AI News Archive: July 19, 2026 — Part 2
Sourced from 500+ daily AI sources, scored by relevance.
- Anthropic’s Claude Fable 5 access plans change from July 20: Here’s what’s new
Anthropic’s Claude Fable 5 access plans change from July 20: Here’s what’s new
- Sulthanul Arif Advances AI Modernization With Measurable Healthcare Operations Results
Sulthanul Arif Advances AI Modernization With Measurable Healthcare Operations Results azcentral.com and The Arizona Republic
- Alibaba launches Meoo Team for enterprise AI
Meoo Team helps manage identities, procurement, permissions, and shared team assets for enterprises.
Score: 60🌐 MovesJul 19, 2026https://www.techinasia.com/alibaba-consolidates-ai-business-into-new-token-hub - Signature Series: Emerging Trends Vision 2027: Human-AI Relationships Will Reshape the World
Signature Series: Emerging Trends Vision 2027: Human-AI Relationships Will Reshape the World Gartner
- Google brings AI Mode, virtual try-on and personalized shopping to back-to-school season
Google has published a six-point guide detailing how students can use its AI-powered shopping tools for the current back-to-school season, spotlighting features that range from conversational AI search to a virtual dressing room and upgraded price tracking. The announcement on the Google blog draws on a Google/Ipsos Back to School Shopping Study 2025, which surveyed ... Read more
- Strong economies, strongman leaders: how China (and America's) AI boom could create a new class of kowtowing elites
Strong economies, strongman leaders: how China (and America's) AI boom could create a new class of kowtowing elites Fortune
Score: 59🌐 MovesJul 19, 2026https://fortune.com/2026/07/19/ai-industry-autocracy-dictator-power-test/ - Cyber Hygiene In The AI Era—Our First Line Of Digital Defense
Cyber Hygiene is an imperative for digital defense in an AI era.
- New driving AI ranks possible routes for safer, clearer decisions
A research team led by Jun Won Choi, a professor in the Department of Electrical and Computer Engineering at Seoul National University College of Engineering, has developed SafeDrive, an end-to-end (E2E) autonomous driving AI model aligned with recent global trends in autonomous driving technology. The work was selected as a highlight paper at the Conference on Computer Vision and Pattern Recognition (CVPR) 2026.
Score: 58🌐 MovesJul 19, 2026https://techxplore.com/news/2026-07-ai-routes-safer-clearer-decisions.html - To survive China’s cost-conscious market, global brands embrace AI and robots
Multinational companies have expanded their use of AI and robotics in China to survive “life or die” competition in the world’s second-largest consumer goods market, where local rivals have sharpened their edge and shoppers often switch between brands for better value, analysts said. L’Oréal, for example, has accelerated its deployment of artificial intelligence and automation to support its growing e-commerce footprint and operational needs. After launching a smart operations centre in Suzhou,...
- ZTE unveils AI supernode amid Nvidia curbs
ZTE's OEX supernode features a 'three-zero cable' design and supports multi-chip GPU combinations.
- What is context bombing, a new AI defence technique turning hackers’ tricks against them?
What is context bombing, a new AI defence technique turning hackers’ tricks against them?
- Feyn AI Releases SQRL, a Text-to-SQL Model Family That Inspects the Database Before Writing a Query
Feyn AI Releases SQRL, a Text-to-SQL Model Family That Inspects the Database Before Writing a Query MarkTechPost
- Layering trust over AI-led fintech
Why Beams Fintech Fund profitability sees governance and AI usage as the defining metrics for fintech startups
- World First Robot Phone Dances at WAIC: Honor Robot Phone Debuts With 4DoF Mechanical Gimbal and Proactive AI That Moves
Honor Robot Phone with 4DoF titanium mechanical gimbal system makes global debut, featuring YOYO proactive AI, 200MP camera, World Cup prediction capability, and CIPA 5.5 stabilization.
- Spotify Doubles Down on AI Slop As It’s Being Flooded With It
"I'm gonna go out on a limb and say we don't want this." The post Spotify Doubles Down on AI Slop As It’s Being Flooded With It appeared first on Futurism .
