「UNBOUND x SHIFT UP」 Two powerhouses unite to develop a brand-new title. Stay tuned for the incredible synergy between Shinji Mikami's directorship and SHIFT UP! #MikamiShinji# #UNBOUND# #SHIFTUP#

It is a profound honor to work alongside the top-tier creative team at Unbound. I am deeply moved to support the visionaries behind the legendary games I’ve cherished my whole life, especially a project directed by Shinji Mikami himself. We are currently preparing a diverse lineup—from massive blockbusters that will shock the world to experimental projects fueled by brilliant, sparkling ideas. Please look forward to what’s coming! #MikamiShinji# #UNBOUND# #SHIFTUP#

EA has confirmed that Need for Speed is on hold for the foreseeable future. Criterion Games, the studio behind Need for Speed Unbound, is now fully focused on the Battlefield franchise. Rebecka Coutaz, Vice President and General Manager of Battlefield Studios Europe, made it clear that the team’s priority is Battlefield, saying they are “solely focused on Battlefield” rather than looking back at past projects. As a result, there is currently no new Need for Speed game in development at Criterion. EA has not cancelled the franchise but has not announced any plans for its return.

Golf is the worst sport for AI. And that's exactly why it's the best sport for AI. Worst because: • Unbounded state space (every lie is unique) • Delayed rewards (one bad decision shows up 3 holes later) • Partial observability (you can't see wind at the green) • Non-stationary environment (conditions change continuously) Best because: • Solve golf, solve generalization • Every challenge maps to real-world AI problems • The benchmark is clear: beat the best human We're not building golf bots. We're building general intelligence that happens to play golf.

For agent memory, the real question isn't "how to store" — it's "what to remember" 🧠 A fresh take that learns what to memorize via reinforcement learning. Title: Task-Focused Memorization for Multimodal Agents URL: 🧠 Overview This work proposes TaskMem, which treats long-term memory for multimodal agents as a learnable policy optimized with reinforcement learning, focused on deciding what to memorize. From an unbounded stream of observations, it selectively retains only the content relevant to the agent's role and task. ❓ Challenges Solved A multimodal agent operating in the real world continuously receives an unbounded stream of observations. ・Most prior work focused on how to store memories (designing memory modules) ・But the essential problem is what to memorize — without a principled way to select role-relevant content from an endless stream, memory simply fails This work starts from that shift in perspective. 💡 Methodology & Proposed Approach TaskMem treats memorization as a learnable policy, optimized in two phases. ・Phase 1: learn high-quality memorization under fidelity requirements ・Phase 2: post-deployment fine-tuning that uses task rewards to align memorization with the environment's demands ・It builds on the MLLM Qwen3-VL-30B-A3B and optimizes the policy lightly via adapter tuning ・Reward models derived from real tasks steer the policy toward selecting relevant content 🌍 Use Cases / Experimental Results On reformulated streaming benchmarks, it delivered clear accuracy gains. ・VideoMME: 67.9% VQA accuracy (+6.3%) ・EgoLife: 45.4% VQA accuracy (+7.0%) ・EgoTempo: 27.6% VQA accuracy (+5.3%) ・Strong precision across all benchmarks (80.5-85.6%) It charts a practical path for long-running, always-on agents to selectively remember the right things while keeping context bloat in check. #AIAgents# #Memory#

Shift Up says the next Stellar Blade game is going well. The company will publish it itself worldwide. The first Stellar Blade was published by Sony. This time, Shift Up will handle everything. That means it will control marketing, global sales, and fan engagement. The company wants full control to do a better job and sell more copies, expect no censorship from Sony Shift Up expects the sequel to perform much better than the first game thanks to its large fan base The original game is still selling well on PC and consoles. In the first three months of 2026, it generated 12.9 billion Korean won, or 27.3% of the company’s total revenue for that period. The company plans to share more details about the Stellar Blade sequel and another game called Project Spirits later in 2026. Shift Up is also building its publishing team. It hired about 60 new developers through a partnership with UNBOUND.

# Neo4j Features and Practical Usage 🕸️ "If this server goes down, what chain of things stops with it?" Cypher's variable-length patterns answer that by traversing relationships across any number of hops. 🏷️ Title: Variable-length patterns (`*1..5`) / SHORTEST / Quantified Path Patterns 🔗 URL: 📘 Overview Variable-length patterns let you express paths whose hop count is not fixed, all in a single declarative pattern. Questions like "trace suppliers up to 5 levels deep" or "find one shortest path" become natural one-liners. Shortest-path finding and Quantified Path Patterns (QPP) belong to the same family. ⚙️ How It Works ・Quantified relationships (legacy syntax): `[:REL*1..5]` matches 1 to 5 repetitions. `*2` is exactly 2, `*3..` is 3 or more, `*..10` is up to 10, `+` is one or more, and `*` is zero or more. ・Quantified Path Patterns (QPP): wrap a whole pattern in parentheses and quantify it, e.g. `((a)-[r:NEXT]->(b)){1,3}`. It is GQL-conformant and supports inline `WHERE` predicates inside the repetition. ・Group variables: variables declared inside a QPP become lists when referenced outside (e.g. `r` becomes an array of relationships). Combine with `reduce()` to compute, say, total distance along a path. ・Shortest paths: `SHORTEST k` returns k shortest paths, `ALL SHORTEST` returns every tied-shortest path, and `ANY` returns any one. The legacy `shortestPath()` / `allShortestPaths()` functions still work but the keyword syntax is faster. 🛠️ Practical Usage Trace a part's suppliers up to 5 levels deep: ```cypher MATCH (p:Part {sku: $sku})-[:SUPPLIED_BY*1..5]->(supplier:Company) RETURN DISTINCT ``` Find the single shortest introduction chain between two people: ```cypher MATCH path = SHORTEST 1 (a:Person {id:$a})-[:KNOWS]-+(b:Person {id:$b}) RETURN [n IN nodes(path) | AS intro_chain ``` 💡 Use Cases ・Impact analysis: how far a server failure cascades. ・Money-laundering detection: tracing fund-movement paths between accounts. ・Org charts and social graphs: boss-of-boss lookups, friend-of-friend shortest reach. ⚠️ Caveats ・Unbounded `*` or `[:REL*]` causes path explosion and can stall on millions of matches. Always set an upper bound such as `*1..5`. ・Pruning early with labels, relationship types, and inline predicates makes queries dramatically faster. QPP is built for this kind of pruning. ・By default Cypher does not allow re-traversing the same relationship; be aware that this behavior can be modified. #Neo4j# #Cypher#

