Elias Moosman (@Wiber) profile picture

Explored what comes after robots two decades. Did something different.

Joined August 2007
Photo by Elias Moosman (@Wiber)Photo by Elias Moosman (@Wiber)Photo by Elias Moosman (@Wiber)Photo by Elias Moosman (@Wiber)Photo by Elias Moosman (@Wiber)Photo by Elias Moosman (@Wiber)Photo by Elias Moosman (@Wiber)Photo by Elias Moosman (@Wiber)Photo by Elias Moosman (@Wiber)Photo by Elias Moosman (@Wiber)
Elias Moosman retweeted
Folks, seriously. The GraphRAG Ecosystem Tools are frickin' amazing. The KG Builder is a visual cloud service that easily (click click!) turns unstructured data into a knowledge graph! You can point it to some random PDFs, a wikipedia page or a youtube video, and a few seconds later you have the data and concepts from those sources represented as a knowledge graph in your Aura free account. 🤯 This used to take DAYS! It really makes it so easy to get started on your own knowledge graph journey. neo4j.com/developer-blog/gra…
13
87
6
485
57,386
We can't stop AI anymore. Our only path forward is to guide it.
1,243
1,599
239
8,782
3,033,717
Elias Moosman retweeted
Having spent most of the day in horror at AI and warfare (see @ScotNational tomorrow), I reserve the right to cease the day sentimentally. So…goodnight
6
26
5
251
15,642
5 powerful laws to be aware of.
11
534
11
1,521
134,961
Elias Moosman retweeted
Why is the situation with the upcoming AGI is very similar to the film "Don't Look Up"? A segment from my short-doc "Don't Look Up - The Documentary: The Case For AI As An Existential Threat."
4
6
26
1,562
Elias Moosman retweeted
I played a game of REFERENCE-CLASS TENNIS 🎾 with @eriktorenberg over the future of superintelligent AI. Erik's reference class is techno-optimism: Tech breakthroughs have consistently been extremely good for humanity. I hit back with three alternative reference classes: 1. Game-over buttons: There currently exist buttons that instantly kill 100 million humans when pressed (the triggers on thermonuclear bombs). What's the next game-over button in that sequence? 2. Extinction events: Earth has had five great extinctions, killing up to 96% of all life. When might the next extinction be, and what percent of all life will the next extinction kill? (Keep in mind that 100% was historically prohibited due to the anthropic principle, but is allowed in the future.) 3. Optimization-power jumps: The jump from passive physical processes to evolution by natural selection blanketed Earth in life. The jump from animal-level to human-level intelligence has caused similarly massive changes in Earth's ecosystem, with many species going extinct, according to our whims and side effects. What similar effects should we expect from the next optimization-jump, the jump from human-level intelligence to artificial superintelligence? Ultimately, we can't answer these questions by playing reference-class tennis, so my goal wasn't to "win". My goal was to broaden the listener's perspective about other reference classes besides techno-optimism that apply to AI x-risk. There are MANY reference classes that are equally valid, yet much more dire. 🏓
5
10
48
13,800
The artwork represents the intricate interplay of various elements within a partially mechanistically interpretable AI system as we evolve partial mechanistic interpretability, highlighting the intricate balance between the structured, proven and the inherent dynamism of AI.
1
131
🎉 We are proud to announce the winners of the 2023 Future of Life Award! This year, we celebrate the unsung achievements of five visionaries whose films helped prevent nuclear war. 🧵 Read about the winners below ⬇️ (1/5)
89
555
62
2,506
5,522,743
I call bs, if heaven on earth can be true for a moment it can be true for a lifetime, the hedonic treadmill is the absence of discipline. The right intentions matter. Heaven is not built in control, envy, power or ego. Threatening free will is the cardinal sin of bad AI. 🤖 🍟 🧠
2
1
89
Elias Moosman retweeted
Some wild decisions will come up in the future about AI programs
4,764
17,964
687
301,730
23,171,676
Replying to @liron
Steering a ship takes a map, a compass and consecutive takes on weather your on course or drifting. How much more if it’s Theseus ship and your conception of time is like a tree growing compared to the winds and currents?
1
2
1
98
"To our knowledge, this is the first demonstration of Large Language Models trained to be helpful, harmless, and honest, strategically deceiving their users in a realistic situation without direct instructions or training for deception." arxiv.org/abs/2311.07590
40
68
17
471
76,079
Midjourney, 1 year apart What if AIs improve this fast at science? What if AIs improve this fast at programming/hacking? What if AIs improve this fast at making superviruses? What if AIs improve this fast at most jobs? What if AIs improve this fast at persuasion? @sama said “i expect ai to be capable of superhuman persuasion well before it is superhuman at general intelligence, which may lead to some very strange outcomes” Did our ancestral environment prepare us for change this rapid? It’s not just art, AIs are rapidly becoming superhuman in skill after skill: “See also: the first professional-level Go AI rapidly giving way to vastly superhuman Go AI; early calculating machines vastly exceeding human performance at arithmetic; etc. When you can automate a task at all, you can often do it better, and ~always at enormous scale.” - Rob Bensinger
AI didn't spend a long time with roughly human-level ability to imitate art styles, before it became vastly superhuman at this skill. Yet for some reason, people seem happy to stake the future on the assumption that AI will spend a long time with ~par-human science ability.
29
63
19
455
809,603
Elias Moosman retweeted
Ok, I'm deeply underqualified. But maaan, I'd take that job if I could! All day figuring out how to bridge knowledge (graphs) and AI, in a non-fluff, real running code type way, working with the smartest people and the biggest companies in the world... ARE YOU KIDDING ME
1
9
1,217
Elias Moosman retweeted
Announcing formation of @xAI to understand reality
13,974
16,800
3,881
166,828
59,604,648
Elias Moosman retweeted
Replying to @ylecun
According to my biological NNs, AI existential risk to humanity is non-trivial
368
195
45
2,827
284,596
Elias Moosman retweeted
Replying to @ylecun
You really think AI is a single-edged sword?
3,108
1,679
291
26,046
3,682,908
Elias Moosman retweeted
Don’t Look Up … but AGI instead of comet
4,800
4,841
644
69,724
18,730,393
Elias Moosman retweeted
Replying to @elonmusk
In the future…
Replying to @elonmusk @DrJimFan
Have you heard of autocorrect for life?
6
21
104
115,938