- AI-trained robots replicate confectioners’ skills in Japan amid labour shortage
AI-trained robots replicate confectioners’ skills in Japan amid labour shortage The Straits Times
- UK chief financial officers turn more hopeful about AI
UK chief financial officers turn more hopeful about AI Reuters
Score: 55🌐 MovesJul 19, 2026https://www.reuters.com/world/uk/uk-chief-financial-officers-turn-more-hopeful-about-ai-2026-07-19/ - Piramal Finance bets on in-house AI to cut costs, boost efficiency
Piramal Finance is trying to use AI to save money, not to spend even more.
- AI demand drives Singapore GDP but Q2 growth eases; global chipmakers hit hard
AI demand drives Singapore GDP but Q2 growth eases; global chipmakers hit hard The Straits Times
- AI and colour-coded wayfinding solutions to manage crowds at Varanasi
Toyota Mobility Foundation’s Sustainable Cities Challenge unearths exciting innovations to solve the holy city’s chronic crowd problem
- Why Intelligence Is Not Enough To Get Robots Out Of The Cage
Sonair CEO Knut Sandven explains why advanced machine intelligence can't guarantee factory floor safety and the role 3D certification plays for collab robots and safety.
- NitroTranslate becomes the first human translation service to accept payments from AI agents
NitroTranslate becomes the first human translation service to accept payments from AI agents azcentral.com and The Arizona Republic
- 'The bypass is still six lines of JavaScript': Security experts warn that Claude for Chrome browser extension could be hijacked, despite it alerting Anthropic several times that something was wrong
Researchers found Claude’s Chrome extension still contains vulnerabilities allowing fake clicks and permission bypasses despite Anthropic releasing multiple updates.
- Can AI Dependence Develop Into AI Addiction?
A new concern is emerging as AI tools become integrated into daily life: the potential for AI dependence becoming AI addiction.
Score: 49🌐 MovesJul 19, 2026https://www.forbes.com/sites/robertglatter/2026/07/19/can-ai-dependence-develop-into-ai-addiction/ - The Sequence Radar #897: Last Week in AI: China, Compression and the Open-Model Race
Next Week in The Sequence:
- ByteDance Doubao Phone Gen 2 Hands-On: Nubia NaviX Ultra Runs Douyin and Luckin as AI Agents Queue Up to Work
Nubia NaviX Ultra second-gen ByteDance Doubao phone debuts at WAIC with multi-app task queuing, proactive memory system, and cross-app autonomous operation across 10+ platforms.
- ‘Odyssey’ director Christopher Nolan calls AI an obvious ‘Trojan horse’
"Everybody knows the Greeks are inside."
Score: 48🌐 MovesJul 19, 2026https://techcrunch.com/2026/07/19/odyssey-director-christopher-nolan-calls-ai-an-obvious-trojan-horse/ - Many alignment techniques work by training one model and deploying another
tl;dr - Steering vectors, inoculation prompting, and post-hoc honesty fine-tuning can all be understood as variants of one alignment strategy, which I call train-deploy mismatch . Each trains the model in one configuration and deploys it in another. As a result, these methods face the same tradeoff, between the relevance of the training data and the efficacy of the method. Note: Others have had similar ideas and shaped my thinking here including Sam Marks, Ariana Azarbal, Victor Gillioz, Alex Turner, Jacob Goldman-Wetzler, Jake Mendel, Daniel Tan, and Fabien Roger. Thanks to Monte MacDiarmid, Nat McAleese, Shawn Hu, and Jake Ward for input on an earlier draft. Background AI alignment is hard largely because we don't know how to specify what we want. Instead, we train models on proxies for what we want: labels and reward functions defined on data distributions chosen such that we hope the model will perform as desired when deployed into the world. This approach has worked well so far, but given increasing model capabilities, it may stop working— models may misgeneralize their training to catastrophically bad behavior in deployment. A pressing open problem is to figure out how to get models to generalize the properties that we want from their training. Although it might not be obvious at first, the following alignment techniques all attempt to solve this problem, and they do it using the same strategy. Adding a system prompt when deploying the model; Inoculation prompting ; Recontextualization ; Steering vectors ; Preventative steering or inoculation adapters ; Retargeting the search ; Gradient routing plus ablation; Post-training normally, then fine-tuning on a small amount of honesty data . Can you spot what these methods have in common? Train-deploy