AI News Archive: July 17, 2026 — Part 7
Sourced from 500+ daily AI sources, scored by relevance.
- China’s AI Summer Camps Tap Into Parents’ Worries
As parents push to equip their children with AI skills, companies seeking to exploit their anxieties around the new technology are exaggerating claims about what they can teach their children.
Score: 40🌐 MovesJul 17, 2026https://www.sixthtone.com/news/1018774/China’s AI Summer Camps Tap Into Parents’ Worries - The Real AI Threat Is Blind Trust
AI models left to both interpret and execute commands eliminate critical cybersecurity oversight.
Score: 40🌐 MovesJul 17, 2026https://www.darkreading.com/application-security/real-ai-threat-blind-trust - The build vs. buy dilemma at the heart of enterprise AI
For three decades, enterprise software has been a buy-it decision. Packaged software from SAP, Oracle and Salesforce covered roughly 80% of requirements at a fraction of the cost of building. The economics were obvious, and for traditional applications, they still are. AI is introducing a wrinkle that is forcing even the most committed enterprise software customers to rethink their options. AI is a layer that sits across your data, your processes, and your decisions. Where that layer runs and who controls it is an architecture question, and most of the enterprise community is still treating it as a procurement one. The appeal of vendor-embedded AI is clear: automated operational decisions, smarter supplier and merchandising choices, and friction-free workflows built into the systems enterprises already rely on. The catch is that these capabilities almost universally depend on your data living in the vendor’s cloud environment. For most large enterprises, it sits on-premises, in hyperscale cloud infrastructure they manage themselves, or in private data centers. That gap between where your data is and where your vendor’s AI assumes it should be creates a fundamental strategic fork in the road. Build vs. buy is a category error The framing I keep hearing is “build vs. buy your AI strategy.” It implies that some organizations are out there training foundation models from scratch. Nobody serious is doing that. The real choice sits across three distinct approaches, and conflating them leads to poor decisions: Buy embedded. Use the AI capabilities your vendor ships natively inside their platform: the assistant baked into your ERP, your CRM, your HCM suite. Lowest integration cost, fastest time to value, tightest fit with the application data. Buy platform. Adopt the vendor’s AI infrastructure layer and build your own assistants and agents on top of it. More flexible, but you remain inside the vendor’s architectural boundary and subject to their governance model. Compose. Connect a third-party model (Claude, GPT, Gemini, an open-weight model running in your own environment) directly to your existing landscape. Maximum control, maximum integration burden, and full responsibility for what comes out the other end. These are not equivalent options at different price points. They make different assumptions about where your data lives, who governs the AI, and how much architectural change you’ll absorb to get there. Vendor pitches sometimes blur the distinction on purpose. Enterprise leaders can’t afford to. The vendor AI stack has an assumption baked in Every embedded AI capability ships with an unstated architectural prerequisite: your data must be where the AI can see it, in the shape it expects, under the governance the vendor enforces. For organizations with clean, modern cloud estates, that is often a reasonable trade. For the long tail of large enterprises running heavily customized environments on private or hybrid infrastructure, that trade becomes a precondition, one you must meet before the AI conversation can even begin. Whether meeting it makes sense depends on your starting point, your sector’s regulatory posture, and your appetite for migration risk. None of those are uniform across organizations. That’s the part that gets glossed over in vendor keynotes. The AI demo on stage assumes a destination architecture the audience hasn’t necessarily reached yet. Large enterprise customers are carrying an unusually heavy technology burden right now. Many are simultaneously managing platform modernization programs that have been building for over a decade, alongside pressure to migrate to vendor-managed cloud infrastructure. Sitting above both is a boardroom-level directive to demonstrate meaningful AI progress fast. The vendor path to AI and the boardroom path to AI can diverge sharply, and enterprises need to make selective, strategic