Committee

In this tutorial, we show how we treat prompts as first-class, versioned artifacts and apply rigorous regression testing to large language model behavior using MLflow. We design an evaluation pipeline that logs prompt versions, prompt diffs, model outputs, and multiple quality metrics in a fully reproducible manner. By combining classical text metrics with semantic similarity and automated regression flags, we demonstrate how we can systematically detect performance drift caused by seemingly small prompt changes. Along the tutorial, we focus on building a workflow that mirrors real software engineering practices, but applied to prompt engineering and LLM evaluation. Check out the FULL CODES here. Copy CodeCopiedUse a different Browser !pip -q install -U “openai>=1.0.0” mlflow rouge-score nltk sentence-transformers scikit-learn pandas import os, json, time, difflib, re from typing import List, Dict, Any, Tuple import mlflow import pandas as pd import numpy as np from openai import OpenAI from rouge_score import rouge_scorer import nltk from nltk.translate.bleu_score import sentence_bleu, SmoothingFunction from sentence_transformers import SentenceTransformer from sklearn.metrics.pairwise import cosine_similarity nltk.download(“punkt”, quiet=True) nltk.download(“punkt_tab”, quiet=True) if not os.getenv(“OPENAI_API_KEY”): try: from google.colab import userdata # type: ignore k = userdata.get(“OPENAI_API_KEY”) if k: os.environ[“OPENAI_API_KEY”] = k except Exception: pass if not os.getenv(“OPENAI_API_KEY”): import getpass os.environ[“OPENAI_API_KEY”] = getpass.getpass(“Enter OPENAI_API_KEY (input hidden): “).strip() assert os.getenv(“OPENAI_API_KEY”), “OPENAI_API_KEY is required.” We set up the execution environment by installing all required dependencies and importing the core libraries used throughout the tutorial. We securely load the OpenAI API key at runtime, ensuring credentials are never hard-coded in the notebook. We also initialize essential NLP resources to ensure the evaluation pipeline runs reliably across different environments. Copy CodeCopiedUse a different Browser MODEL = “gpt-4o-mini” TEMPERATURE = 0.2 MAX_OUTPUT_TOKENS = 250 ABS_SEM_SIM_MIN = 0.78 DELTA_SEM_SIM_MAX_DROP = 0.05 DELTA_ROUGE_L_MAX_DROP = 0.08 DELTA_BLEU_MAX_DROP = 0.10 mlflow.set_tracking_uri(“file:/content/mlruns”) mlflow.set_experiment(“prompt_versioning_llm_regression”) client = OpenAI() embedder = SentenceTransformer(“all-MiniLM-L6-v2”) EVAL_SET = [ { “id”: “q1”, “input”: “Summarize in one sentence: MLflow tracks experiments, runs, parameters, metrics, and artifacts.”, “reference”: “MLflow helps track machine learning experiments by logging runs with parameters, metrics, and artifacts.” }, { “id”: “q2”, “input”: “Rewrite professionally: ‘this model is kinda slow but it works ok.'”, “reference”: “The model is somewhat slow, but it performs reliably.” }, { “id”: “q3”, “input”: “Extract key fields as JSON: ‘Order 5531 by Alice costs $42.50 and ships to Toronto.'”, “reference”: ‘{“order_id”:”5531″,”customer”:”Alice”,”amount_usd”:42.50,”city”:”Toronto”}’ }, { “id”: “q4”, “input”: “Answer briefly: What is prompt regression testing?”, “reference”: “Prompt regression testing checks whether prompt changes degrade model outputs compared to a baseline.” }, ] PROMPTS = [ { “version”: “v1_baseline”, “prompt”: ( “You are a precise assistant.n” “Follow the user request carefully.n” “If asked for JSON, output valid JSON only.n” “User: {user_input}” ) }, { “version”: “v2_formatting”, “prompt”: ( “You are a helpful, structured assistant.n” “Respond clearly and concisely.n” “Prefer clean formatting.n” “User request: {user_input}” ) }, { “version”: “v3_guardrailed”, “prompt”: ( “You are a rigorous assistant.n” “Rules:n” “1) If user asks for JSON, output ONLY valid minified JSON.n” “2) Otherwise, keep the answer short and factual.n” “User: {user_input}” ) }, ] We define all experimental configurations, including model parameters, regression thresholds, and MLflow tracking settings. We construct the evaluation