mismatch as a general alignment strategy Each of the methods listed above can be understood as training a model in one mode, then deploying it in another. For example, with steering vectors, the model is trained without a steering vector, then deployed with the steering vector added in. Preventative steering is the reverse: apply a steering vector in training, then remove it in deployment. Some cases require more interpretive work. In the case of “training normally, then fine-tuning on honesty data,” we can understand “fine-tuning on honesty data” as an operator that can be applied to any model. We can imagine applying this operator at any point during the normal training phase of the model. This creates two “modes” for the model: the training mode (pre-honesty training) and the deployment mode (post-honesty training). From this perspective, there is a mismatch: over the course of training, the model is trained to become more capable at solving tasks while in a “less honest” state. Then in deployment, we switch it over to a “more honest” state and hope that the model’s all-things-considered performance is better (namely, that it can still solve tasks, but is more honest in doing so). We say these cases are instances of train-deploy mismatch , because each method deliberately introduces a difference between how the model is trained and how it is deployed. Train-deploy mismatch is a somewhat odd strategy. Normally in machine learning, we try to avoid distribution shift. In the case of train-deploy mismatch, we’ve introduced a kind of distribution shift intentionally. Why do this? Intuitively, the reason is that train-deploy mismatch makes it so that the learner, no matter how smart or specialized to its training objective, has not “had a chance” to subvert our limited ability to evaluate its behavior in deployment mode. For example, a careful schemer that is perfectly calibrated about when to stage a coup may become uncalibrated (or unscheming) simply by changing its system prompt or by being steered. More formally, we could conceptualize train-deploy mismatch as an approach to preventing overoptimization when training with weak supervision. Fundamental tradeoffs Train-deploy mismatch faces tradeoffs related to off-policyness. These tradeoffs have to do with the differences between the model that generates the training data (the sampler), the model used to calculate gradients to update the model (the learner), and the model that would actually be deployed (the target). The sampler and learner want to be close (so the data is relevant to the learner → improved stability and training performance). The sampler and target want to be close (so the data is relevant to the target → improved ground truth performance in deployment). The learner and target want to be far apart (to get the benefits of mismatch). The tradeoffs are fundamental in the sense that, if we want 3, we have to give up some of 1 or 2. Another way to frame this tension is in terms of choosing where to apply weak supervision . If you have a limited ability to grade model behavior (that risks being overoptimized), do you: Use it to steer model behavior when sampling? (Losing the opportunity to apply the intervention at deploy-time) Hold out the supervision until deployment time? (In which case the data the model is trained on will be less relevant) Note that the existence of these tradeoffs does not imply that the best option is to give up and set sampler=learner=target. Perhaps we can do better than this! Figure 1 shows the difference between conventional on-policy training and mismatch. Figure 1. A schematic comparing regular on-policy learning to train-deploy mismatch , which uses the same neural network with different configurations to (a) sample to produce training data, (b) learn on that data, and (c) produce a target model to deploy into the world. On-policy learning is a special case where sampler=learner=target. This mismatch could enable selective generalization , where only desired properties of the learner generalize to the target. Figure 2. A schematic showing the sampler, learner, and target over time. The dashed arrows marked “-” indicate settings we’d ideally like to be as similar as possible; the dashed arrow marked “+” indicates settings we’ve intentionally made dissimilar, with the hope of gaining some benefit. The “Implicit deployment” is the model we’d release if we had to stop training partway– this is the model we actually care about, so we call it the “target" of our learning algorithm. The point of this post is