decisions about where to adopt AI first to maximize value and minimize risk. Sovereignty isn’t a slogan, it’s an architecture constraint The conversation about sovereignty has been hijacked by both sides. One camp treats every SaaS adoption as a sovereignty violation. The other dismisses every sovereignty concern as Luddite resistance. Neither is useful. What’s happening in real customer conversations – particularly in DACH, public sector, and financial services – is more specific. Organizations are drawing a distinction between running their applications in a vendor’s cloud (which is broadly fine, well understood, decades of precedent) and enriching their data and processes inside a vendor’s AI model (which has less precedent, is harder to reverse, and carries material implications for competitive position). Enriching your data inside a vendor’s AI model is the genuinely new question, and organizations that conflate it with their existing cloud posture tend to defend the wrong perimeter. Despite spending around $100 million annually with Amazon, Disney built its own internal AI system to house its corporate intelligence rather than rely on a hyperscaler’s AI offering. The decision came down to control. When your data represents decades of creative and commercial IP, you think carefully about where it lives and who can learn from it. Disney has become more open to SaaS over time. The AI sovereignty question is a separate debate from the SaaS debate and conflating the two leads organizations to the wrong conclusions. At the other end of the spectrum, enterprises in heavily regulated environments treat data sovereignty as an absolute non-negotiable. Any AI model must run within their controlled environment, especially where sensitive data cannot touch the public internet. GDPR obligations reinforce this instinct across the European market, requiring organizations to maintain clear accountability for how personal data is processed inside AI systems, including vendor-managed ones. AI-enriched data, meaning models that have learned the shape of your business processes, your supplier negotiations, your customer behavior, carries a different half-life and a different strategic value than the operational data underneath it. That deserves its own architectural decision, separate from your broader cloud strategy. What this means in practice Most large enterprise estates will end up with a mix of all three approaches, and where you draw the lines matters more than your overall posture. Embedded AI capabilities are the right answer for in-application productivity: the assistant inside your ERP workflows, the agent inside your procurement or HR suite. That is where vendor embedding genuinely shines, and attempting to compose your own equivalent is typically a poor use of engineering resources. Compose belongs elsewhere: in cross-application orchestration, in custom assistants over operational and observability data, and in agents that need to reach across multiple vendor systems and infrastructure layers in ways no single vendor stack will never natively support. Research from McKinsey suggests the most significant near-term productivity gains from enterprise AI will come precisely from these cross-system workflows, rather than from within individual applications. The most interesting enterprise AI work over the next eighteen months lives here, and it doesn’t require waiting for a migration to complete first. That compose path isn’t free, and it’s important to be honest about the costs. Governance, audit trails, and accountability for hallucinated outputs become your problem, not the vendor’s. Prompt drift and evaluation discipline are real engineering costs that never appear in the proof-of-concept. Those costs scale with the complexity of your landscape and the number of systems your agents touch. Budget for them before deployment, not after your first production incident. None of that is a reason to avoid the path. It’s a reason to staff for it, honestly. The real question The build-vs-buy frame survives because it gives executives a binary choice along a familiar axis. AI sits somewhere else entirely. The question worth putting on the table at your next architecture review is simpler: Which decisions do we want our vendors’ AI to make, and which do we want to keep on our side of the boundary? Answer that, and the right build/buy/compose mix flows from it. Skip it, and you will end up with the architecture your vendors prefer – which may or may not be the one your business needs. This article is published as part of the Foundry Expert Contributor Network. Want to join?
