Test Information Space

9210 bookmarks
Custom sorting
Fei-Fei Li's World Labs speeds up the world model race with Marble, its first commercial product | TechCrunch
Fei-Fei Li's World Labs speeds up the world model race with Marble, its first commercial product | TechCrunch
(How to combine setting models with characterization? Games would likely be about threats to the order, or relationships thereinr. Although other recent offerings explore alternative artifacts as scenarios, e.g. following insurgency. Is the system hardened, absorbent, or able to transition to a research mode? How to backtrack from the resultant pattern to identify the cause? Or in the hyper-realistic genres, how it is used as an aesthetic.)
·techcrunch.com·
Fei-Fei Li's World Labs speeds up the world model race with Marble, its first commercial product | TechCrunch
Claim Your Thumb Drive: First Artefact Mint Drops 11.11 on Lamina1
Claim Your Thumb Drive: First Artefact Mint Drops 11.11 on Lamina1
(For survivors of the Singularity Apocalypse. Of course, hardware is going to target those thumbs after Stuxnet and all. And one recent definition of machine consciousness was memory across systems or substrates. Not unlike TIS. The argument was that the Turing Test was social where current tech includes xenomorphs, anthromorphs, or morphogenetic layers thereof. What is the next form? Perhaps this can dredge up everything ever strained by games.)
LAMINA1
·medium.com·
Claim Your Thumb Drive: First Artefact Mint Drops 11.11 on Lamina1
From Words to Worlds: Spatial Intelligence is AI’s Next Frontier
From Words to Worlds: Spatial Intelligence is AI’s Next Frontier
(Note manifesto style, e.g. rhetorical triples. Likely for the early academic reader. Not specifically for machine parsing as in the other Digital Twin movement. This is the memory going in as well as call to action. Published papers tend to push the cognitive limits out toward 10. Should there be any additional temporary AI freezes and need to reconstitute the motivation if not ideology at the roots level. It is a question whether research is like democracy, but they tend to be individualists looking for consensus local at least to the new domain. Another issue is whether there will be a hierarchy among such projects and purges from the top down as in politics Funding is scarce from other sectors at the moment. Markets skew the directions. Deep Learning bypassed all that last time. What is next for AI?)
·a16z.substack.com·
From Words to Worlds: Spatial Intelligence is AI’s Next Frontier
Meta chief AI scientist Yann LeCun plans to exit and launch own start-up
Meta chief AI scientist Yann LeCun plans to exit and launch own start-up
(LeCun reportedly was an expert in France in the 80s while AT&T was breaking up, then he joined Bell Labs and figured out OCR using neutral networks. People had been working on print and character generation and recognition in prior decades, but the techniques were very limited. Lucent eventually acquired the Labs. He later was at NEC before NYU. Alcatel merged with Lucent afterward. Nokia has the Labs now. Incidentally, the upcoming Gemini 3 was said to be getting really good at handwriting rec, so there may be hope for all those notes. Of course, cognitive architecture is a different problem set.)
·ft.com·
Meta chief AI scientist Yann LeCun plans to exit and launch own start-up
Opinion | Use A.I. to Reinvigorate Democracy — Not Replace It
Opinion | Use A.I. to Reinvigorate Democracy — Not Replace It
(Opponents tend to use the same arguments, but select the other valence. This is what causes hallucinations. Seems to beg the alignment issue, and will sacrifice a lot of models along the way. Again the issue of tech evolution, if not posthuman existentialism. How to think about throwing a lot more voices at society? Could be like cluster analysis if not cellular automata. Or wandering through the chaos in Future Games. Would appear to appeal to human nature, but at the same time that is undergoing fast-track genetic editing advances. Recruit the new demographics. Born of tragedy as in classic narrative. Or is it Luv wondering what she would look like in the Shoulder of Orion outfit?)
