HN user
HackerSmacker profile: https://www.hackersmacker.org/user/Kydlaw?hs=hpSX58FooTyWlMBVt6
I have a love-hate relationship with GitHub Actions. Love because they are right there in my GitHub repository. Hate because they are very brittle once you move out of the happy path.
It seems GitLab has a much better experience in this department, but their pricing is hard to justify for us...
Genuinely curious if folks here had better experiences or recommendations for a smooth CI/CD experience.
You might have accidentally described what accounting is.
You should give a try to https://pixi.sh/latest/ (I am not involve in the project).
They are a little more focus on scientific computing than uv, which is more general. They might be a better option in your case.
We explored LangGraph last November and were pleasantly surprised by the difference with LangChain. The framework had much more care put in it. It was much easier to iterate and the final solutions felt less brittle.
BUt the pricing model and deployment story felt odd. The business model around LangGraph reminded us of Next.js/Vercel, with a solid vendor lock-in and every cent squeezed out of the solution. The lack of clarity on that front made us go with Pydantic AI.
I discovered Marimo a couple weeks/months ago here iirc. This really lands on a sweet spot for me for data exploration. For me the features that really nails it are the easy imports from other modules, the integrated UI components, and the app mode.
Being able to build model/simulations easily and being able to share them with others, who can then even interact with the results, as truly motivated me to try more stuff and build more. I've been deploying more and more of these apps as PoCs to prospects and people really like them as well.
Big thanks to the team!
In the list that you filtered and hand picked yourself there is literally "War crimes committed in Serbia". The article linked is mentioning a researcher that spoke badly about the US president.
Help me understand how these two things are on same level, please?
In fact they are mentioning LangGraph (the agent framework from the LangChain company). Imo LangGraph is a much more thoughtful and better built piece of software than the LangChain framework.
As I said, they already mention LangGraph in the article, so the Anthropic's conclusions still hold (i.e. KISS).
But this thread is going in the wrong direction when talking about LangChain
I am adding a blog post, shared on HN a couple months ago, that show an architectural cross section of the city.
https://cohost.org/belarius/post/6677850-architectural-cross
(I am not the author of the blog, nor the original poster, but I just want to share the link because I found this incredibly cool)
I am so puzzled by this attitude. People are free to make any distinction they want, the way they want, but the representation you have of a scientific literature is erroneous.
Scientific articles are informational tools that report results of experiments and nothing more. If the results are interesting to the peers, they are published. By they are not world's laws made paper unless sufficient replications are made. This means that each article need to be read with the context of the literature in mind and with a critical eye. Each are a single point of evidence to a phenomena.
Hence, there are subjective informational tools, written toward a specific audience (the experts of the domains) to inform of a specific result in a specific case.
On top of that, their are specific journal/issues where these types submissions are allowed. Don't read these submissions if you are looking for serious "information tools"
Scientific literature must be handled the same way as legal literature. If you are not a law expert, you ask a lawyer. If you think you are a legal expert when you are not, surprising consequences may arise.
In universities, they are classes dedicated to handling the scientific literature. They are provided for a reason.
So please, don't use cat's physic for liquid simulation in game engine... or please do?
Concerning the currency symbol positioning, yes, it is indeed country-dependent. Countries of the Commonwealth usually place the currency sign in front of the figure, but most european do not, and instead place the symbol after the figure.
See the Use section in https://en.wikipedia.org/wiki/Euro_sign
Everybody is talking about the existential risk posed by AI but this guys release this tool in the wild without any rail guards... concerning, really.
A bit more seriously, it can be really useful to have a graph of ArXiv tabs instead of a linear range of tabs, this can be very handy when doing a dive in scientific literature.
Hi! Informative piece to the ones that need it. Also, your last paragraphs 4 are duplicated, I guess?
> often to the point where it's one of the strings attached to the money, that you can't publicly disclose who made the donation.
TIL. Philanthropy is not big where I live so I don't know the ins and outs of it.
Same here... except maybe in the opposite direction? Does it implies that their other philanthropy would not benefit from more public support? So what are they in this case?
Anyway, they do whatever they want with their philanthropy in the end, but I found that was an odd phrasing.