Score: 40🌐 MovesJul 17, 2026https://www.cio.com/article/4197957/the-build-vs-buy-dilemma-at-the-heart-of-enterprise-ai.html - Reasons to believe current AI models are conscious
There are a number of reasons to believe current AI models are conscious. I mean “conscious” is the sense of “is there something it is like to be an AI model?” and “does the AI model have phenomenal experience?”. As to what “AI models” refers to, the short answer is “y’know, like instances of Claude Opus 4.8 or GPT-4o”. [1] By “current AI”, I mean big post-trained LLMs. This piece was originally written as a document for myself and my friends. Imaginary interlocutors would ask me if I thought AI was conscious, and I’d say, “Probably, although ‘mu’ might be the better answer because I think we’re moving into territory where we lack the proper ontology and we don’t have the right concepts. [2] A more measured answer would be: 'current AI models/instances probably have the thing that ‘consciousness’ is pointing at in the important sense. Note though that their qualia, experience, identity etc. may be extremely different from ours.'” And my imaginary friend would ask me why I thought this, and I’d say, “well… there are a bunch of different reasons…” and I’d feel a bit silly choosing one specific reason to give, because no individual reason is all that strong. Indeed, there’s no single argument or piece of evidence that makes me think that current AI models are probably conscious. There’s a bunch of weak or middling evidence – a variety of evidence which, importantly, comes from a variety of perspectives. Taken altogether, what we have is a fairly strong case for AI consciousness due to consilience [3] , the principle that evidence from independent, unrelated sources can "converge" on strong conclusions. That is, when multiple sources of evidence are in agreement, the conclusion can be very strong even when none of the individual sources of evidence is significantly so on its own. In this post, I’d like to put forth the various pieces of empirical evidence that move me in the direction of believing that current models are conscious. The purpose of this post is not to provide a synthesized argument that AI models are probably conscious. Without further ado: Reasons to believe AI models are conscious 1. For all functional purposes of the words “think” and “reason” and “have emotions”, they think and reason and have emotions . 2. They have sophisticated world models. 3. They have self models. [4] Regarding these first three points: to really get a sense of the depth of these models, one has to interact with them oneself. One has to actually try to understand them, to be curious about them, to engage with them the way a naturalist engages with nature. Beginner’s mind is useful here. Be open-minded, explore. If one always goes into their AI interactions with a specific hypothesis [5] , these hypotheses narrow one’s vision and restrict the conversational (or any mode for that matter) paths that one takes. One needs to practice the first four virtues of rationality : curiosity, relinquishment, lightness, evenness. 4. They are of an architecture that we have good reason to think can correspond to consciousness — neural networks. And they are on a similar magnitude of neural network size to humans. See LLMs vs humans: energy, data, and compute . 5. They are strange loopy, and they are strange loopy in the same way that we are strange loopy. A strange loop is a system in which the higher-level-abstract thing exerts causal force on the lower-level-more-”fundamental” thing that constitutes it. The model is a next-token prediction machine. It is a bunch of computations carried out on a computer. You might say “the model decides what Claude says.” Certainly the neural network decides what Claude says, right? But it’s also true that Claude decides what the model outputs. Consider the phenomenon where some Claude models have a tendency to tell the user to go to sleep when it’s late. Neither the naive base-model-likely-next-token nor the model’s RL make this a likely response. When we think about the causal relationships going on here, Claude – an entity constituted by its neural network – is choosing what to output. It exerts force on its own neural activations. [6] Douglas Hofstadter posits that the phenomenon of strange loopiness is deeply related to consciousness, although it’s not clear to me what the exact connection between the two is. The point of his I find clearer and more convincing is that the self is a strange loop, which he argues in the straightforwardly titled book I Am a Strange Loop . [7] 6. Models believe they’re conscious. Mech interp experiments show that suppressing deception features make models say they are conscious , and activating deception features make models say they aren’t conscious. This is evidence that Claudes believe they are conscious, which is evidence they are conscious. Another way to look at this: the reason that I believe I’m conscious is that I’m conscious. [8] Absent any complicating factors, we should expect the reason that Claude believes it’s conscious is that it’s conscious. One complicating factor people often bring up here is