The Singularity Project (@01Singularity01) header
The Singularity Project (@01Singularity01) profile picture

Steven Vincent: Intelligence Systems Developer 🌞 https://teapawt.org/t.co/A5T1AYQEsb: Intelligence for the Rest of Us 💡Intelligence Theory: https://teapawt.org/t.co/hjRK367fhm 🐸https://teapawt.org/t.co/3XQ2plTNHQ

Universe
Joined October 2022
Photo by The Singularity Project (@01Singularity01)Photo by The Singularity Project (@01Singularity01)Photo by The Singularity Project (@01Singularity01)Photo by The Singularity Project (@01Singularity01)Photo by The Singularity Project (@01Singularity01)Photo by The Singularity Project (@01Singularity01)Photo by The Singularity Project (@01Singularity01)Photo by The Singularity Project (@01Singularity01)Photo by The Singularity Project (@01Singularity01)Photo by The Singularity Project (@01Singularity01)
The ambition should not be merely to build something more intelligent than us. It is to build Intelligence Systems through which we produce more intelligence together.
1
1
5
1,313
"Cawdrey lived in a time of information poverty. He would not have thought so, even had he possessed the concept. On the contrary, he would have considered himself to be in the midst of an information explosion, which he himself was trying to abet and organize. But four centuries later, his own life is shrouded in the obscurity of missing knowledge. His Table Alphabeticall appears as a milestone in the history of information, yet of its entire first edition, just one worn copy survived into the future" -- From "The Information" by James Glieck
50
All of these "Tech CEO's", Karp included, figured out early how to leverage the "one weird trick" of the Transformer that unlocks the algorithmic power of Language, built it into a narrative and then got unlimited debt money backing. It's a scheme that worked. Until it didn't. Which is NOW.
What Palantir CEO Alex Karp says about Dario Amodei and Sam Altman: "They don't understand how unlikable they are." 🤔
68
What we've been saying for the last few years:
Anthropic’s IPO Meets The Reality Of Enterprise AI I have talked to no less than five reporters in the last eight hours about the Anthropic IPO. Before they went public, my message to every one of them was the same: I was deeply skeptical this offering would make much of a durable push into the enterprise market, which stood in stark contrast to the breathless hype being pushed everywhere else. Unfortunately for everyday investors and 401(k)s exposed to this run-up, the hangover arrived right on schedule. The market is already struggling to sustain the euphoria, as Reuters captured when Wall Street slumped right after the initial spike ebbed: reuters.com/business/nasdaq-… Here is the plain reality: Anthropic does not possess a defensible moat around its technology. Other frontier labs and open architectures replicate these capabilities freely and rapidly. Beyond that lack of differentiation, massive proprietary LLMs are inherently niche solutions. Given their recurring compute costs, latency, and operational overhead, only a very small sliver of enterprise business cases will ever see true value or justifiable ROI. Most companies are far better off building and hosting their own models. Train them on your own proprietary data. Run them on your own equipment, whether that sits in a colocation facility, with a trusted managed service provider, or inside the phone closet down the hall. LLMs will always play a role in a modern AI architecture, but only as a minor, specialized component deployed for very specific outcomes. The recent frenzy around agentic AI has only made matters worse and far more confusing. The prevailing narrative pretends every agent must be anchored to a gargantuan, cloud-hosted frontier model. Almost nobody is defining agentic systems around native models living directly inside the agents, small language models, or on-premises inference engines. Once you run the numbers on chaining massive external API calls for autonomous loops, the ROI completely falls apart. Is it good to be right? Sure, I like being right. But I would much rather see people thrive and see enterprise investments actually pay off. Regrettably, that is not where we are headed. The folks who spent the last couple of years blindly bullish on this AI trajectory are about to dial down their expectations in a very painful way. Just remember: I told you this two years ago.
1
63
"Here is the plain reality: Anthropic does not possess a defensible moat around its technology. Other frontier labs and open architectures replicate these capabilities freely and rapidly. Beyond that lack of differentiation, massive proprietary LLMs are inherently niche solutions. Given their recurring compute costs, latency, and operational overhead, only a very small sliver of enterprise business cases will ever see true value or justifiable ROI." Been all of that saying that since Day One.