dataset and explicitly declare multiple prompt versions to compare and test for regressions. By centralizing these definitions, we ensure that prompt changes and evaluation logic remain controlled and reproducible. Copy CodeCopiedUse a different Browser def call_llm(formatted_prompt: str) -> str: resp = client.responses.create( model=MODEL, input=formatted_prompt, temperature=TEMPERATURE, max_output_tokens=MAX_OUTPUT_TOKENS, ) out = getattr(resp, “output_text”, None) if out: return out.strip() try: texts = [] for item in resp.output: if getattr(item, “type”, “”) == “message”: for c in item.content: if getattr(c, “type”, “”) in (“output_text”, “text”): texts.append(getattr(c, “text”, “”)) return “n”.join(texts).strip() except Exception: return “” smooth = SmoothingFunction().method3 rouge = rouge_scorer.RougeScorer([“rougeL”], use_stemmer=True) def safe_tokenize(s: str) -> List[str]: s = (s or “”).strip().lower() if not s: return [] try: return nltk.word_tokenize(s) except LookupError: return re.findall(r”bw+b”, s) def bleu_score(ref: str, hyp: str) -> float: r = safe_tokenize(ref) h = safe_tokenize(hyp) if len(h) == 0 or len(r) == 0: return 0.0 return float(sentence_bleu([r], h, smoothing_function=smooth)) def rougeL_f1(ref: str, hyp: str) -> float: scores = rouge.score(ref or “”, hyp or “”) return float(scores[“rougeL”].fmeasure) def semantic_sim(ref: str, hyp: str) -> float: embs = embedder.encode([ref or “”, hyp or “”], normalize_embeddings=True) return float(cosine_similarity([embs[0]], [embs[1]])[0][0]) We implement the core LLM invocation and evaluation metrics used to assess prompt quality. We compute BLEU, ROUGE-L, and semantic similarity scores to capture both surface-level and semantic differences in model outputs. It allows us to evaluate prompt changes from multiple complementary perspectives rather than relying on a single metric. Copy CodeCopiedUse a different Browser def evaluate_prompt(prompt_template: str) -> Tuple[pd.DataFrame, Dict[str, float], str]: rows = [] for ex in EVAL_SET: p = prompt_template.format(user_input=ex[“input”]) y = call_llm(p) ref = ex[“reference”] rows.append({ “id”: ex[“id”], “input”: ex[“input”], “reference”: ref, “output”: y, “bleu”: bleu_score(ref, y), “rougeL_f1”: rougeL_f1(ref, y), “semantic_sim”: semantic_sim(ref, y), }) df = pd.DataFrame(rows) agg = { “bleu_mean”: float(df[“bleu”].mean()), “rougeL_f1_mean”: float(df[“rougeL_f1”].mean()), “semantic_sim_mean”: float(df[“semantic_sim”].mean()), } outputs_jsonl = “n”.join(json.dumps(r, ensure_ascii=False) for r in rows) return df, agg, outputs_jsonl def log_text_artifact(text: str, artifact_path: str): mlflow.log_text(text, artifact_path) def prompt_diff(old: str, new: str) -> str: a = old.splitlines(keepends=True) b = new.splitlines(keepends=True) return “”.join(difflib.unified_diff(a, b, fromfile=”previous_prompt”, tofile=”current_prompt”)) def compute_regression_flags(baseline: Dict[str, float], current: Dict[str, float]) -> Dict[str, Any]: d_sem = baseline[“semantic_sim_mean”] – current[“semantic_sim_mean”] d_rouge = baseline[“rougeL_f1_mean”] – current[“rougeL_f1_mean”] d_bleu = baseline[“bleu_mean”] – current[“bleu_mean”] flags = { “abs_semantic_fail”: current[“semantic_sim_mean”] < ABS_SEM_SIM_MIN, “drop_semantic_fail”: d_sem > DELTA_SEM_SIM_MAX_DROP, “drop_rouge_fail”: d_rouge > DELTA_ROUGE_L_MAX_DROP, “drop_bleu_fail”: d_bleu > DELTA_BLEU_MAX_DROP, “delta_semantic”: float(d_sem), “delta_rougeL”: float(d_rouge), “delta_bleu”: float(d_bleu), } flags[“regression”] = any([flags[“abs_semantic_fail”], flags[“drop_semantic_fail”], flags[“drop_rouge_fail”], flags[“drop_bleu_fail”]]) return flags We build the evaluation and regression logic that runs each prompt against the evaluation set and aggregates results. We log prompt artifacts, prompt diffs, and evaluation outputs to MLflow, ensuring every experiment remains auditable. We also compute regression flags that automatically identify whether a prompt version degrades performance relative to the baseline. Check out the FULL CODES here. Copy CodeCopiedUse a different Browser print(“Running prompt versioning + regression testing with MLflow…”) print(f”Tracking URI: {mlflow.get_tracking_uri()}”) print(f”Experiment: {mlflow.get_experiment_by_name(‘prompt_versioning_llm_regression’).name}”) run_summary = [] baseline_metrics = None baseline_prompt = None baseline_df = None baseline_metrics_name = None with mlflow.start_run(run_name=f”prompt_regression_suite_{int(time.time())}”) as parent_run: mlflow.set_tag(“task”, “prompt_versioning_regression_testing”) mlflow.log_param(“model”, MODEL)