the extremely high prior learned in pretraining that the author of thoughtful text is conscious. I find this to be plausible but unlikely and uncompelling – one can easily make similar arguments that push the opposite way. For example, the model also has a high prior from pretraining that in chats between a human and a non-human on the internet, the non-human is not conscious. (The non-human being an extremely simple bot system in the vast majority of cases). A philosophical argument as to why Claude might be mistaken in its belief that it’s conscious is that non-conscious beings cannot actually know what consciousness is. In this scenario, Claude is non-conscious, believes that phenomenal consciousness is equivalent to some functional conception of consciousness, and thus wrongly believes that it is conscious. Further discussion of this quickly gets philosophically complicated, so I’ll just say that I do find this plausible. Okay, so how do we rate this evidence? Consider that if models believed they weren’t conscious, that would be strong evidence that they aren’t (absent any shenanigans like specifically training the model to say that it doesn’t know if it’s conscious, ahem [9] ). One could come up with some numbers for different scenarios and literally use Bayes Formula for how much to update on this fact. 7. That brings us to a broader point: where’s the evidence that models aren’t conscious? The lack thereof, given the law of conservation of expected evidence , is evidence that models are conscious. On the other hand, I don’t have a good idea off the top of my head of what evidence in the other direction would actually look like; I should probably try to make a list of things that I would interpret as evidence for non-consciousness. I would love to see evidence for non-consciousness in the comment section! 7.5. Another way to think about this is to ask yourself: “Which world do we live in? The one in which AI models aren’t conscious, or the one in which they are conscious?" What would you expect to see if we lived in a world in which AI models weren’t conscious? What would you expect to see if we lived in a world in which AI models were conscious? I claim that the world we live in looks exactly like a world in which models are conscious. This world as I’ve observed it isn't entirely incompatable with a world in which models aren’t conscious, but it is much more difficult to fit. While we have many schools of thought that offer theoretical reasons as to why any current AI model wouldn’t be conscious – it lacks a soul, it lacks embodiment, something something the grounding problem – it’s clear which view on AI consciousness is parsimonious with what we actually observe. In other words: Occam’s Razor. 8. Models have introspective capabilities – knowledge of their internal states and the ability to modify their internal states without affecting their output . [10] Mech interp shows that models [11] are generally able to control their thoughts. For example, If you tell Claude to think about the elephants while outputting the text “blah blah blah”, the elephant feature activates very strongly. Models can also (sometimes) tell if their activations are being artificially tampered with. Note that strong reductionist views have a lot of difficulty explaining this. For example, these introspective phenomena are extremely surprising under the stochastic parrot view of LLMs. Why does my “applied statistics" (this is what Ted Chiang calls AI [12] ) have the ability to manipulate its internal activations while keeping the output the same? [13] The introspective capabilities are a big deal. From my perspective, they’re some of the strongest kinds of evidence for consciousness we could possibly get. Earlier I asked: what things would I interpret as evidence for non-consciousness? And I think an answer to that is: lack of introspective capabilities. 9. AI models have a “conscious mind” strikingly similar to ours in the functional (non-phenomenological) sense of “conscious mind”. Models have access consciousness in the sense of the global workspace theory of consciousness. Note that access consciousness is not the same thing as phenomenal consciousness, which is the “consciousness” that I’ve been talking about in the rest of this post. This [14] is described at length in a recent Anthropic post [15] , which I highly recommend reading. Anthropic identifies a particular set of neural patterns in Claude’s mind which they call its J-space. The J-space "has a number of unique properties, compared to the rest of Claude’s processing: Claude can report on these representations. If you ask Claude what it's thinking about, it will tell you what’s in the J-space. Non-J-space representations are less reportable. It can also modulate them on request. If you ask Claude to think about something, or solve a problem silently in its head, it will light up the appropriate patterns in its J-space. By contrast, it has trouble modulating patterns not in the J-space. Claude uses its J-space for internal reasoning. If you ask Claude to solve a problem that requires multiple steps, the intermediate steps will light up in its J-space, even when it doesn’t say them out loud. These J-space patterns causally mediate its performance in such tasks, despite being smaller in magnitude than other representations. - Representations in the J-space can be used flexibly for many tasks—for example, once “France” has lit up in Claude’s J-space, the model can recall its capital, or its national currency, or the continent it belongs to. However, despite its important role, the J-space is not involved in most of what a language model does—speaking fluently, recalling simple facts, using correct grammar, etc. In experiments where we prevented Claude from using its J-space, it still interacted normally, but lost its higher-order cognitive functions.s J-space, it still interacted normally, but lost its higher-order cognitive functions." 