πΌπ“‰π“ˆ π’Ÿπ’Ύπ‘œπ“‡ (@ImPurchasable) header
πΌπ“‰π“ˆ π’Ÿπ’Ύπ‘œπ“‡ (@ImPurchasable) profile picture

World of pure imagination

Joined January 2016
5/ Its forecast: within a few years, AI could match top experts across most disciplines and work autonomously for hours, days or weeks. That’s management’s expectationβ€”not a demonstrated result.
1
2
6/ Public listing. Unusual control. A founders’ entity gets 50.1% of general voting power. Separately, the Long-Term Benefit Trust generally elects and removes a majority of directors.
3
Does this mean they’ve done it?…
Time to scale that SSI:
1
134
πΌπ“‰π“ˆ π’Ÿπ’Ύπ‘œπ“‡ retweeted
Time to scale that SSI:
We are announcing a long-term strategic partnership with NVIDIA. NVIDIA is making a substantial investment in SSI that will let us 10x our compute in the next 12 months. We reached the point where our research is worth scaling and with this partnership we will be able to. We are honored by NVIDIA’s conviction.
650
1,145
380
15,504
2,548,965
First signs of AGI in the streets of Brooklyn
First signs of AGI in Amsterdam
1
2
351
πΌπ“‰π“ˆ π’Ÿπ’Ύπ‘œπ“‡ retweeted
My five year plan is to marry Mira Murati
122
70
15
4,099
356,822
πΌπ“‰π“ˆ π’Ÿπ’Ύπ‘œπ“‡ retweeted
Replying to @OfficialLoganK
You know how to make a man hard Logan. @sama you use to be able to do that, where’s the passion
26
8
1
474
22,597
πΌπ“‰π“ˆ π’Ÿπ’Ύπ‘œπ“‡ retweeted
RT to help Simon raise awareness of prompt injection attacks in LLMs. Feels a bit like the wild west of early computing, with computer viruses (now = malicious prompts hiding in web data/tools), and not well developed defenses (antivirus, or a lot more developed kernel/user space security paradigm where e.g. an agent is given very specific action types instead of the ability to run arbitrary bash scripts). Conflicted because I want to be an early adopter of LLM agents in my personal computing but the wild west of possibility is holding me back.
If you use "AI agents" (LLMs that call tools) you need to be aware of the Lethal Trifecta Any time you combine access to private data with exposure to untrusted content and the ability to externally communicate an attacker can trick the system into stealing your data!
102
504
46
2,966
436,476
πΌπ“‰π“ˆ π’Ÿπ’Ύπ‘œπ“‡ retweeted
the job related meaning crisis has already started and will soon go full swing. this may sound insane but my only hope is that it happens quickly and at large enough scale such that everyone is forced to rebuild rather than painfully hanging onto the old structures
160
226
131
3,604
1,517,288
This was the best car I ever owned hands down
a few years ago, I painstaikingly scanned my old 4Runner right before selling it. Today I reprocessed the data as a gaussian splat It was a great little offroader. I'll visit it in VR some day and feel like I'm sitting back in the driver seat.
3
3
165
Why…
JUST IN - Pentagon wants AI to fabricate deep fake online personas for use on social media platforms β€” The Intercept
1
103
"The future is going to be so bright that no one can do it justice by trying to write about it now"
50
πΌπ“‰π“ˆ π’Ÿπ’Ύπ‘œπ“‡ retweeted
Possible secret recipe for next-gen agents > Finetuned ADAS + AutoToS/GoT/AoT/CoV/MCTS + GraphRAG w/LGGMs for causal reasoning + turing complete planning > Search trees + agent tools continuously evolving from accumulated knowledge > Optimizable agent graphs w/ DSPy LinksπŸ‘‡
3
4
23
1,458
what is the purpose of mosquitoes in this particular simulation…
2
1
3
147
πΌπ“‰π“ˆ π’Ÿπ’Ύπ‘œπ“‡ retweeted
You can learn anything if you sit down and focus for long enough
As a tech professional, what has been your greatest learning?
2
2
9
489