·nytimes.com·
Opinion | Use A.I. to Reinvigorate Democracy — Not Replace It
Baidu just dropped an open-source multimodal AI that it claims beats GPT-5 and Gemini
Baidu just dropped an open-source multimodal AI that it claims beats GPT-5 and Gemini
(Issue becomes which materials immediately lend themselves to this type of analysis, or what prior treatments might be useful. Then, if time-travel is no longer censored, how generative can introduce elements for a contextual vision, whether past or future. Hypothesize for falsification or predictions BTW, this is multimodal, and in analysis of Big Tech dialectics on that front, or how China may try to beat the iPhone, charts with a variety of these models yield a framework like Coordination, Cognitive, Compute, and Civic, Baidu made its name in a leading LLM and autonomous driving and is affecting regulations. ChatGPT called China wired body, where the US was scaled brain, and predicted a Continental-like synthesis. Although it may still seem like a poker game bto the user ased on CoreWeave's situation. Is it AI's responsibility to solve society's aporias, and vice-versa? Where's the lightning, or is it all toast?)
·venturebeat.com·
Baidu just dropped an open-source multimodal AI that it claims beats GPT-5 and Gemini
Baby Shoggoth Is Listening - The American Scholar
Baby Shoggoth Is Listening - The American Scholar
(Article's style is not an exact match for the suggestion since it gets personal background into the content, which is a criticism, in preparation for an eventual Digital Twin since AI is no longer your grandfather's search engine. Carnegie vs. Convergence. Or for drone download of Google Maps outline, top, street, and POV. Right now those include timestamps, but eventually they may also show preferences, like for Franklin before Mather, the Privvy Council, or Adams.)
·theamericanscholar.org·
Baby Shoggoth Is Listening - The American Scholar
Stop Coding and Start Planning
Stop Coding and Start Planning
(Caveat: organizations may consider this Shadow AI if they have standards to meet But, in cases of information gathering, decision-making, problem-solving, and so forth, the practice is growing.. The long shutdown may have teased this for the government sector as well. Perhaps an agent for the DOGE memo. Looking back, without the rise of the Perceptron in '57, AI would be very different.)
·every.to·
Stop Coding and Start Planning
How to Build a Career in AI Policy
How to Build a Career in AI Policy
(Uncertainties invite proxies, e.g. models. Those are used by professions, startups, organizations, sovereignties, or agents. They become more refined and specialized over time. Some of the current excitement is that individuals are now able to make these to scale up their skill or complexity. One issue is how this relates to tech evolution. Fitness appears to be ruthless, but somehow things are still distributed efficiently. The quantum methods look at the waves of probability, but hint that there are imaginary universes acting in parallel independent of the observer. Will machines see the same reality? Expect people to be running around trying to fix their limits.)
·youtube.com·
How to Build a Career in AI Policy
Google Workspace on X: "Available now: Gemini in @googledocs provides answers and writing suggestions based on the sources linked in your Doc. So you get answers that are grounded in the decks, reports, and spreadsheets you trust - all without ever leaving your Doc. Learn more → https://t.co/D9dMfVcwaA https://t.co/8Vqhr1p8bT" / X
Google Workspace on X: "Available now: Gemini in @googledocs provides answers and writing suggestions based on the sources linked in your Doc. So you get answers that are grounded in the decks, reports, and spreadsheets you trust - all without ever leaving your Doc. Learn more → https://t.co/D9dMfVcwaA https://t.co/8Vqhr1p8bT" / X
·x.com·
Google Workspace on X: "Available now: Gemini in @googledocs provides answers and writing suggestions based on the sources linked in your Doc. So you get answers that are grounded in the decks, reports, and spreadsheets you trust - all without ever leaving your Doc. Learn more → https://t.co/D9dMfVcwaA https://t.co/8Vqhr1p8bT" / X
Sam Altman on X: "Yesterday we did a livestream. TL;DR: We have set internal goals of having an automated AI research intern by September of 2026 running on hundreds of thousands of GPUs, and a true automated AI researcher by March of 2028. We may totally fail at this goal, but given the" / X
Sam Altman on X: "Yesterday we did a livestream. TL;DR: We have set internal goals of having an automated AI research intern by September of 2026 running on hundreds of thousands of GPUs, and a true automated AI researcher by March of 2028. We may totally fail at this goal, but given the" / X
·x.com·
Sam Altman on X: "Yesterday we did a livestream. TL;DR: We have set internal goals of having an automated AI research intern by September of 2026 running on hundreds of thousands of GPUs, and a true automated AI researcher by March of 2028. We may totally fail at this goal, but given the" / X