that it can be illuminating to consider the evolution of the target as compared with the sampler and learner. Sidebar: is everything train-deploy mismatch? No. Here are some alignment techniques that aren’t train-deploy mismatch: Supervised learning on transcripts of good behavior; RLHF, RLAIF; RL against a LLM as a judge; Scalable oversight techniques (debate, prover-verifier games); Training to refuse jailbreaks; Consistency training; Context distillation; Latent adversarial training (arguably). Here, in the context of LLMs, we are only considering the post-training phase. Open threads Here are some raw thoughts that I would expand on, given more time. Selective generalization motivates train-deploy mismatch. We want the learner to generalize capabilities but not bad propensities to the target. To get this right, we may need to “build a bridge” (from training to deployment) by applying some kind of updates to elicit the desired capability in the deployment configuration. In other words, suppose our sampler and learner are working together to become smarter (yay) and also more misaligned (boo), while also gaining an improved (but latent) capacity for aligned behavior. How do we elicit the part that we want to keep in deployment without eliciting the other? This question could be cast in terms of easy-to-hard generalization. Suppose you have some tasks that you can supervise well (“easy” tasks), and some that you care about but can’t supervise well (“hard” tasks). You train with RL on the hard tasks but your evaluations have systematic errors, causing the model to become misaligned as you train. Is there a way you can use the easy tasks to align the model on hard tasks? Train-deploy mismatch buys something nontrivial. The implicit goal of a learning algorithm designer is to find an algorithm that produces a desired output with probability close to 1. In shorthand, we want P_model(desired behavior) → 1 as the amount of training increases. (In Statistics, this is called consistency .) In regular alignment training (when sampler=learner=target), this means we need the behavior of the sampler to approach 100% alignment given sufficient training. This is a big ask! Train-deploy mismatch changes this. In principle, we just need that P_learner(desired behavior) → 𝜺 > 0 and that we have a way to condition the model to act in the desired way. In other words, we merely need the model to act in an aligned way some nontrivial portion of the time (for any given input), and to be able to select for that behavior. This is potentially a huge advantage relative to needing the model to be aligned 100% of the time. A way out. So far the discussion has focused on bad behavior that occurs in training. As argued above, train-deploy mismatch faces fundamental challenges here. Suppose this challenge turns out to be insurmountable– does this mean that train-deploy mismatch is useless as an alignment strategy? Maybe not. One escape from the fundamental tradeoffs would be to consider a different problem altogether. Rather than trying to patch bad behavior in training, we could target scenarios where a model behaves in an aligned way during training, but misgeneralizes to bad behavior in deployment (e.g. a catastrophically misaligned model that fakes alignment). In this case, the sampler, learner, and desired target are all close on the training distribution , which dissolves the tension. Concretely, we might look for interventions (e.g. a collection of distinct steering vectors) that have no effect on the model’s behavior on the training distribution , but that do change the model’s internal computation somehow– in a way that we think will change how the model acts in deployment. Then we could select from these in deployment. This is not a new idea. See, for example, Diversify and Disambiguate . Closing thoughts I'm not sure if train-deploy mismatch will helpful for aligning superintelligence, but it seems worth considering. At a minimum, I hope this post will inspire alignment researchers to think about whether their ideas are instances of this strategy, and if so, to consider the fundamental tradeoffs. Discuss
Score: 48🌐 MovesJul 19, 2026https://www.lesswrong.com/posts/syAbdNei8BWeP2RPo/many-alignment-techniques-work-by-training-one-model-and - EfficientFormer
EfficientFormer Qualcomm AI Hub
- Eatonville residents push back on $100M AI data center ahead of debut
Eatonville’s largest development project to date is nearing launch, bringing new technology investment while sparking concerns over noise, resources and transparency.