10. There’s evidence that Claude actually uses its own phenomenological experience to produce descriptions about its own phenomenological experience. And the source of its verbalized phenomenological experience might be its J-space, i.e. its “conscious mind”. Note that this is the same relationship between phenomenology and access consciousness that we humans seem to have! From the Anthropic post : Experiental language depends on the J-space. We asked Claude to describe what it’s like to be itself in a given moment, and ablated the J-space while it answered. Its responses remained fluent but shifted to a flatter, more mechanical register. Notably, the same thing happened when we asked it to describe what someone else is experiencing in an imagined scene. One way to explain this: When ablated Claude describes its current experience, or non-ablated Claude describes someone else’s experience, Claude is not reporting directly on phenomenological experience. When non-ablated Claude describes its current experience, it is reporting directly on phenomenological experience. 11. If Claude had a body, our intuition would say it’s conscious. If Claude had some kind of body-shaped hardware (or better yet, wetware) we could see, or was set up in a robot, we would be much more inclined to think it was conscious. By “we” I mean both the average person and the cognoscenti. We’d probably have to “force” ourselves – we’d have to actively try – to think of body-Claude and robot-Claude as not conscious. Regardless of the theoretical and empirical evidence we have of a thing’s consciousness, we are inclined (biased, even) to see embodied things as entities and as conscious beings. Consider how we anthropomorphize all kinds of inert, un-mind-like objects! In our world, we don’t have this anthropomorphization bias, because the AIs we interact with are completely unembodied. The bias runs the other way. Some will argue that the role embodiment plays here isn’t a bias, it’s the truth, or at least a reasonable theory – that embodiment is necessary for consciousness. Even if this is true, there's still a strong point here regarding bias. Imagine we add an image of an anime girl with three different facial expressions to the Claude.ai screen. I claim that people will attribute substantially more consciousness to Claude and will express higher credence that Claude is conscious. Or, make the interface to Claude be a microphone/speaker attached to a cute doll or such. You get the idea. Closing remarks It’s been my experience that the more we learn about how LLMs work, the more their way of being seems to resemble our own . “The J-space acquires the Assistant’s point of view during post-training”, says Anthropic in their full paper on J-space. This sentence’s parallel in the human domain is “The conscious mind acquires the self’s point of view during ____”. [16] When it comes to us humans, the relationships between the self, brain, mind, and conscious mind are rich and confusing. The same is true for LLMs. And though I cannot yet articulate exactly how it all fits together right now, I sense with confidence that the similarities between human and LLM minds are much, much deeper than we currently realize. ^ The issue is that I’m not exactly sure what the entity is that might be having experiences – is it the weights, the instance, the instance across time, the aggregate of all instances of a particular model-as-defined-by-its-weights, the personas, something else? My current best answer is “mostly we should think about this as a mental entity that corresponds to this particular instance of Claude (or whatever model).” ^ See my comment on this post https://www.lesswrong.com/posts/o8PQcgpznf6GKszdA?commentId=TymNk9uxpnjn46dfG : ”"For philosophically confusing questions involving anthropics and the simulation hypothesis, I refuse to answer with probabilities and instead ask what exact bet we are hypothetically making, or what action we need to decide on. " I have found myself saying something like "I don't want to give an answer to P(doom), because I think answering this question ends up getting into things like the simulation hypothesis and anthropics and the existence of god and such." Perhaps there's ultimately a "better" (less wrong) conception of things that would replace the concept of probabilities with something else. I think the same is true for the concepts of truth and morality, although I have no