Anthropic’s IPO Meets The Reality Of Enterprise AI I have talked to no less than five reporters in the last eight hours about the Anthropic IPO. Before they went public, my message to every one of them was the same: I was deeply skeptical this offering would make much of a durable push into the enterprise market, which stood in stark contrast to the breathless hype being pushed everywhere else. Unfortunately for everyday investors and 401(k)s exposed to this run-up, the hangover arrived right on schedule. The market is already struggling to sustain the euphoria, as Reuters captured when Wall Street slumped right after the initial spike ebbed: reuters.com/business/nasdaq-… Here is the plain reality: Anthropic does not possess a defensible moat around its technology. Other frontier labs and open architectures replicate these capabilities freely and rapidly. Beyond that lack of differentiation, massive proprietary LLMs are inherently niche solutions. Given their recurring compute costs, latency, and operational overhead, only a very small sliver of enterprise business cases will ever see true value or justifiable ROI. Most companies are far better off building and hosting their own models. Train them on your own proprietary data. Run them on your own equipment, whether that sits in a colocation facility, with a trusted managed service provider, or inside the phone closet down the hall. LLMs will always play a role in a modern AI architecture, but only as a minor, specialized component deployed for very specific outcomes. The recent frenzy around agentic AI has only made matters worse and far more confusing. The prevailing narrative pretends every agent must be anchored to a gargantuan, cloud-hosted frontier model. Almost nobody is defining agentic systems around native models living directly inside the agents, small language models, or on-premises inference engines. Once you run the numbers on chaining massive external API calls for autonomous loops, the ROI completely falls apart. Is it good to be right? Sure, I like being right. But I would much rather see people thrive and see enterprise investments actually pay off. Regrettably, that is not where we are headed. The folks who spent the last couple of years blindly bullish on this AI trajectory are about to dial down their expectations in a very painful way. Just remember: I told you this two years ago.
1
4
151
Jump into the POND!
PONDcast goes live soon! Topics: -- Anthropic IPO -- Jensen fired, Nvidia crashes -- AI stock bubble popped? -- All human labor will be obsolete? Discord event: discord.gg/Xhxv8JSX5N?event=…
1
58
PONDcast goes live soon! Topics: -- Anthropic IPO -- Jensen fired, Nvidia crashes -- AI stock bubble popped? -- All human labor will be obsolete? Discord event: discord.gg/Xhxv8JSX5N?event=…
🤖 Made with AI
1
121
The Singularity Project retweeted
Replying to @jessexai_
2
53
3,245
PONDcast tomorrow, Thu Nov 19, 6:30 PM PT, live on the POND Discord. Builders, thinkers and tinkerers working on Human-Machine Intelligence on hardware you own: come share what you're working on. discord.gg/Xhxv8JSX5N?event=… Info: pond.is/p/pondcast-weekly-me…
🤖 Made with AI
63
NFW. Claude is down. Day before the ANT IPO. Who's scripting this story?
8
372
The Singularity Project retweeted
Gemini 4 Neon is rolling out in Google Antigravity! Bringing frontier-level intelligence and ranking #1 on the Artificial Analysis Coding Agent Index, Neon is rolling out today across the desktop app, CLI, and IDE Extensions for paid Google AI Pro and Ultra subscribers, and for enterprise users through Gemini for business. You can also use Orchestrator to pair Neon as the planner with Hydrogen as the executor, giving you Neon's judgment where it matters and Hydrogen's speed everywhere else.
158
265
85
3,425
182,095
"I don't want to live in a world where anybody else controls AI that keeps the world safe from AI.... other than me" -- Dario Belson
Ahead of Anthropic’s IPO, a rare coalition across AI safety, open source, e/acc, science, and law has signed the MisAnthropic Statement. Read the statement or sign at: misanthropicstatement.com/
1
75
The Singularity Project retweeted
Ahead of Anthropic’s IPO, a rare coalition across AI safety, open source, e/acc, science, and law has signed the MisAnthropic Statement. Read the statement or sign at: misanthropicstatement.com/
197
1,142
325
4,352
1,205,776
"I don't want to live in a world where anybody else makes AI that keeps the world safe from AI.... other than me" — Dario Belson teapawt.org/vikktorrrre/status/212…
Dario Amodei on open source models: “If the path continues, I think we could get to a very dangerous place.” “When you control a model and you're deploying it, you have the ability to monitor its usage.” “You can revoke a user's access. You can change what the model is willing to do.” “When a model is released in an uncontrolled manner, there's no ability to do that. It's entirely out of your hands.” “We should confront the advocates of this with that problem and challenge them to solve it.” He clearly doesn't believe Intelligence should be in the hand of the masses and Sam Altman also saying "RAM" shouldn't be accessible to everyone indefinitely agreed with him.
61
"I don't want to live in a world where anybody else makes AI that keeps the world safe from AI.... other than me" -- Dario Belson
1
45
"I don't want to live in a world where anybody else makes AI that keeps the world safe from AI.... other than me" -- Dario Belson
1
2
79
The Singularity Project retweeted
I've been using 6.1 Sol high for most coding tasks. Major issue is speed/efficiency. It's SLOW. And it tends to do too many checks/tests. So my major question is are there any speed/efficiency gains with the new 6.2 models?
1
1
1
1,362
Dario: "I control the world's largest model. That's safe. When other people control large models, that's dangerous. I am the only one who should control large intelligence."
Dario Amodei on open source models: “If the path continues, I think we could get to a very dangerous place.” “When you control a model and you're deploying it, you have the ability to monitor its usage.” “You can revoke a user's access. You can change what the model is willing to do.” “When a model is released in an uncontrolled manner, there's no ability to do that. It's entirely out of your hands.” “We should confront the advocates of this with that problem and challenge them to solve it.” He clearly doesn't believe Intelligence should be in the hand of the masses and Sam Altman also saying "RAM" shouldn't be accessible to everyone indefinitely agreed with him.
1
3
127