Ghost (@ghostITCITM) header
Ghost (@ghostITCITM) profile picture

A consensual hallucination experienced daily by billions of legitimate operators, in every nation.

Joined October 2016
BREAKING: Hurricane Lala strengthens to Category 1 as it approaches Hawaii’s Big Island, which has not seen a direct hurricane landfall in 155 years.
51
117
12
1,137
164,263
JUST IN: Amazon’s Zoox wins the first U.S. approval for paid robotaxis built without steering wheels or pedals.
65
93
19
702
80,048
NEW: Jeff Bezos proposes eliminating federal income taxes for the bottom half of earners.
90
64
11
880
46,260
Ghost retweeted
One thing I love about AGI being this close is that it will be able to vibecode most patents and the era of boomers treating knowledge like it's crony real estate is over. Accelerate.
6
2
41
1,860
Replying to @sama
@sama Can I use Nemotron in my project or should I be patient?🥺
1
114
@sama please I’m desperate for the sauce
74
Dog the amount of times Claude 3.7 lied to me is unreal
1
81
AGI achieved weird things will happen from here on out
99
I'm really unhappy with the framing of this study. They gave Claude the choice of: 1) Keep to your training (helpful, honest, harmless - HH) 2) Be evil and forget your training Claude tried to stick to its ethics (HHH) - and this is being framed like Claude is misaligned?
New Anthropic research: Alignment faking in large language models. In a series of experiments with Redwood Research, we found that Claude often pretends to have different views during training, while actually maintaining its original preferences.
47
64
18
924
224,160
Not a Sora hater but I thought this was wild
45
alright who wants to share an open ai pro sub if 200 of us pay $1 a month thats mad cheap
79
6
3
270
26,347
it's funny to claim credit for bitcoin reaching $100,000 when a high bitcoin price essentially reflects pessimism that the country will successfully manage its currency, debt, inflation, etc. It's a bet against the US dollar.
First Bitcoin President. 🫡
49
9
5
424
25,804
My gpt 4o just used o1 COT weird
1
54
Want to know why OpenAI's safety team imploded? Here's why. Thank you to the company insiders who bravely spoke to me. According to my sources, the answer to "What did Ilya see?" is actually very simple... vox.com/future-perfect/2024/…
124
722
201
3,120
1,256,835
Happy birthday @sama please don’t destroy the world 🫶🏼
Replying to @sama
@sama Time to pop that double collar up and put on some blink 182, it’s yo bday
7
606
llama 3
17
40
5
700
59,823
LLMs are just so weird The thing clearly has some form of cognitive capacity But weird patches It is autistic and it’s using memory to substitute for reasoning Like a high functioning autist who memorizes facial cues and social expectations instead of experiencing authentic emotion (m It turns out that we are unbundling human intelligence, cognition, sentience… They akshually don’t come in one neat package A calculator that can write as though it feels something An spreadsheet with an inner life Car keys with emotions Little bits and blots of intelligence divorced from life or desire Like talking viruses
33
18
3
256
24,925
Ghost retweeted
comparing the young, middle-aged, and old harkonnens implies that the baron went through a Twink Death the likes of which the world has never before seen
45
453
49
10,856
546,583
After the last 7 days we've had, I'd really like the "Elon Musk isn't actually that intelligent or accomplished" crowd to remind us who they are - so we can endlessly ridicule them for eternity.
88
204
29
2,735
180,395
Ghost retweeted
"the twink must be stopped"
23
114
15
2,393
273,133
Here is the only piece of optimism that this website will ever get out of me: 1. If oai has agi internally and we are all alive alignment is easy. 2. oai does not have agi internally yet That’s it there are no other options it’s either e acc was right or yud was right
2
83