I recently moved to data engineering role where everything uses GCP services (think BigQuery, DataProc, Cloud Storage, ...) and wondered is all that was really necessary?
What would be the simple yet robust infra for data eng? Not thought a lot about it for now, so I am curious if some of you have would have any insights.
The next paragraph explains that: "When looking at the test coverage numbers alone, Sail’s capability may seem limited. But we have found that there is a long tail of failed tests due to formatting discrepancies, edge cases, and less-used SQL functions, which we will continue tackling in future releases."
I am with you that it is still very very early. I'll personally keep an eye on the project.
I wasn't expecting Elon Musk as a proofreader for PG's essays. To me that's interesting just by itself.
That might be personal bias, but I think the idea that PG is developing here already stem a lot in the Agile movement (not the software one, the general one). The problem is that, as PG mention, there is probably not one founder mode, but more like a set of them. This makes their academic description harder to conceive.
I personally find this information frightening... we are now in a world where management no longer even listens to its engineers or technical experts when they warn of serious dangers. Obviously everyone thinks of Boeing, and today it's “only” a train station, but imagine if the same culture developed in an industrial chemical plant or a nuclear power station?
Where do you go? What do you do? Unless you have an Engineer Association or an Union to back you up, you are doomed to be crushed by your upper management and beyond...
If I understand correctly what you are looking for, Ollama might be a solution (https://ollama.com/)?. I have no affiliation, but I lazily use this solution when I want to run a quick model locally.
This type of comment is one of the two reasons I'm still opening comments on HN. Thank you for the laugh.
IMO LangChain provides very high level abstractions that are very useful for prototyping. It allows you to abstract away components while you dig deeper on some parts that will deliver actual value.
But aside from that, I don't think I would run it in production. If something breaks, I feel like we would be in a world of pain to get things back up and running. I am glad they shared their experience on that, this is an interesting data point.
I am genuinely curious of how they are going to design the matching system. Which features are going to be used to describe individuals? I feel like the government has access to many more data than a private company, and the temptation to tap into PII for that might be high.
Thank you for sharing, I love the dictionary format, only mentioning the main concepts and links to learn more about them. Sources all look great btw also! But as always, Umar Jamil[1] is missing...
If you think Andrej's videos are useful, people should definitely go learn with this guy took. Great in-depth explanation of concepts where he doesn't shy away on theory while still providing hands-on code exploration. Not affiliated, just wanted to give him and his work a shoutout.
They do have an extension for vector similarity search (https://duckdb.org/docs/extensions/vss).
But Milvius might propose more features, as they have been in this specific space for longer.
Interesting line of work, but as (too?) often on LessWrong, there is no literature review or even a related work section to go back to seminal research or dig deeper into the topic. It might just be me, but I find this really frustrating when I'm unable to hop references to skim a field.
Does anyone have cool links or references to share about related work?
Really cool and impressive work! I'm more on the software (high level) side than on the hardware, but I always wondered how USBs and other ports were truly working. It is an area of knowledge that I'm shamefully lacking...
Does anyone have nice resources to share like this one that focuses on a specific port/connection and implement a driver/reader/parser? I'd very like to learn more of this.
They state in their report that they filter evaluation data off their training data, see p.3 - Filtering:
"Further, we filter all evaluation sets from our pre-training data mixture, run targeted contamination analyses to check against evaluation set leakage, and reduce the risk of recitation by minimizing proliferation of sensitive outputs."
Yep, I noticed it immediately too. Yet it is subtle in reality. I'm not that good to spot imperfections on picture but on the video I immediately felt something was not quite right.
It takes into account the recent changes. At the top of the window it indicates the Python's version used.
Lots of great resources listed here. But I think there is "Understanding Deep Learning" missing from the list [0]. In my opinion, Simon J.D. Prince accomplished a true feat with his book, not only through the material itself but also with the notes attached to each chapter, linking directly to advanced references (free literature review), exercises that really challenge your understanding of the material, and great notebooks with code that truly materialized the concepts learned (free exercises to give to students if you teach a DL class, but this community is probably not the targeted audience for that).