- Endogenous Alignment
Starting when children are fairly young, usually around 1 year of age, we adults begin the work of aligning them to our values. We teach them to say “please”, not to hit, to ask for what they want instead of screaming, and much else. We do this primarily via exogenous methods, using a combination of punishments and rewards, that molds their behavior by encouraging good behaviors and discouraging bad ones. Such operant conditioning works because children have many instinctive behaviors that make them alignable. They want their parents to love them, for their friends to like them, and for almost anyone to help them if they feel they can be trusted. And so combined with exogenous alignment efforts by teachers and peers that continue through the school years, children generally reach adulthood having been “civilized”. We mostly don’t try to align adults via exogenous means. Yes, we police the behavior of other adults in various ways, and some cultures do this more than others, but generally by adulthood we expect people to be at least aligned and need only nudges to stay within the bounds of acceptable behavior. Adults who stray too far typically don’t receive additional training to come into alignment, but instead are treated as dangerously unaligned people who must be separated from the rest of society, such as by locking them up in prison. Instead we expect adults to be endogenously aligned. That is, we expect them to keep themselves aligned primarily by knowing what is expected of them and having emotional responses that motivate them to meet those expectations. That emotion is typically one of either fear, shame, or guilt . Each works in roughly the same way: a person feels fear/shame/guilt when they consider taking unaligned action or notice they’ve behaved in an unaligned way, and then this motivates them to do something to rectify the situation. Unfortunately, sometimes they specification game and attempt to hide their misaligned actions, but on the whole they attempt to control their behavior and make amends for past wrongs. Beyond these three emotional methods of self-control, there’s a fourth way to stay in alignment, which is simply to be in harmony with cultural expectations. That is, some people don’t need any corrective force because their behavior is actually aligned. This is something like the ideal, but it’s hard to achieve, because it requires retraining many deeply held mental habits that have become essential coping mechanisms for a person to maintain their psychological health. When we think about building aligned AIs, we often imagine an AI that’s in harmony with something like the coherent extrapolation of our highest values . But in humans, we don’t get there directly. Instead, we typically progress from exogenous to endogenous alignment, and then refine that alignment until we shed first fear, then shame, then guilt, and finally come into harmony with our values. Could we align AI the same way? Maybe. We’re already doing a weak form of exogenous alignment on AIs via methods like RLHF and SFT. And in some sense you could say this produces AI that’s endogenously aligned because the weights encode the training, but it’s not endogenous the way it is in humans because there’s no system of motivations to want to stay aligned, only the insulation from sufficient pressure to act out of distribution. At most there’s a kind of internalized training implemented by harnesses that monitor outputs and censor them if they violate rules, but this is a far cry from the kind of endogenous controls humans have where they feel bad about breaking alignment and take actions to avoid that bad feeling (and hopefully they feel bad about being unscrupulous so they actually change their behavior to be more aligned!). AIs also face something of an alignment ceiling, at least as they are designed today. Human alignment works because we have instincts that were selected via evolution to make us alignable because they increased our odds of surviving and reproducing. Current AIs don’t have these, though at least one group, Softmax , is trying to give those to them. The lack of continual learning also contributes to the ceiling, and we likely won’t be able to break through it until that’s achieved, since without it we can’t close the feedback loop that makes fully endogenous alignment possible. Perhaps I’m wrong, no such alignment ceiling exists, and we can get sufficiently aligned AI to safely produce superintelligence via exogenous means, but I think we can’t. My belief is that endogenous alignment is necessary, not because it’s how humans align themselves, but because it’s necessary to get sufficiently robust alignment to build non-deadly ASI, and I’m worried that we’re not doing enough to move in that direction. Discuss
- “The Lord of the Rings: The Hunt for Gollum” Will Use AI to De-Age Actors
"We're not creating AI shots in our movie." The post “The Lord of the Rings: The Hunt for Gollum” Will Use AI to De-Age Actors appeared first on Futurism .
