great_psy

6 ms·

31.
by great_psy 8mo ago
Yes and no. Regular folks will not pay Anthropic, but NSA, NASA or research labs might. I’m not implying this will be a good time for AI companies. I am saying AI as a technology can provide value without it being controlled by only 3 compa
32.
by great_psy 8mo ago
Can this method be extended to go down to the sentence level ? In the example it shows how much of the reason for an answer is due to data from Wikipedia. Can it drill down to show paragraph or sentence level that influences the answer ?
33.
by great_psy 8mo ago
Actually I think there’s another option. There’s the scenario where LLMs get more efficient in size, and to get 2026 SOTA performance you will be able to get it from consumer grade laptop. Sure with a 1000B parameter you will get better per
34.
by great_psy 8mo ago
Maybe I’m not creative enough to see the potential, but what value does this bring ? Given the example I saw about CRISPR, what does this model give over a different, non explaining model in the output ? Does it really make me more confiden
35.
by great_psy 8mo ago
Any place where we can test the llm output without loading it on n64? Curious of what we can get out of those constraints.
36.
by great_psy 1y ago
The introduction to the article is not paywalled. But the actual 2027 ai story is paywalled
37.
by great_psy 1y ago
Can you provide some names of AI apps who’s revenue > cost ?
38.
by great_psy 1y ago
But leaving a light on 2x the time will equal very close to 2x the price. Asking “what day is today” vs “create this api endpoint to adjust the inventory” will cost vastly different. And honestly I have no clue where to start to even estima
39.
by great_psy 1y ago
A likely catalyst would be a DeepSeek like model where it has the capability of top model but in a much smaller footprint. NVDA and cloud providers that are highly staked in AI would likely take a hit if a 70B model could do what Sonnet4 do
40.
by great_psy 1y ago
I have not delved to deep in the code, but is there any functional differences it has over Java other than the size ? Presumably Java would also be pretty tiny if we wrote it in bytecode instead of higher lever Java.
41.
by great_psy 1y ago
Is this an indication of the peak of the AI bubble ? In a way this is saying that there are some GPUs just sitting around so they would rather get 50% than nothing for their use.
42.
by great_psy 2y ago
I wouldn’t be so dismissive. Research is just a loop of hypothesis, experiments, collect data, make new hypothesis. There’s so creativity required for scientific breakthroughs, but 99.9% percent of scientists don’t need this creativity. Ju
43.
by great_psy 2y ago
Why would you want to have an ever growing memory usage for your Python environment? Since LLM context is limited, at some point the LLM will forget what was defined at the beginning so you will need to reset/ remind the LLM whats in m
44.
by great_psy 2y ago
Before or after the bubble popped ?
45.
by great_psy 2y ago
With how healthcare is handled in America … good luck to poor people getting access to anything like that.
46.
by great_psy 2y ago
Yes it’s possible, but it’s not something you can easily scale. I had a similar project a few years back that used OSX automations and Shortcuts and Python to send a message everyday to a friend. It required you to be signed in to iMessage
47.
by great_psy 2y ago
When’s the last time you personally had a bit flip on you?
48.
by great_psy 2y ago
We focused so much on how that we didn’t stop to ask why. We already have an interface for agents, we call that an API. Why do we need to have AI click buttons, that eventually call an API. Just skip the middle layer and go straight to an A
49.
by great_psy 2y ago
I think a lot of conspiracies people have are out in the open, it’s just that they are hidden behind articles that take 30 minutes to read and understand.
50.
by great_psy 2y ago
When you say, “monitor red flags” what is actually happening ? How are you determining red flags ? Will an LLM be running every x hours analyzing news, so I can give very specific things to look for ?
51.
by great_psy 2y ago
Does this lower the protection iCloud offers ? Currently you need to use biometric or password to access your keys. If you use the extension will it make a copy of the passwords locally, or does it still require biometric on every password
52.
by great_psy 2y ago
I found LLMs to be quite useful, so the time saved is worth the effort to double check answers. Granted it might have to do with how I use LLMs. If you just blindly ask a question you increase the chance of hallucinations. If you give a le
53.
by great_psy 2y ago
This was for some internal tooling. Not sure about router deployments.
54.
by great_psy 2y ago
I have heard the deployment process at Rogers is something along the lines of: 1) get approval for your changes 2) ssh into a server 3) copy a backup of current jar file 4) copy the new instance 5) run some command to refresh
55.
by great_psy 2y ago
Is there more than the cover to this ? On mobile I only see one page of the PDF.
56.
by great_psy 2y ago
I can’t speak for what OpenAI is doing, but I’ve noticed those types of hallucinations occurring when I quantize a model beyond a certain point. Maybe they are trying to cut down on memory usage ?
57.
by great_psy 2y ago
I’m excited to hear work is being done on models that support function calling natively. Does anybody know if performance could be greatly increased if only a single language was supported ? I suspect there’s a high demand for models that a
58.
by great_psy 2y ago
Does not ‘seem’ like the right shape for it, but could it use ocean currents ? Similar idea to those weather balloons we were talking about last year ? Is that even possible in the ocean ?
59.
by great_psy 2y ago
Seems like a waste of money to me. Training some NN is just matrix multiplications. Why would you need a custom silicon to take advantage of the software ? I suspect a GPU is optimized for this operation all the way to software level. I c
60.
by great_psy 2y ago
The issue I see with the screenshots from the article is that it changed the contrast and overall feel of the scene. I think there’s plenty of opportunity to exchange old videos, but I think it requires some human touch or deeper understand

More ›