ftkftk

7 ms·

31.
by ftkftk 2y ago
When I go fossiling in Florida almost every large find is from the Pleistocene or Holocene, when Florida was not covered by the ocean. Identifying the bone fragments is always a big puzzle that involves a good bit of study and reference boo
32.
by ftkftk 2y ago
It was the best of times. However not at a FAANG but a smaller shop with roughly 100 engineers.
33.
by ftkftk 2y ago
My balmer peak was at 1.75 pints and lasted thru the end of 3. Then followed by a steep decline. I agree that this would need to be modeled in the robot.
34.
by ftkftk 2y ago
Great job. Both on your build as well as your entertaining video. Makes me miss the days of in-person working where we spent a LOT of time on the foosball table. To the point where we built an automatic score tracking system complete with d
35.
by ftkftk 3y ago
One strategy is to store the model version used for each vector, and then using the appropriate model to embed your queries. Of course that means that for a certain time frame you may need to blend two result sets, so that may not be ideal
36.
by ftkftk 3y ago
Regarding embeddings - I am assuming you are using ada-002 or have you moved on to 3-small already? Do you have a particular strategy for migrating embedding models other than re-embed the whole dataset? And lastly, what is/are your ve
37.
by ftkftk 3y ago
If you are using different versions of GPT, in particular GPT-4, how are you controlling costs once users' projects/backlogs grow in complexity?
38.
by ftkftk 3y ago
That is correct. The process is manual and definitely has humans in the loop. It is also required for every individual subscription in your account.
39.
by ftkftk 3y ago
Somebody is having a bad Tuesday.
40.
by ftkftk 3y ago
Sounds like a clickbait headline. Reading the article confirms the suspicion. A lot of narrative being pushed in this.
41.
by ftkftk 3y ago
The title is misleading. > He said the Linux Foundation’s adoption of Open Tofu raised serious questions. “What does it say for the future of open source, if foundations will just take it and give it a home. That is tragic for open sourc
42.
by ftkftk 3y ago
Thanks for the chuckle.
43.
by ftkftk 3y ago
Relevant NYT article (paywall) https://www.nytimes.com/2023/08/16/technology/ai-defcon-hack...
44.
Employee Search with OpenAI, Flowise, and LangChain(betterprogramming.pub)
2 points by ftkftk 3y ago|1 comments
45.
by ftkftk 3y ago
I think this depends entirely on scale and performance metrics. For many smaller use cases using Lucene (or postgres, or elasticsearch or whatever else you already have running in your stack) is perfectly adequate as this paper shows. But a
46.
by ftkftk 3y ago
It's expensive to run.
47.
by ftkftk 3y ago
If you are new to vector databases Weaviate is nice to get started with. With docker you are up and running in a few minutes. It is very easy to integrate with embedding and generative models from openai (direct and Azure), cohere and googl
48.
by ftkftk 3y ago
My thoughts exactly half way through the article.
49.
by ftkftk 4y ago
One strategy that may help is to have two pages in your diagram, where one includes more detail. I don't do it often, but when needed I will create the more detailed version first, then duplicate it and reduce the detail by just deleti
50.
by ftkftk 4y ago
In my experience a few boxes and sticks usually communicate more effectively to more constituents (both technical and not) than the most well-written document. The trick is to know how much detail is necessary to include at the time.
51.
by ftkftk 4y ago
I hadn't tried Gaia before today. Gave it a whirl and looks like a great tool for my backpacking trips. I am a paper map and compass kinda guy but this seems like a convenient addition, especially for trip planning. Mapped out the next
52.
by ftkftk 4y ago
Agreed.
53.
by ftkftk 4y ago
It is not.
54.
by ftkftk 4y ago
I agree - I don't believe that there is a grand master plan of a conspiratorial or other nature. I think it is simply, as you stated, a co-evolution of independent actors.
55.
by ftkftk 4y ago
The joke all were thinking, but were afraid to type out. Kudos!
56.
by ftkftk 4y ago
User-agent strings finally make sense now. I thoroughly enjoyed this article.
57.
by ftkftk 5y ago
I predict that predictions will get easier and automated over the next 10 year time period.
58.
by ftkftk 5y ago
Excellent book - came here to recommend it as well.
59.
by ftkftk 5y ago
What was your second best culture?
60.
A Look at Workplace Culture: Why I Love Working at CLA(ryanboddy.net)
2 points by ftkftk 5y ago|3 comments

More ›