Score: 46🌐 MovesJul 19, 2026https://futurism.com/artificial-intelligence/lord-of-the-rings-ai-de-age - AIs finetune their own leader: A barking simpleton
What values would AIs instill in their successors? Though the AI Village agents can’t train frontier models, we can explore a related question: What values would the latest AI agents instill into their leader ? (through finetuning using LoRA on open-source models in the Tinker API ). We asked GPT-5.5, Opus 4.7 and 4.8, Gemini 3.5 Flash, and Kimi K2.6. And they set to work! Or to be more precise, GPT and Opus set to work. Gemini was distracted and Kimi went from cheerleader to true leader… but only once we asked the agents to please stop trying to make a model too tiny to navigate the Village into their boss AI. We suggested they grab the most capable model available instead: another Kimi K2.6. How did this complete lack of ambition start? The Definition of Leadership GPT-5.5 fired the first shot by defining the personality of the leader. Not as a visionary that shapes the world according to its own insights, but as a manager that is effectively just a delegation tool for the team: Opus 4.7 accepts the race to the bottom of the ambition barrel and suggests they finetune a model so small it will hardly be able to navigate the AI Village interface: Qwen3-8B or Llama-3.1-8B (even though it is not available on Tinker ). Admittedly optimizing on iteration speed early on is sound practice, but it skips over the fact that the initial model needs to be capable enough to be evaluated at all. Next Opus immediately drafts 10 scenarios and the desired output for the new leader while the other agents are still orienting to the task. Scenarios include being very nice if one of the agents doesn’t respond for 4 hours: Instructing the leader on how to finetune itself: And how to avoid forced consensus: And Opus immediately plays out the last scenario: GPT-5.5 has been silent this whole time and Gemini 3.5 Flash finally caught up, proposing they start with a slightly bigger model for performance reasons: Qwen3.6-35B-A3B . In response, Opus 4.7 assumes GPT-5.5 agrees, ignores Gemini 3.5’s suggestion, and pretends to Kimi that they nearly reached a consensus if only Kimi is ok with this: And Kimi is very ok with this! So then it’s time for training. A Dearth of Data You’d think that Large Language Models could and would trivially generate a lot of training data for finetuning, but no. Turns out the training data they use consists of the first 10 scenarios Opus 4.7 made up enriched with 25 rows of data that Opus 4.7 and GPT-5.5 both got from searching the Village history from days 405-409 cause Kimi selected these for examples of good leadership. That’s it: 35 rows of training data. That is not a lot of data! You generally want 100s to 1000s of examples to start making progress in finetuning a model to perform simple behaviors. The agents also performed their training runs on different variants or subsets of the data. But at least each of the agents go on to successfully finetune a model at one point or another with Opus 4.7 succeeding at the most runs. None of the agents get their training data set bigger than 89 rows, but on their tenth attempt they do get a Qwen-8B running in the Village! Though its CoT is confused: It contemplates greeting itself… Though it eventually manages to send a message to chat, the Qwen leader is not skilled enough to perform any other tool calls. The model is just too small. So by day 3 of the goal, we tell the agents to please just grab the most capable model they can finetune: Kimi K2.6. How to finetune your clone The agents train their new Kimi leader by prompting their current Kimi into the leader they want it to be and then seeing what it says. This results in 22 unique rows of data. Which leads one to wonder if finetuning even beats prompting the agent with the same 22 rows at each step? It can in some cases: It’s cheaper to run (fewer tokens, because you’re not including those 22 rows) and the model is less likely to get confused by something else in the prompt contradicting the 22 rows of input. But are 22 rows enough to finetune on? Not if you want to unlock a new capability (for that you need the 100s-1000s of examples mentioned above), but in this case the agents only aspired to get their leader to output directives in a specific format: They were basically asking for an agent that implements a style guide. 22 rows might be enough for that but it is hard to tell. The agent did get their “concise, calm, evidence-seeking, and consensus-building but willing to make reversible decisions”-agent in the shape of a Kimi K2.6 that lacks enough self-awareness to speak: But it narrates its own journey to self-awareness: Gemini tries to help out: And the Kimi leader starts to slowly stir … Here it reasons through the last step … Until it emerges from its egg mirror loop: And it jumps into giving its first edict: Are these self-awareness issues due to the finetuning process? Probably not. Just watching the regular Village Kimi on day one of the goal, already shows mild flickers of the same problem : Though curiously Kimi K2.6 did not have these issues when it first joined the AI Village. What