying that the jobs numbers may not be perfectly representing the actual job market. That is not exactly news; it’s a very hard thing to measure. He’s not accusing the BLS of lying.

Beyond the in-office and the N-days-a-week-hybrid, you have within _actual_ remote roles:

- Country remote (mostly for taxation/regulatory)

- Time zone remote (remote first companies but constraint to within 2 or 3 hours of HQ time zone)

- Anywhere remote (actual remote but often as a contractor or EoR)

People needing freelancers for few weeks/months to complete projects where the requirements are glueing the usual APIs and solving the usual Safari bugs asking me Leetcode questions are out of their mind.

I am not applying for a full time position, I'm not a cofounder that is going to make or break your startup and no sorry I am not sharing your vision/mission whatever.

Discord is by far better for finding work in your domain and related to technologies you like, and you can ask for much more money because people already know you're experienced on the topics you share on that discord.

Another good source of work are local meetups/communities chats.

I've tried to interact on a variety of gamedev discords, and the results are about as dry as LinkedIn. But I suppose probing based on the 2020's market won't garner typical results. (still, open to checking out any suggestions. Far from a census here).

Money, though... ha. Less money and usually very few VC's so you're taking a hit compared to trying to grind interviews with EA/Activision. But that's games for you.

Pick a framework or library, get involved.

Used to be very specific instructions about mailing a resume to an address with a reference number. And advertised only in the newspaper. But Immigration said they can't do that one anymore; has to be the same submission methods (email, webform, whatever) as an actually open position and advertised/listed in the same places too.

But they'll still have the other tells, which is very specific experience and education requirements which happen to line up exactly with their preferred candidate. Sorry, we did our best, but we can't find any local candidates with a 4 year BS degree, a minor in Clown Studies, and 3 years experience with very specific software that isn't used many places (experience most likely obtained at the hiring company during internship or while on OPT; or while on H1-B if this is in support of a green card, rather than in support of H1-B).

The HN job board is much more likely to produce a qualified candidate.

If they don't suck, the guy fails the labor certification and may have to wait a year to apply for the green card again. However, he doesn't lose his job, and the qualified candidate never gets an interview.

So yes, it's pretty standard for the companies to post openings they don't have.

But I really doubt they post on the HN jobs board. Most will go through more formal channels.

Or any public sector job where positions have to be posted but you already have your internal hire, son-in-law, or cousin lined up.

But, it's mostly just rumor I've heard.

We did have many work visa employees though.

We worked with Healthcare clients. Big ones.

I’d love to see a similar analysis to “Who Wants to be Hired”. What trends exist in folks struggling to find work? That can help point people to how to target their career growth.

Been there. Hackers will do what hackers will do.

1. Awesome new tech,

2. Tech gets more and more popular,

3. Tech gets more and more features to better cope with new use cases,

4. People get sick of bloat and create awesome new tech to replace it,

5. Go to 1.

In 5 years time earnest conversations will be had about trying to find somebody who is willing to maintain the thing, or do we have resource for a complete rewrite?

Also seems a bit silly to include the JS runtime itself in comparison with the JS frameworks. In fact it is almost 100% of the time used along with a front-end framework/library at least on a build-level. I would not put a ton of trust in that frameworks demand data ;)

My vague memory of this is that AngularJS from v1 to v2 was such a huge difference (and arguably some percentage of the community qualify this update as a colossal f up), that they decided to go with a different name, and they also came to their senses and their backwards incompatible changes are no longer so bad.

I only used OpenAI to determine location and then some (failed) analysis of candidate’s HN comments. The rest of the tools I wrote didn’t need LLMs (just string matching).

Nit: There appears to be both MongoDB and Mongo in the database graph.

Also, could you do a write up on how you actually used GPT-4o to assist you in this exploration? Or share some of the main prompts you used? Thanks!

Chat GPT seems to have statistical awareness of this fact, yet people generally choose not to figure it out

How we can have a world where this statement exists, and anyone is bullish about LLMs?

The lack of precision is appalling. I can see how a gpt-4o can be fun for toy projects, but can you really build meaningful real-world applications with this nonsense?

Once bank of america and every other banking app hooks up one of these to my accounts, i better not have to dispute a charge with "HAS TO BE FRAUD, REALLY!"

I think bytes is generally a better metric than number of entries, naturally 10,000 books will have a lot more insight than 10,000 one-liners.

I'm not. But I don't think my industry is too represenative on what affects HN hirings. I've seen maybe 5-ish game positions on Ycomb job boards ever (they don't even have it as a category to filter into).

Layoffs across the industry were already well underway when Musk completed his first round. I think some tech company execs were happy to use Musk as a high profile excuse, but investors everywhere were demanding layoffs/corrections as early as Summer 2022.

FYI, you could've just used the hackernews API, and get all posts by the user `whoishiring`, which submits all these who is hiring posts. And then filter out only the posts where the title starts with "Ask HN: Who is hiring?", as this bot also submits the 'Who wants to be hired?' and 'Freelancer? Seeking Freelancer?' posts.

https://hacker-news.firebaseio.com/v0/user/whoishiring.json?...

I seem to have noticed that Rust has gotten more common than golang in the job postings, but it's hard to verify without code because there are so many false positives for "go" on any post.

Hardly scientific, but shows the risk of using Hacker News to draw overly strong conclusions of language popularity.

> it's hard to verify without code because there are so many false positives for "go" on any post.

Hence grep being insufficient in this case.

I should write a bot to analyze users and hide those who are actively and regularly negative online or on a particular forum and hide their comments.

Yes, please! I'd sponsor this work if you want donations.

A browser plugin would be even better.

I want to filter out (or at least highlight): ads, sponsored content, clickbait, rage, negativity, trolling, crypto (much of it anyway), assholes, rudeness, politically partisan, spam emails, etc.

If this could work on every website, and eventually every pane of glass (phone, etc.), it would be truly magical and absolutely game changing for the trajectory of the internet.

I want more signal and less negativity. And fewer people trying to sell me on things (unless those things are interesting ideas).

This is the correct way to fight the algorithm and the walled off platforms.

The article says it cost ~$50. That's astonishing. That's 25 minutes of my billable time.

On topic: things people can run on their computers or hold in their hands. For hardware, you can post a video or detailed article. For books, a sample chapter is ok.

Off topic: blog posts, sign-up pages, newsletters, lists, and other reading material. Those can't be tried out, so can't be Show HNs. Make a regular submission instead.

Things that can't be Show HN can be perfectly fine regular HN posts, just like this thing already is.

Edit: Maybe I'm figuring it out, is it that you thought that self-posting is strongly discouraged and that 'Show HN' is a label you're supposed to use to identify self-posting? They're not directly related, Show HN is just a special HN category with its own subsection and its own extra rules. Nobody has to use it but if you do use it, the post has to be within those rules. The other, somewhat orthogonal thing is that you can post your own stuff (in or outside Show HN), just don't overdo it.