Sabitlenmiş Tweet
MedKits101
7K posts
MedKits101
@MedKits101
Medic. Lefty. Shitposter. Playlist to get you started on basic first aid skills: https://t.co/ytmhAKXOVr He/Him
Katılım Eylül 2019
1.1K Takip Edilen1.2K Takipçiler
Ok, cool. What's the battery life and how long does it take to fully charge on a standard outlet? Cause until battery tech improves you're gonna have a 200lb paperweight taking up space in your house to do an hour of work every eight hours
Brett Adcock@adcock_brett
Oh really? This entire task was done with a neural net as well
English
I called out of work today cause I'm sick, and I notice that 1) I have a bunch of notifications and 2) people are talking about something with AWS?
I have no intention of either checking Teams nor googling what's happening so if any of you want to lmk that'd be cool
English
It's so simple. You just need to solve the two biggest problems in power infrastructure that experts have been battling since the field was first created. You idiots. You absolute morons. GOD!
(This is genuinely how the people in charge of everything think, and we are screwed)
English
This prompt in any AI chatbot can really, really help the output:
"Amazing. Please go back over your output and red-team it, generating a second draft that fine-tunes your claims. Act as if you're debating them, then output the modified version that you can fully defend against red-team you! Thank you!"
English
"When a man sees a xerox output what took him hours to write by hand, he joins the team"
Wes Roth@WesRoth
When a black hole expert watches GPT-5 Pro solve in 30 minutes what took him days of hand calculation, he joins the team. Alex Lupsasca just announced he’s joined OpenAI for Science to help push AI to the edge of physics and beyond. A major win for the future of automated reasoning.
English
This is easily one of the best adaptations of this play that's ever been filmed
Gabe@melchizedek420
Why did my 9th grade english teacher make us watch this shit
English
@MedKits101 Ya one of the other concerns is insecure design and implementation using it. But LLM CAN be a useful tool. I just want the bubble to burst so everyone can do legit analysis of its beneficial, harm-reductionist uses
English
People getting pissed at this need to understand that LLM's are not the future of data/coding, they are the present. That's just how it works now. They shouldn't be used for creative tasks, or art, and how they're used by most people sucks, but in industry this just where we are
Culture Crave 🍿@CultureCrave
Hideo Kojima on AI "A lot of people use AI in creative work to come up with ideas, but I think of AI as more of a friend ... I would lead the creative part and use AI to boost efficiency" "I'd like AI to handle the tedious tasks that would lower cost and cut down on time ... co-creating with AI instead of just using it" (via @WIRED)
English
@KADB627 Yeah, you really have to stay on top of it for sure. If you're not reviewing every pr it puts out then you're asking for trouble. But that said, it's still saved me dozens of hours or work, so worth it over all. Should probably still burn OpenAI to the ground on principle though
English
@MedKits101 Right right. Ya the only part I took umbrage with is LLM for coding.
The other big downside for LLM-Coding is I hear it makes debugging a nightmare. Not a dev tho, and I avoid all AI shit where I can
English
@KADB627 Depends on what you're trying to do and how robust public repos are for that thing. I find it incredibly useful for doing low skill, time consuming, easy to predefine tasks. Write up a deployment plan document and have it execute it step by step. Wouldn't trust it to plan though
English
@MedKits101 Ehhh, I’d be very hesitant to say that. As a tool in general they’re not going anywhere but LLM coding is a nightmare
English
MedKits101 retweetledi
Any path forward requires the democrats to pack the court. This is the moderate position, we have justices taking bribes, we have white nationalists, we have cranks, they have to be marginalized.
USA TODAY@USATODAY
Supreme Court asks if Black voter protections discriminate against White people usatoday.com/story/news/pol…
English
Most people use chatgpt to write shitty papers & stroke their ego. The free models do that just fine. The paying user base was always gonna be small and OpenAI is discovering in real time they're not going to be profitable if they expect normies to pay for it (thus, adding porn)
Max Spero@max_spero_
Is it just me or is 40 million paying ChatGPT users kind of low? Spotify has 276 million paid subscribers. Just kind of surprised that there are 760 million people who don't feel like they'd derive an additional $20 of value from a more capable model.
English
@deredleritt3r How was it confirmed that it's not present in the existing literature? Exact methods please
English
Just to recap:
We found out today that an LLM that fits on a high-end consumer GPU, when trained on specific biological data, can discover a novel method to make cancer tumors more responsive to immunotherapy.
Confirmed novel discovery (not present in existing literature). Experimentally validated in living cells.
This is AI generating novel science. The moment has finally arrived.
prinz@deredleritt3r
Google and Yale scientists have trained an LLM that has generated a novel hypothesis about cancer cellular behavior. This prediction was confirmed multiple times in vitro. - "What made this prediction so exciting was that it was a novel idea. Although CK2 has been implicated in many cellular functions, including as a modulator of the immune system, inhibiting CK2 via silmitasertib has not been reported in the literature to explicitly enhance MHC-I expression or antigen presentation. This highlights that the model was generating a new, testable hypothesis, and not just repeating known facts." The model that generated this prediction is a 27B-parameter LLM based on the Google Gemma open source models, and trained on a corpus comprising >1B tokens of transcriptomic data, biological text, and metadata. Quite remarkable that a small (just 27B) LLM trained on specialized data is able to make novel scientific discoveries. "Teams at Yale are now exploring the mechanism uncovered here and testing additional AI-generated predictions in other immune contexts. With further preclinical and clinical validation, such hypotheses may be able to ultimately accelerate the path to new therapies."
English
But yeah, there's no way this doesn't IMMEDIATELY lead to the largest pile of revenge porn and csam that as ever existed
English
Hearing multiple independent whispers that OpenAI just had a breakthrough too big to announce casually. Something emergent. People inside calling it “Phase 2.” Others say “Level 4.”
English
@MustacheBob2 Yeah, fair. That's probably more a statement of defeat than anything else
English
@MedKits101 It's understandable coming from the artist driven 70s to seeing and largely becoming part of a system even larger and more absurd than the parodies of corpo culture he once railed against
English
Is it just me, or is the a weirdly cynical take coming from a guy who made a fortune making art about small groups of people actively attempting to overthrow the evil in their world, which he based in part on real world revolutionary groups?
Culture Crave 🍿@CultureCrave
Hayden Christensen recalls going on a walk with George Lucas one evening while filming 'Revenge of the Sith' "George ended the conversation by saying to me, ‘Hayden, you can't change the world. It's not possible. All you can do is try to make your own world and then invite other people to be a part of it' ... It was one of the most profound things that anyone's ever said to me, and it has really informed the way I live my life"
English
Making them impotent/uninterested in sex, and combining sex drive, the desire to procreate, hunger, thirst, and a deep addiction into a single primal urge, which requires that a human die to sate, makes them more interesting, monstrous, and terrifying, regardless of who's the POV
Vampire Posting 🏠🦇🍉@Vampire_Posting
What's a vampire opinion that'll get you like this?
English