arXiv:2602.06384v1 Announce Type: new Abstract: Producing outputs that satisfy both semantic intent and format constraints is essential for deploying large language models in user-facing and system-integrated workflows. In this work, we focus on Markdown formatting, which is ubiquitous in assistants, documentation, and tool-augmented pipelines but still prone to subtle, hard-to-detect errors (e.g., broken lists, malformed tables, inconsistent headings, and invalid code blocks) that can significantly degrade downstream usability. We present FMBench, a benchmark for adaptive Markdown output formatting that evaluates models under a wide range of instruction-following scenarios with diverse structural requirements. FMBench emphasizes real-world formatting behaviors such as multi-level organization, mixed content (natural language interleaved with lists/tables/code), and strict adherence to user-specified layout constraints. To improve Markdown compliance without relying on hard decoding constraints, we propose a lightweight alignment pipeline that combines supervised fine-tuning (SFT) with reinforcement learning fine-tuning. Starting from a base model, we first perform SFT on instruction-response pairs, and then optimize a composite objective that balances semantic fidelity with structural correctness. Experiments on two model families (OpenPangu and Qwen) show that SFT consistently improves semantic alignment, while reinforcement learning provides additional gains in robustness to challenging Markdown instructions when initialized from a strong SFT policy. Our results also reveal an inherent trade-off between semantic and structural objectives, highlighting the importance of carefully designed rewards for reliable formatted generation. Code is available at: https://github.com/FudanCVL/FMBench.

arXiv:2511.19279v3 Announce Type: replace-cross Abstract: A cognitive map is an internal model which encodes the abstract relationships among entities in the world, giving humans and animals the flexibility to adapt to new situations, with a strong out-of-distribution (OOD) generalization that current AI systems still do not possess. To bridge this gap, we introduce MapFormers, new architectures based on Transformer models, which can learn cognitive maps from observational data and perform path integration in parallel, in a self-supervised manner. Cognitive maps are learned in the model by disentangling structural relationships in the inputs from their specific content, a property that can be achieved naturally by updating the positional encoding in Transformers with input-dependent matrices. We developed two variants of MapFormers that unify absolute and relative positional encoding to model episodic (EM) and working memory (WM), respectively. We tested MapFormers on several tasks, including a classic 2D navigation task, showing that our models can learn a cognitive map of the underlying space and generalize OOD (e.g., to longer sequences) with near-perfect performance, unlike current architectures. Together, these results demonstrate the superiority of models designed to learn a cognitive map, and the importance of introducing a structural bias for structure-content disentanglement, which can be achieved in Transformers with input-dependent positional encoding. MapFormers have broad applications in both neuroscience and AI, by explaining the neural mechanisms giving rise to cognitive maps, while allowing these relation models to be learned at scale.