As agents become the biggest users of software, then all software has to be available in a headless fashion. Agents won’t be using your UI, they’ll be talking to your APIs. So the question becomes what is the business model of software and this headless approach in the future? Here are a few thoughts on how everything plays out based on what we’re seeing and doing at Box, but also conversation with other platforms. 1) Seats don’t go away for *people*. Seats are still a convenient and efficient way to have a customer use technology predictably for a set of users within a baseline set of usage. The key, though, is that when the customer pays for a seat, it has to come with a set of usage of APIs on behalf of that user that the agent can use on their behalf. The user will need to be able to interact with their data and the underlying tool via any agent they work with, and an embedded amount of usage will come with the seat. I would imagine most software -Box included- will enable seats to work with their data at a relatively high volume via systems like ChatGPT, Codex, Claude, Gemini, Cursor, Copilot, Perplexity, Factory, Cogniton, et al. quite seamlessly. If you don’t do this, you’re DOA. 2) Agents may have “seats” if they are doing stateful work in the system, but they will be priced very differently than people. Seats (or the equivalent) can make sense when you have an agent that has its own workspace, stores its own data, needs a different set of permissions compared to the user, and so on. If a company wants this agent to be around for long period of time, that may very well look like another “user” in the system. Openclaw-style agents highlight what this future could look like. The only issue on pricing here is that one customer could decide to do all their work in 1 agent, and another might split it into 1,000 agents. So pricing like a human seat is nearly impossible and impractical; each company will have a different approach for this as it gets tricky perfectly trying to capture all the value within an agent seat. 3) The dominant pricing for headless use that goes above the seat allotment, or when an agent is firmly acting on their own, will be a consumption model. Many enterprises software platforms have previously operated like this with PaaS options, and agents will look like another machine user of their system. In some cases the APIs might get priced just as they did previously, but in other cases there may need to be new types of APIs that represent the work an agent would do in one go -more akin to an outcome- instead of a series of API calls. This is especially germane when the headless software also has an agentic use-case embedded within in, such as orchestrating the process within their own system via AI. Overall the growth of this usage pattern is effectively unbounded as the use-cases for agents operating on data in these systems will dramatically exceed what people do with their data and tools today. Every platform that goes headless (which will be anyone that wants to take advantage of agents) will need to adopt a model like this. Some may fight it initially but it’s an inevitably as there will always be more and more agents outside your platform than people. Overall, there’s a lot of really interesting changes left to come in software due to headless use of these systems. Early days.

Love and Deepspace | Fallen Crown Through immortality and oblivion, my path leads to your light. 👑Trailer of Xavier's New Event [Fallen Crown] Is Coming! ┈┈▾Giveaway▾┈┈ 💫Follow @Love_Deepspace and Repost for a chance to win $70 (5 winners) before Aug. 7. 👑5-Star Memory Pair from [Fallen Crown] Event From 05:00 on Jul. 25 to 04:59 on Aug. 8 (server time), the limited 5-Star Solar-Slot Memory Pair [Xavier: Nightvow Mooncurse] and [Xavier: Nightvow Requiem] will be available in the Limited Wish Pool. During the event, a maximum of 150 pulls guarantees two event-limited 5-Star Memories and unlocks the Companion [Xavier: King of Darknight]. *See the specific wish rules in the following event preview. 👑Myth: Where Stars Scatter After 05:00 on Jul. 25 (server time), you can read the new Myth of Xavier: Where Stars Scatter in-game. "When their names are spoken once more... Only then will their unfulfilled ambitions, unexpressed love, and unsolved mysteries transform into stories that are passed down for generations. And to endure the test of time can be seen as a divine miracle. 'Eternity' has always been about stories that can immortalize humanity's joys and sorrows." The prophesied Chosen One, summoned by fate... The Mad King, held captive by the crown... The longing to return to Philos draws the threads of their destiny together. When the fated moment arrives amidst the ruins of time—will you still stand by his side? 👑Event: Cycle Unbound Participate in the event to claim 4-Star Solar-Slot Memory Pair [Xavier: Gilded Mirage], [Xavier: Gilded Dreams], [Deepspace Wish: Limited*10], [Diamond*500], and other rewards for FREE! #LoveandDeepspace# #XavierKingofDarknight# #Xavier#