lysecret

22 ms·

301.
LangChain Announces 10M Seed Round(blog.langchain.dev)
235 points by lysecret 4y ago|167 comments
302.
by lysecret 4y ago
Oh very interesting didn’t expect so much scepticism on parquet. We use it to store around 100 tb of data in a partitioned way on GCS and query it using BigQuery. Absolutely fantastic solution and incredibly cheap too.
303.
by lysecret 4y ago
Hmm I am sure this is well intentioned, but I am a bit scared this will just further chip away on FireFoxes market share which doesn't look good to begin with.
304.
by lysecret 4y ago
Id say it has approximate as high a chance to work as "random code from the internet". Also, it is often just relying on an old version of an API or a Library.
305.
by lysecret 4y ago
Thank you so much. I am so excited about Ai and there are much negative sentiment around (which I also understand). But it is refreshing to read some optimism around the topic.
306.
by lysecret 4y ago
Hey just wanted to take this chance to thank the Retool people. Retool was absolutely essential for our last startup. And they have been super responsive with bugs. A great product!
307.
by lysecret 4y ago
Yes, I am so curious how they did it, I know about flash attention but there is no way this gets us all the way there.
308.
by lysecret 4y ago
Take meteroits then. Ah do me a favour and google meteroits.
309.
by lysecret 4y ago
I mean this is the beauty of our market economy, in 9/10 cases when someone asks "why is this the price?" The answer is supply and demand. The other answer is incentives. Here you got a free econ degree.
310.
by lysecret 4y ago
I have used them extensively and am a big fan, but none of their models are even close to the type of thing ChatGPT can do (for now). They are built for a world where you take a pre-trained model (say Bert) and fine tune it to your specific
311.
by lysecret 4y ago
I’m highly curious of hugging faces role in the new world of huuuge models and few shot learning. Seems to me to be quite an existential thread to them.
312.
by lysecret 4y ago
As a general feedback I have written it down here before: https://news.ycombinator.com/item?id=34304376 Sorry for being a bit negative but I was very excited for AlloyDB as a potential serverless offering for Postgres. Espe
313.
by lysecret 4y ago
Hugely disappointed by Alloydb. Price and performance for a write heavy workload where off the charts horrible. It didn’t even support disabling an instance and it did only storage autoscaling much worse overall experience than using Aurora
314.
by lysecret 4y ago
This is what LangChain is about right? I think these kind of things have to be open source and self hostable, maybe with a paid and hosted offering as a side. Of course until Apple will eventually integrate this.
315.
by lysecret 4y ago
I mean, all the China scariness aside probably a good move, western regulators could take a look at.
316.
by lysecret 4y ago
Yes. Also storing each individual state can be seen as caching there is a place for both of course. Each derived state is a pain and has to be maintained but might be essential.
317.
by lysecret 4y ago
Fascinating, this reminds me a lot of a conversation I had about 4 years ago with a translator, they were working at a translation company which just introduced AI. And their initial implementation of AI looked like this: We give you an ini
318.
by lysecret 4y ago
Ah yes, Monads it couldn't teach me well but i got it good enough through this old school post https://fsharpforfunandprofit.com/series/computation-express...
319.
by lysecret 4y ago
Absolutely nothing besides teaching me how awesome tailwind typography is.
320.
by lysecret 4y ago
I am extremely bullish on this. I use ChatGPT every day. I am learning at a rate that I haven‘t in years and I love it, the pure fact I can ask follow up question or ask it to shorten things. This is the core problem of learning: I already
321.
by lysecret 4y ago
It might be slightly too early but after using ChatGPT to learn functional programming I am a huge believer. Just the fact you can ask follow up questions. Of course it’s not always correct but neither are teachers.
322.
by lysecret 4y ago
Congrats! I had something quite similar in mind (also working a lot in python based streaming ETL). I am unfamiliar with Ray. A few questions: 1. Ray seems to be focussed on ML usecases, are you as well or are you a more generic streaming
323.
by lysecret 4y ago
Probably mostly targeting financial services companies with those prices.
324.
by lysecret 4y ago
All decent replies but, the truth is that it is mostly a quirk of transformer architecture which scales quadratically in the length of the sequence because transformers look at all pairwise combinations of the input tokens. So you can get t
325.
by lysecret 4y ago
I think it’s crazy to think about that we scraped all this data from the web to train this crazy LLM to be able to interpret the web and scrape more data from it.
326.
by lysecret 4y ago
Yea LLM for data Extraction will be huuuuuge. I had so many Data Projects in the past which boiled down to: Damn the data is too unstructured.
327.
by lysecret 4y ago
Well in my experience it is generally true that when you start optimising things get less clean. Let me explain: most of the optimisation situations I had looked like this. Hey this query is pretty slow and costs us quite a bit. Oh look for
328.
by lysecret 4y ago
I think a lot of people are interpreting this (implicitly) with a kind of Generative AI lense. However, this article was much more about "classical" ML type work, e.g. training your own Neural Net while I think the paradigm seems
329.
by lysecret 4y ago
I am super excited about ChatGPT even just yesterday it helped me fix a really difficult bug. (I was so excited that I even wanted to pay for it but their payment was broken haha) However, I am starting to think that moving it to bing alrea
330.
by lysecret 4y ago
Please, I am very curious about the kind of "non boring" classes you are talking about. If you could show an example. I couldn't come up with anything that I would define as reasonable where we couldn't define the types

More ›