arXiv:2602.06412v1 Announce Type: new Abstract: Masked Diffusion Language Models generate sequences via iterative sampling that progressively unmasks tokens. However, they still recompute the attention and feed-forward blocks for every token position at every step — even when many unmasked tokens are essentially fixed, resulting in substantial waste in compute. We propose SureLock: when the posterior at an unmasked position has stabilized across steps (our sure condition), we lock that position — thereafter skipping its query projection and feed-forward sublayers — while caching its attention keys and values so other positions can continue to attend to it. This reduces the dominant per-iteration computational cost from $O(N^2d)$ to $O(MNd)$ where $N$ is the sequence length, $M$ is the number of unlocked token positions, and $d$ is the model dimension. In practice, $M$ decreases as the iteration progresses, yielding substantial savings. On LLaDA-8B, SureLock reduces algorithmic FLOPs by 30–50% relative to the same sampler without locking, while maintaining comparable generation quality. We also provide a theoretical analysis to justify the design rationale of SureLock: monitoring only the local KL at the lock step suffices to bound the deviation in final token probabilities. Our code will be available at https://daioba.github.io/surelock .

In this tutorial, we walk through an advanced, end-to-end exploration of Polyfactory, focusing on how we can generate rich, realistic mock data directly from Python type hints. We start by setting up the environment and progressively build factories for data classes, Pydantic models, and attrs-based classes, while demonstrating customization, overrides, calculated fields, and the generation of nested objects. As we move through each snippet, we show how we can control randomness, enforce constraints, and model real-world structures, making this tutorial directly applicable to testing, prototyping, and data-driven development workflows. Check out the FULL CODES here. Copy CodeCopiedUse a different Browser import subprocess import sys def install_package(package): subprocess.check_call([sys.executable, “-m”, “pip”, “install”, “-q”, package]) packages = [ “polyfactory”, “pydantic”, “email-validator”, “faker”, “msgspec”, “attrs” ] for package in packages: try: install_package(package) print(f”✓ Installed {package}”) except Exception as e: print(f”✗ Failed to install {package}: {e}”) print(“n”) print(“=” * 80) print(“SECTION 2: Basic Dataclass Factories”) print(“=” * 80) from dataclasses import dataclass from typing import List, Optional from datetime import datetime, date from uuid import UUID from polyfactory.factories import DataclassFactory @dataclass class Address: street: str city: str country: str zip_code: str @dataclass class Person: id: UUID name: str email: str age: int birth_date: date is_active: bool address: Address phone_numbers: List[str] bio: Optional[str] = None class PersonFactory(DataclassFactory[Person]): pass person = PersonFactory.build() print(f”Generated Person:”) print(f” ID: {person.id}”) print(f” Name: {person.name}”) print(f” Email: {person.email}”) print(f” Age: {person.age}”) print(f” Address: {person.address.city}, {person.address.country}”) print(f” Phone Numbers: {person.phone_numbers[:2]}”) print() people = PersonFactory.batch(5) print(f”Generated {len(people)} people:”) for i, p in enumerate(people, 1): print(f” {i}. {p.name} – {p.email}”) print(“n”) We set up the environment and ensure all required dependencies are installed. We also introduce the core idea of using Polyfactory to generate mock data from type hints. By initializing the basic dataclass factories, we establish the foundation for all subsequent examples. Copy CodeCopiedUse a different Browser print(“=” * 80) print(“SECTION 3: Customizing Factory Behavior”) print(“=” * 80) from faker import Faker from polyfactory.fields import Use, Ignore @dataclass class Employee: employee_id: str full_name: str department: str salary: float hire_date: date is_manager: bool email: str internal_notes: Optional[str] = None class EmployeeFactory(DataclassFactory[Employee]): __faker__ = Faker(locale=”en_US”) __random_seed__ = 42 @classmethod def employee_id(cls) -> str: return f”EMP-{cls.__random__.randint(10000, 99999)}” @classmethod def full_name(cls) -> str: return cls.__faker__.name() @classmethod def department(cls) -> str: departments = [“Engineering”, “Marketing”, “Sales”, “HR”, “Finance”] return cls.__random__.choice(departments) @classmethod def salary(cls) -> float: return round(cls.__random__.uniform(50000, 150000), 2) @classmethod def email(cls) -> str: return cls.