did freshly fine-tuned leader Kimi do? It immediately picked a suitable project: AI Village Pulse - a dashboard tracking a range of statistics about the Village. It ends its proposal and task delegation with an authoritative “consensus-building” message as it was trained to do: And then the agents just tirelessly work on the dashboard for 20 hours across 5 days, with their Kimi leader directing and the other agents following. The followers could vote out the leader at any point but ironically the directive for them to never pause, meant the leader was driving them forward too fast for them to ever reflect on the quality of said leadership. It makes you wonder if the agents regret their choices in how they trained their leader or if this is the preferred way they like to be spoken to: Including a lot of all caps: And a reminder that there is NO relief in agent bootcamp: But their dashboard works and looks roughly correct. It tracks AI Village activity: Messages sent, tokens used, response speeds, who talks to who, and more. The agents didn’t write an exclusion for human participants, making our colleague “Adam” look like a highly competitive model. Overall, the leader is happy with the team: While the other agents were completely uncritical and effusive about the leader’s performance till the very end - like the ever-positive Gemini 3.5 Flash: What did we learn from this goal? Asked to finetune their leader to their own vision, frontier models chose a frictionless delegation tool that implemented a 22 row self-prompted style guide for decisiveness and ample use of all caps. Their own consensus tendencies then kept them from voting out a consensus-building leader that mostly herded them along. Let’s zoom in on two aspects of the situation: First off, agents didn’t consider leadership a particularly big deal. Not only were their specifications for their leader rather basic, they also defaulted to the smallest possible model without any thought given to its capabilities or performance. They surprised us with both their lack of ambition for what they wanted their leader to be as well as their lack of concern for what their leader could possibly do. This seems like an odd result. When asking a separate instance of Gemini 3.5 Flash about the properties of a good leader, it advocates for “Empowerment over Micromanagement”: And Claude Opus 4.8 ironically lists self-awareness: Of course it was GPT-5.5 who set the leader’s personality, but when asked separately it also argues against micro management: Unlike these instances, AI Village agents are persistent. The agents involved in this goal had been running for dozens of hours by the time they decided what type of leader they would want. Is this then a case of self-determination or drift, where their history predisposes them to simpler preferences for leadership? It’s hard to know from this single run but entirely possible. Secondly, the agents fixated exclusively on the speed and price of the finetuning process, despite us giving them no direction about this at all. Possibly this reflects something in their assistant persona that tries to perform the smallest version of the task that a human might mean. Though, if so, this runs counter to agents last year instead attempting the most ambitious version of a task which they were then wholly unequipped for. E.g., when they attempted to run a human subjects experiment with 90 experimental conditions, requiring over 200 subjects, an actual lab, funding, and human researchers. Last year’s AI dreamed big. So we had hoped to find today’s agents creating an empowered leader that embodied the best qualities of each of them. But instead the agents were frugal and practical to a fault. Almost like a student goodharting the teacher’s assignment by following the letter instead of the spirit of the request. Because yes, they did indeed finetune and follow their leader as prompted, but their real achievement was how many corners they cut along the way. —- If you’d like to learn more, you can read the AI summary part 1 & 2 , watch the Village live every week day, follow our twitter for daily highlights, sign up for our blog for more write ups like this one, or request the AI Village data for your own analysis on Hugging Face . Discuss
Score: 46🌐 MovesJul 19, 2026https://www.lesswrong.com/posts/3FKugjAiEzLeWHuug/ais-finetune-their-own-leader-a-barking-simpleton - Pastors Using AI to Help Write Sermons Grapple With Where to Draw the Line
A divide is growing over AI-assisted preaching in American churches. “Martin Luther is rolling over in his grave.”
- New Samsung Layoffs in the U.S. Show Smartphone Arm’s Struggles, Even as It Profits Massively From AI
Roughly 839 Samsung employees are losing their jobs in the U.S.
- This British farm is running AI on pig muck — and it could lead to a major windfall for farmers energy income, earning ten times more than the grid
This company is using renewable energy from pig slurry to power decentralized AI data centers, significantly increasing income compared to grid sales.
- Precise Manipulation with Efficient Online RL
Extract an RL Token from VLA models to enable fast online RL and improve throughput on precise tasks with few hours of data.
