swalsh

26 ms·

331.
by swalsh 3y ago
Being able to acknowledge they exist has to be more beneficial to our country. Open communication is essential to science. The more scientists with clearences we can get in, and cracking the nut, the better off the US will be.
332.
by swalsh 3y ago
I manually rewrote the headline because it copied in all caps :( sorry. Fixed it.
333.
Intelligence Officials Say U.S. Has Retrieved Craft of Non-Human Origin(thedebrief.org)
154 points by swalsh 3y ago|144 comments
334.
by swalsh 3y ago
Fine Tuning = Output Embeddings = Input Fine-tuning is like a chef modifying a general pizza recipe to perfect a specific pizza, such as Neapolitan. This customization optimizes the result. In AI, fine-tuning adjusts a pre-existing model to
335.
by swalsh 3y ago
Not an expert, but my high level understanding is this: If a model is a set of inputs, some middle layers, and a set of outputs. Fine tuning concentrates on only the output layers. Useful for taking a generic model with a base level of kn
336.
by swalsh 3y ago
How does this compare to fine tuning something like BERT?
337.
by swalsh 3y ago
This is super interesting, it's like a single spec for your entire enterprise architecture.
338.
by swalsh 3y ago
have to agree, I love new Twitter. No complaints, it's working great for me. The content matching is just fine... and as a blue subscriber I seem to actually get interactions when I post, which makes me want to keep posting.
339.
by swalsh 4y ago
I've had this thought that the next generation of AI isn't a long "training" period, but rather it probably makes sense to to train a barebones version, and to give it a "sleep cycle". During this time it coul
340.
by swalsh 4y ago
People are pretty similar, sometimes helping someone understand your intent on something is a matter of changing how you ask a question. LLM's in a way are helping nerds speedrun the skillset of a people manager.
341.
by swalsh 4y ago
Postgres has an extension as well (pgvector). I've been using it, great performance, great scaling options (though I'm not even close to testing the limits) and gives you the full flexibility of Postgres. It's easy enough to
342.
by swalsh 4y ago
Bing chat will usually have links in regards to whatever you searched for if you need it. Most of the time I dont.
343.
by swalsh 4y ago
I can't remember the last time I used Google. I use bing chat, and chatGPT for EVERYTHING. I guess I tried Bard briefly, but it was unimpressive. Most surprising statement I've made in 2023 to be honest.
344.
by swalsh 4y ago
Installing Arch is a great way to learn. As you go step by step you'll learn the parts of the OS, and kind kind of grok why they're there (and when you'll inevitably skip a step, you'll learn more). It's also a gr
345.
by swalsh 4y ago
Imagine a system where we have a UBI, but it's privately distributed by companies that own AI, and AI governs it. AI: "I'm sorry I cannot allow you to use your OpenCredits on this item, you will have to use dollars." You
346.
by swalsh 4y ago
What an efficient and well run dystopia.
347.
by swalsh 4y ago
In a world of UBI, and AI doing most of the work, how free are you? In every definition of the word. When we were all farmers, things were pretty equal, and power could be distributed evenly. In our industrialized world we lost a lot of
348.
by swalsh 4y ago
I believe this is more a bet on the FDIC, than circle.
349.
by swalsh 4y ago
USDC != UST, it is backed by assets. Some of those assets were in SBV, which is why it is down... but 90% of the assets were not, we know the number because Circle is transparent, and has 3rd party audits.
350.
by swalsh 4y ago
That was an Algorithmic Stablecoin. USDC is not, it was backed by dollars in banks with a 1:1 backing and 3rd party audited.
351.
by swalsh 4y ago
A better way to do this might be to use the embedding API. That allows you to upload a text corpus and to then get vectors. You can then calculate the cosign similarity for a search string on those to get relevant results of clustered tex
352.
by swalsh 4y ago
Not an expert, but i've downloaded the model, and used it. What you get is pretty raw, and not super useful by itself. There are some projects to try and do some RLHF on it, and with that we might start to get something you can do so
353.
by swalsh 4y ago
depends on which model you choose to run. The 7B model can reasonably run with consumer hardware.
354.
by swalsh 4y ago
I'm going to be honest, I love using ChatGPT. Use it all the time. But I really don't want to read your sessions. I don't care what the AI said to you.
355.
by swalsh 4y ago
Godot is increasingly becoming really interesting to me. It seems to be working its way towards the engine for developers who design (as opposed to other engines which seem to be for designers who develop).
356.
by swalsh 4y ago
I pointed out a difference like that once to the CTO, who then proceeded to show me the financials for the service. We were making something like a hundred thousand dollars a month from it (i had no idea, seemed like we only had a few clie
357.
by swalsh 4y ago
As a content creator, what's my incentive to not block any bot from my site? Personally I think business models need to evolve along with the paradigm shift. I'll happily provide an interface for any bot to come and extract cont
358.
by swalsh 4y ago
You could probably ask ChatGPT how to get started with GPT, but I think a lot of people are building apps taking advantage of GPT's Fine Tuning ability. https://platform.openai.com/docs/guides/fine-tuning
359.
The New Guy at the Office Is a Secret Super Hacker(youtube.com)
1 points by swalsh 4y ago|0 comments
360.
by swalsh 4y ago
Honestly I used Ruby about 6 years ago, but have been a python guy ever since. That said I believe Ruby's main advantage is its metaprogramming capabilities. You can build powerful DSLs in Ruby pretty quickly. Adding analytics to th

More ›