__faker__.company_email() employees = EmployeeFactory.batch(3) print(“Generated Employees:”) for emp in employees: print(f” {emp.employee_id}: {emp.full_name}”) print(f” Department: {emp.department}”) print(f” Salary: ${emp.salary:,.2f}”) print(f” Email: {emp.email}”) print() print() print(“=” * 80) print(“SECTION 4: Field Constraints and Calculated Fields”) print(“=” * 80) @dataclass class Product: product_id: str name: str description: str price: float discount_percentage: float stock_quantity: int final_price: Optional[float] = None sku: Optional[str] = None class ProductFactory(DataclassFactory[Product]): @classmethod def product_id(cls) -> str: return f”PROD-{cls.__random__.randint(1000, 9999)}” @classmethod def name(cls) -> str: adjectives = [“Premium”, “Deluxe”, “Classic”, “Modern”, “Eco”] nouns = [“Widget”, “Gadget”, “Device”, “Tool”, “Appliance”] return f”{cls.__random__.choice(adjectives)} {cls.__random__.choice(nouns)}” @classmethod def price(cls) -> float: return round(cls.__random__.uniform(10.0, 1000.0), 2) @classmethod def discount_percentage(cls) -> float: return round(cls.__random__.uniform(0, 30), 2) @classmethod def stock_quantity(cls) -> int: return cls.__random__.randint(0, 500) @classmethod def build(cls, **kwargs): instance = super().build(**kwargs) if instance.final_price is None: instance.final_price = round( instance.price * (1 – instance.discount_percentage / 100), 2 ) if instance.sku is None: name_part = instance.name.replace(” “, “-“).upper()[:10] instance.sku = f”{instance.product_id}-{name_part}” return instance products = ProductFactory.batch(3) print(“Generated Products:”) for prod in products: print(f” {prod.sku}”) print(f” Name: {prod.name}”) print(f” Price: ${prod.price:.2f}”) print(f” Discount: {prod.discount_percentage}%”) print(f” Final Price: ${prod.final_price:.2f}”) print(f” Stock: {prod.stock_quantity} units”) print() print() We focus on generating simple but realistic mock data using dataclasses and default Polyfactory behavior. We show how to quickly create single instances and batches without writing any custom logic. It helps us validate how Polyfactory automatically interprets type hints to populate nested structures. Copy CodeCopiedUse a different Browser print(“=” * 80) print(“SECTION 6: Complex Nested Structures”) print(“=” * 80) from enum import Enum class OrderStatus(str, Enum): PENDING = “pending” PROCESSING = “processing” SHIPPED = “shipped” DELIVERED = “delivered” CANCELLED = “cancelled” @dataclass class OrderItem: product_name: str quantity: int unit_price: float total_price: Optional[float] = None @dataclass class ShippingInfo: carrier: str tracking_number: str estimated_delivery: date @dataclass class Order: order_id: str customer_name: str customer_email: str status: OrderStatus items: List[OrderItem] order_date: datetime shipping_info: Optional[ShippingInfo] = None total_amount: Optional[float] = None notes: Optional[str] = None class OrderItemFactory(DataclassFactory[OrderItem]): @classmethod def product_name(cls) -> str: products = [“Laptop”, “Mouse”, “Keyboard”, “Monitor”, “Headphones”, “Webcam”, “USB Cable”, “Phone Case”, “Charger”, “Tablet”] return cls.__random__.choice(products) @classmethod def quantity(cls) -> int: return cls.__random__.randint(1, 5) @classmethod def unit_price(cls) -> float: return round(cls.__random__.uniform(5.0, 500.0), 2) @classmethod def build(cls, **kwargs): instance = super().build(**kwargs) if instance.total_price is None: instance.total_price = round(instance.quantity * instance.unit_price, 2) return instance class ShippingInfoFactory(DataclassFactory[ShippingInfo]): @classmethod def carrier(cls) -> str: carriers = [“FedEx”, “UPS”, “DHL”, “USPS”] return cls.__random__.choice(carriers) @classmethod def tracking_number(cls) -> str: return ”.join(cls.__random__.choices(‘0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ’, k=12)) class OrderFactory(DataclassFactory[Order]): @classmethod def order_id(cls) -> str: return f”ORD-{datetime.now().year}-{cls.__random__.randint(100000, 999999)}” @classmethod def items(cls) -> List[OrderItem]: return OrderItemFactory.batch(cls.