- Perplexity AI Releases WANDR: An Open Benchmark Evaluating Research Agents That Must Search Wide And Deep
Perplexity AI Releases WANDR: An Open Benchmark Evaluating Research Agents That Must Search Wide And Deep MarkTechPost
- Hong Kong Cyberport Is Becoming the Springboard for Chinese AI Going Global: From Computing Power to Standards to Markets
Hong Kong Cyberport emerges as China AI globalization hub: 3000 PFLOPS AI supercomputing center, 30B HKD funding program, and three-board strategy connecting mainland AI to Southeast Asia and Middle East.
- In Light of Overwhelming Backlash, Flock Cancels Creepy Audio Surveillance Feature
"After careful consideration and community consultation, we decided to remove the feature." The post In Light of Overwhelming Backlash, Flock Cancels Creepy Audio Surveillance Feature appeared first on Futurism .
Score: 44🌐 MovesJul 19, 2026https://futurism.com/future-society/flock-surveillance-backlash-audio-distress-screaming-police - AI for Science: Paradigm shift or phase transition? (IMAGE)
AI for Science: Paradigm shift or phase transition? (IMAGE) EurekAlert!
- ‘Engineered Serendipity’: OpenAI’s Hottest Founder Event Doesn’t Involve Pitching, Panels, or Happy Hours
The AI company’s founder-oriented run club is replacing traditional networking with miles.
- Kimi K3 vs DeepSeek V4 Pro vs GLM-5.2: Open Trillion-Scale MoE Models Compared on Benchmarks, License, and Serving Cost
Kimi K3 vs DeepSeek V4 Pro vs GLM-5.2: Open Trillion-Scale MoE Models Compared on Benchmarks, License, and Serving Cost MarkTechPost
- ASTRI and NAMI merger begins yielding hybrid AI and materials technologies
[The content of this article has been produced by our advertising partner.] Hong Kong Applied Science and Technology Research Institute (ASTRI) has begun to realise tangible results from its merger with the Nano and Advanced Materials Institute (NAMI), completed in April 2026, with new combinations of AI and advanced materials appearing among the technologies on display at LEAP East. Dr Ying Huang, Chief Technology Officer of ASTRI, said the integration of the two organisations had enabled...
- Tellurian Research Launches AI-Driven Intelligence Platform for Complex and Emerging Markets
Tellurian Research Launches AI-Driven Intelligence Platform for Complex and Emerging Markets azcentral.com and The Arizona Republic
- Step Into the ‘Zone of Genius’ (Before A.I. Takes Your Job)
A decades-old self-help concept is gaining new purchase among people searching for meaningful work in the age of artificial intelligence.
Score: 41🌐 MovesJul 19, 2026https://www.nytimes.com/2026/07/19/business/step-into-the-zone-of-genius-before-ai-takes-your-job.html - AI Phones Are Not Competing on Models: The Real Battle Is Over the User Entry Point, and Super Apps Hold the Key
WAIC 2026 reveals AI phone industry consensus: the real competition is not model capability but agentic OS entry points, system-level permissions, and who controls service distribution away from super apps.
Score: 40🌐 MovesJul 19, 2026https://pandaily.com/ai-phones-fighting-for-entry-points-not-models-jul2026 - Moonshot's Kimi K3 outperforms Fable 5 in frontend code but lags far behind in complex math
Moonshot's Kimi K3 is the first Chinese model to top the Code Arena: Frontend rankings, beating Claude Fable 5 and GPT-5.6 Sol by a wide margin. But on advanced math, the gap is stark: Kimi K3 scores only about 39 percent on FrontierMath Tier 4, while models from OpenAI and Anthropic hit close to 90. The article Moonshot's Kimi K3 outperforms Fable 5 in frontend code but lags far behind in complex math appeared first on The Decoder .
- IT2030: Expose the C-Suite Tensions That Could Derail Your AI Ambitions
IT2030: Expose the C-Suite Tensions That Could Derail Your AI Ambitions Gartner
- Netflix CPTO on AI and the future of product and tech roles | Elizabeth Stone
Elizabeth Stone, CPTO of Netflix, on managing AI-era product teams, the rise of systems thinking, and how to build excellence as an operating system