PDDL-based symbolic task planning remains pivotal for robot autonomy yet struggles with dynamic human-robot collaboration due to scalability, re-planning demands, and delayed plan availability. Although a few neurosymbolic frameworks have previously leveraged LLMs such as GPT-3 to address these challenges, reliance on closed-source, remote models with limited context introduced critical constraints: third-party dependency, inconsistent response times, restricted plan length and complexity, and multi-domain scalability issues. We present Gideon, a novel framework that enables the transition to modern, smaller, local LLMs with extended context length. Gideon integrates a novel problem generator to systematically generate large-scale datasets of realistic domain-problem-plan tuples for any domain, and adapts neurosymbolic planning for local LLMs, enabling on-device execution and extended context for multi-domain support. Preliminary experiments in single-domain scenarios performed on Qwen-2.5 1.5B and trained on 8k-32k samples, demonstrate a valid plan percentage of 66.1% (32k model) and show that the figure can be further scaled through additional data. Multi-domain tests on 16k samples yield an even higher 70.6% planning validity rate, proving extensibility across domains and signaling that data variety can have a positive effect on learning efficiency. Although long-horizon planning and reduced model size make Gideon training much less efficient than baseline models based on larger LLMs, the results are still significant considering that the trained model is about 120x smaller than baseline and that significant advantages can be achieved in inference efficiency, scalability, and multi-domain adaptability, all critical factors in human-robot collaboration. Training inefficiency can be mitigated by Gideon's streamlined data generation pipeline.

Lauren: So you are the editor in chief here at WIRED, and you've been talking a lot about AI, so I wanted to see how good you are at telling regular human-made music apart from AI-generated music. Gideon: I mean, I can barely tell music by one human apart from another sometimes. So, uh, you know, you might be disappointed, but I will do my best. Lauren: Let's hear the second one. Lauren: First, I'm curious if you know who the artist is. Gideon: I have no idea.

A quick reminder, every link you need for our materials and articles is right here on our website. Please share this article, so we can keep spreading the word. Join Bill's Twitter here, his Gab here, and our official We Are Sovereign Twitter here. Check the bottom of this post for important notes. The following was collated from a few wide-ranging conversations covering many topics. Bill: My feeling is this is a big year, and that 2022 was this transition or preparation year, and 2023 is going to be a year of implementing plans. Gideon: "The groundwork laid in 2022 will come to fruition in 2023. If a person is not prepared, they are going to become ensnared in the cabal's plans. We also believe CSRQ will come online in 2023." Bill: I hear all the time that 2025 is their big date, it's all going to happen then, all leading up to that, but you've maintained it's coming sooner than that. What will be going on then? Gideon: "By that time, CSRQ will be fully operational. The mass die-off of the vaccinated will be evident, but generally covered-up and dismissed as the effects of Covid or other things. The cabal's main focus at that point will be implementing the new vaccination agenda, which will tie into a new pandemic. Their focus will be the non-vaccinated. They know the hold-outs, or those who resisted, will not comply easily, so this is where we see a lot of focus on using the gatekeeping operations to bring them into a compliance, and possibly a point they may even be unaware they are complying. I also suspect we will see a greater popularization of Alt Media, alternative truths, where that is brought into the light more, which will lead to the deception that we are somehow winning and the cabal is losing. Meanwhile, it will be the cabal doing it in the first place, a sort of false sense of victory and security. This is already happening, on a small scale, and it's cabal-controlled in nature." Bill: That's a very interesting point and you've brought it up more and more with me, in terms of compliance. I guess I see how it's possible. Like how we just accepted debit and credit cards have these RFID chips in them, which is now ubiquitous.

Intel Drop #24 - The Threat Of A.I. Is Already Here A quick reminder, every link you need for our materials and articles is right here. Please share this article, so we can keep spreading the word. Join Bill's Twitter here, his Gab here, and our official We Are Sovereign Twitter here. Check the bottom of this post for important notes. The following was collated from a few wide-ranging conversations covering many topics. Bill: I few months ago you told me we need to keep an eye on A.I., but you didn't get much into it. Then this week you brought it up a lot more. Can you give everyone an update on that? The cabal has moved up the timeline on the use of A.I. in a way I did not expect at all." Bill: What are they doing? Gideon: "Introducing it to the public for the first time, which I did not think was something they had ready yet." Bill: But we already have A.I. surrounding us in some form or another, so what do you mean? Bill: So this is a different type of A.I. they are introducing? It is first important to understand there are two different things we must address. First, there are software-based A.I. programs. They mainly serve as automation technologies, like self-driving cars, or even simple customer support chat bots. In some cases, it is utilized to process data. None of that is a threat at all. It will never become sentient or fully replace human ingenuity, but this is where we see the concern focused. We see people talking about it, but not talking about what is the true threat. The true threat is that something is being brought into this world that the cabal will call'A.I.' but it is not."

Ed Note: Gideon is speaking at ODSC Europe 2019, see his talk "Augmented Programming" there. Over the past decade, deep learning research has led to significant advances in perceptual tasks, such as object detection, face recognition, and speech recognition. In each of these use cases, raw real-world inputs have to be mapped into a normalized representation. Deep learning has also started to lead to significant advances in natural language processing, for example, where contextual embeddings enable multi-task transfer learning. A newer, emerging field for machine learning is in its application within the production, deployment, and maintenance of software.

Stephen Colbert kicks up his feet at the Ed Sullivan Theater, where he tapes "The Late Show With Stephen Colbert." Stephen Colbert kicks up his feet at the Ed Sullivan Theater, where he tapes "The Late Show With Stephen Colbert." (Carolyn Cole / Los Angeles Times) Stephen Colbert's desktop computer monitor is ringed with reminders -- Post-it notes ("Joy is the most infallible sign of the presence of God," Pierre Teilhard de Chardin), keepsakes (cards from musicians Regina Spektor and Jack White), directives ("Ask yourself this question: Is my attitude worth catching?"), When not in use, Colbert's computer screen defaults to a live feed of the Earth taken from the International Space Station. Right now, the view has just crossed the Nile, the sun is setting and clouds are casting long shadows across the Red Sea. Colbert looks at these images whenever he's feeling anxious. There's the whole world, he tells himself.