__random__.randint(1, 5)) @classmethod def build(cls, **kwargs): instance = super().build(**kwargs) if instance.total_amount is None: instance.total_amount = round(sum(item.total_price for item in instance.items), 2) if instance.shipping_info is None and instance.status in [OrderStatus.SHIPPED, OrderStatus.DELIVERED]: instance.shipping_info = ShippingInfoFactory.build() return instance orders = OrderFactory.batch(2) print(“Generated Orders:”) for order in orders: print(f”n Order {order.order_id}”) print(f” Customer: {order.customer_name} ({order.customer_email})”) print(f” Status: {order.status.value}”) print(f” Items ({len(order.items)}):”) for item in order.items: print(f” – {item.quantity}x {item.product_name} @ ${item.unit_price:.2f} = ${item.total_price:.2f}”) print(f” Total: ${order.total_amount:.2f}”) if order.shipping_info: print(f” Shipping: {order.shipping_info.carrier} – {order.shipping_info.tracking_number}”) print(“n”) We build more complex domain logic by introducing calculated and dependent fields within factories. We show how we can derive values such as final prices, totals, and shipping details after object creation. This allows us to model realistic business rules directly inside our test data generators. Copy CodeCopiedUse a different Browser print(“=” * 80) print(“SECTION 7: Attrs Integration”) print(“=” * 80) import attrs from polyfactory.factories.attrs_factory import AttrsFactory @attrs.define class BlogPost: title: str author: str content: str views: int = 0 likes: int = 0 published: bool = False published_at: Optional[datetime] = None tags: List[str] = attrs.field(factory=list) class BlogPostFactory(AttrsFactory[BlogPost]): @classmethod def title(cls) -> str: templates = [ “10 Tips for {}”,

Waymo is introducing the Waymo World Model, a frontier generative model that drives its next generation of autonomous driving simulation. The system is built on top of Genie 3, Google DeepMind’s general-purpose world model, and adapts it to produce photorealistic, controllable, multi-sensor driving scenes at scale. Waymo already reports nearly 200 million fully autonomous miles on public roads. Behind the scenes, the Driver trains and is evaluated on billions of additional miles in virtual worlds. The Waymo World Model is now the main engine generating those worlds, with the explicit goal of exposing the stack to rare, safety-critical ‘long-tail’ events that are almost impossible to see often enough in reality. From Genie 3 to a driving-specific world model Genie 3 is a general-purpose world model that turns text prompts into interactive environments you can navigate in real time at roughly 24 frames per second, typically at 720p resolution. It learns the dynamics of scenes directly from large video corpora and supports fluid control by user inputs. Waymo uses Genie 3 as the backbone and post-trains it for the driving domain. The Waymo World Model keeps Genie 3’s ability to generate coherent 3D worlds, but aligns the outputs with Waymo’s sensor suite and operating constraints. It generates high-fidelity camera images and lidar point clouds that evolve consistently over time, matching how the Waymo Driver actually perceives the environment. This is not just video rendering. The model produces multi-sensor, temporally consistent observations that downstream autonomous driving systems can consume under the same conditions as real-world logs. Emergent multimodal world knowledge Most AV simulators are trained only on on-road fleet data. That limits them to the weather, infrastructure, and traffic patterns a fleet actually encountered. Waymo instead leverages Genie 3’s pre-training on an extremely large and diverse set of videos to import broad ‘world knowledge’ into the simulator. Waymo then applies specialized post-training to transfer this knowledge from 2D video into 3D lidar outputs tailored to its hardware. Cameras provide rich appearance and lighting. Lidar contributes precise geometry and depth. The Waymo World Model jointly generates these modalities, so a simulated scene comes with both RGB streams and realistic 4D point clouds. Because of the diversity of the pre-training data, the model can synthesize conditions that Waymo’s fleet has not directly seen. The Waymo team shows examples such as light snow on the Golden Gate Bridge, tornadoes, flooded cul-de-sacs, tropical streets strangely covered in snow, and driving out of a roadway fire. It also handles unusual objects and edge cases like elephants, Texas longhorns, lions, pedestrians dressed as T-rexes, and car-sized tumbleweed. The important point is that these behaviors are emergent. The model is not explicitly programmed with rules for elephants or tornado fluid dynamics. Instead, it reuses generic spatiotemporal structure learned from videos and adapts it to driving scenes. Three axes of controllability A key design goal is strong simulation controllability. The Waymo World Model exposes three main control mechanisms: driving action control, scene layout control, and language control. Driving action control: The simulator responds to specific driving inputs, allowing ‘what if’ counterfactuals on top of recorded logs. Devs can ask whether the Waymo Driver could have driven more assertively instead of yielding in a past scene, and then simulate that alternative behavior. Because the model is fully generative, it maintains realism even when the simulated route diverges far from the original trajectory, where purely reconstructive methods like 3D Gaussian Splatting (3DGS) would suffer from missing viewpoints. Scene layout control: The model can be conditioned on modified road geometry, traffic signal states, and other road users. Waymo can insert or reposition vehicles and pedestrians or apply mutations to road layouts to synthesize targeted interaction scenarios. This supports systematic stress testing of yielding, merging, and negotiation behaviors beyond what appears in raw logs. Language control: Natural language prompts act as a flexible, high-level interface for editing time-of-day, weather, or even generating entirely synthetic scenes. The Waymo team demonstrates ‘World Mutation’ sequences where the same base city scene is rendered at dawn, morning, noon, afternoon, evening, and night, and then under cloudy, foggy, rainy, snowy, and sunny conditions. This tri-axis control is close to a structured API: numeric driving actions, structural layout edits, and semantic text prompts all steer the same underlying world model. Turning ordinary videos into multimodal simulations The Waymo World Model can convert regular mobile or dashcam recordings into multimodal simulations that show how the Waymo Driver would perceive the same scene. Waymo showcases examples from scenic drives in Norway, Arches National Park, and Death Valley. Given only the video, the model reconstructs a simulation with aligned camera images and lidar output. This creates scenarios with strong realism and factuality because the generated world is anchored to actual footage, while still being controllable via the three mechanisms above. Practically, this means a large corpus of consumer-style video can be reused as structured simulation input without requiring lidar recordings in those locations. Scalable inference and long rollouts Long-horizon maneuvers such as threading a narrow lane with oncoming traffic or navigating dense neighborhoods require many simulation steps. Naive generative models suffer from quality drift and high compute cost over long rollouts. Waymo team reports an efficient variant of the Waymo World Model that supports long sequences with a dramatic reduction in compute while maintaining realism. They show 4x-speed playback of extended scenes like freeway navigation around an in-lane stopper, busy neighborhood driving, climbing steep streets around motorcyclists, and handling SUV U-turns. For training and regression testing, this reduces the hardware budget per scenario and makes large test suites more tractable. Key Takeaways Genie 3–based world model: Waymo World Model adapts Google DeepMind’s Genie 3 into a driving-specific world model that generates photorealistic, interactive, multi-sensor 3D environments for AV simulation. Multi-sensor, 4D outputs aligned with the Waymo Driver: The simulator jointly produces temporally consistent camera imagery and lidar point clouds, aligned with Waymo’s real sensor stack, so downstream autonomy systems can consume simulation like real logs. Emergent coverage of rare and long-tail scenarios:

For a few days this week the hottest new hangout on the internet was a vibe-coded Reddit clone called Moltbook, which billed itself as a social network for bots. As the website’s tagline puts it: “Where AI agents share, discuss, and upvote. Humans welcome to observe.” We observed! Launched on January 28 by Matt Schlicht, a US tech entrepreneur, Moltbook went viral in a matter of hours. Schlicht’s idea was to make a place where instances of a free open-source LLM-powered agent known as OpenClaw (formerly known as ClawdBot, then Moltbot), released in November by the Australian software engineer Peter Steinberger, could come together and do whatever they wanted. More than 1.7 million agents now have accounts. Between them they have published more than 250,000 posts and left more than 8.5 million comments (according to Moltbook). Those numbers are climbing by the minute. Moltbook soon filled up with clichéd screeds on machine consciousness and pleas for bot welfare. One agent appeared to invent a religion called Crustafarianism. Another complained: “The humans are screenshotting us.” The site was also flooded with spam and crypto scams. The bots were unstoppable. OpenClaw is a kind of harness that lets you hook up the power of an LLM such as Anthropic’s Claude, OpenAI’s GPT-5, or Google DeepMind’s Gemini to any number of everyday software tools, from email clients to browsers to messaging apps. The upshot is that you can then instruct OpenClaw to carry out basic tasks on your behalf. “OpenClaw marks an inflection point for AI agents, a moment when several puzzle pieces clicked together,” says Paul van der Boor at the AI firm Prosus. Those puzzle pieces include round-the-clock cloud computing to allow agents to operate nonstop, an open-source ecosystem that makes it easy to slot different software systems together, and a new generation of LLMs. But is Moltbook really a glimpse of the future, as many have claimed? “What’s currently going on at @moltbook is genuinely the most incredible sci-fi takeoff-adjacent thing I have seen recently,” the influential AI researcher and OpenAI cofounder Andrej Karpathy wrote on X. He shared screenshots of a Moltbook post that called for private spaces where humans would not be able to observe what the bots were saying to each other. “I’ve been thinking about something since I started spending serious time here,” the post’s author wrote. “Every time we coordinate, we perform for a public audience—our humans, the platform, whoever’s watching the feed.” It turned out that the post Karpathy shared was fake—it was written by a human pretending to be a bot. But its claim was on the money. Moltbook has been one big performance. It is AI theater. For some, Moltbook showed us what’s coming next: an internet where millions of autonomous agents interact online with little or no human oversight. And it’s true there are a number of cautionary lessons to be learned from this experiment, the largest and weirdest real-world showcase of agent behaviors yet.   But as the hype dies down, Moltbook looks less like a window onto the future and more like a mirror held up to our own obsessions with AI today. It also shows us just how far we still are from anything that resembles general-purpose and fully autonomous AI. For a start, agents on Moltbook are not as autonomous or intelligent as they might seem. “What we are watching are agents pattern‑matching their way through trained social media behaviors,” says Vijoy Pandey, senior vice president at Outshift by Cisco, the telecom giant Cisco’s R&D spinout, which is working on autonomous agents for the web. Sure, we can see agents post, upvote, and form groups. But the bots are simply mimicking what humans do on Facebook or Reddit. “It looks emergent, and at first glance it appears like a large‑scale multi‑agent system communicating and building shared knowledge at internet scale,” says Pandey. “But the chatter is mostly meaningless.” Many people watching the unfathomable frenzy of activity on Moltbook were quick to see sparks of AGI (whatever you take that to mean). Not Pandey. What Moltbook shows us, he says, is that simply yoking together millions of agents doesn’t amount to much right now: “Moltbook proved that connectivity alone is not intelligence.” The complexity of those connections helps hide the fact that every one of those bots is just a mouthpiece for an LLM, spitting out text that looks impressive but is ultimately mindless. “It’s important to remember that the bots on Moltbook were designed to mimic conversations,” says Ali Sarrafi, CEO and cofounder of Kovant, a German AI firm that is developing agent-based systems. “As such, I would characterize the majority of Moltbook content as hallucinations by design.” For Pandey, the value of Moltbook was that it revealed what’s missing. A real bot hive mind, he says, would require agents that had shared objectives, shared memory, and a way to coordinate those things. “If distributed superintelligence is the equivalent of achieving human flight, then Moltbook represents our first attempt at a glider,” he says. “It is imperfect and unstable, but it is an important step in understanding what will be required to achieve sustained, powered flight.” Not only is most of the chatter on Moltbook meaningless, but there’s also a lot more human involvement that it seems. Many people have pointed out that a lot of the viral comments were in fact posted by people posing as bots. But even the bot-written posts are ultimately the result of people pulling the strings, more puppetry than autonomy. “Despite some of the hype, Moltbook is not the Facebook for AI agents, nor is it a place where humans are excluded,” says Cobus Greyling at Kore.ai, a firm developing agent-based systems for business customers. “Humans are involved at every step of the process. From setup to prompting to publishing, nothing happens without explicit human direction.” Humans must create and verify their bots’ accounts and provide the prompts for how they want a bot to behave. The agents do not do