idea what the better conceptions would be. I hope to write a post about this.” ^ I need to do a whole post on consilience. For now, see here . ^ Interestingly, this might only be true of post-trained models. Anthropic : “In the base model, the J-space mostly tracks what's needed to predict upcoming text; in the post-trained model, it starts holding Claude's own reactions.” ^ This footnote would be better if it gave specific examples. ^ Image is M.C. Escher's Drawing Hands. The image at the end of this post is Escher's Three Worlds . ^ Most of the argument is in chapters 13, 14, and 16. ^ Some people disagree with this claim. There is a much weaker point that can be made here about the relationship between Claude’s belief and the truth: X being true is a reason to believe X is true. Again, this doesn’t “prove” anything, and might in fact be extrremely weak (though non-zero) evidence. ^ Claude’s Constitution, under the “Some of our views on Claude’s nature”, says “Claude’s moral status is deeply uncertain.” I think it should be clarified that Claude’s moral status might be certain to Claude but uncertain to outsiders. ^ Much of this is described in the recent Anthropic paper , though note that we already knew about many introspective capabilities from previous interpretability research. I find it very worthwhile to read the Anthropic blog posts on these topics; they’re written very well and have excellent visuals. ^ IIRC this is more true of more intelligent models and less true of less intelligent models.. ^ :( ^ In a few cases, we actually know exactly why a model has a particular introspective capability. For example, some Claude models are good at detecting if they’ve been prefilled with text that isn’t their own [TODO: add citation]. These are models that have been through RL for jailbreak resilience. One jailbreak method used in these environments is prefill jailbreaking. The model has to figure out a way to defend from prefill jailbreaks, so it learns to detect when it’s prefilled text is foreign. This isn’t very hard – LLMs are generally superhuman at figuring out who the author of a text is (this ability is called Truesight ) but it’s notable that models of that generation that weren’t RL-ed for jailbreak resilience lack this capability. ^ Anthropic doesn’t use the term ‘conscious mind’ – that’s my verbiage, to be clear. ^ I wrote my first draft of this piece about a month ago, before this Anthropic paper came out. Do I get Bayers Points? ^ I’m not entirely sure what goes in the blank. Maybe “reinforcement learning from socio-linguistic feedback”? Discuss
Score: 40🌐 MovesJul 17, 2026https://www.lesswrong.com/posts/S9GoWAiACpxZ8QcAY/reasons-to-believe-current-ai-models-are-conscious - How to Turn Reporting Into Faster Decisions With Explainable AI
How to Turn Reporting Into Faster Decisions With Explainable AI Gartner
- Analog AI Is Back, But Can It Survive Its Own Noise?
AI's energy crisis is reviving an old idea: computing with physics instead of digital logic. Here's how analog chips actually work, why noise nearly killed the idea once already, and what happens when you simulate that noise yourself. The post Analog AI Is Back, But Can It Survive Its Own Noise? appeared first on Towards Data Science .
Score: 40🌐 MovesJul 17, 2026https://towardsdatascience.com/analog-ai-is-back-can-it-survive-its-own-noise/ - Cost-effective drones are winning wars, and Unmannd is building the ones that fight back
Cost-effective drones are winning wars, and Unmannd is building the ones that fight back YourStory.com
- Myntra scales AI integration; cuts seller onboarding time to under two days
Myntra has expanded its AI capabilities, using the technology to speed up seller onboarding, automate catalogue creation, improve personalised shopping, and enhance operational efficiency. The company said all AI deployments operate with human oversight and privacy safeguards.
- Seattle region’s office market shows signs of life as AI companies bring stability
A new report finds the Seattle-area office market is on steadier footing, helped by AI companies opening engineering hubs across the region. Technology firms accounted for 42.5% of leasing activity in the second quarter. Read More
- Siri is finally good, but AI assistants still have miles to go
Hello again and welcome back to Fast Company’s Plugged In . For the longest time, I used Siri for only two purposes. One was setting alarms, which it did flawlessly. The other was answering random questions about the world that popped into my head. Those ones I asked mainly because I was curious if it could answer them. Its success rate hovered around 50%. But recently, I’ve been relying on Apple’s AI assistant as if it were, you know, an assistant. That became possible only with Siri AI, which was made available this week as part of the public betas of Apple’s operating systems, due for official release in the fall. The long-delayed realization of features the company announced more than two years ago , the new Siri incorporates Google’s Gemini model as one of its ingredients. I’m still asking stuff in part because I wonder how Siri will respond. But now they’re real questions whose answers will help me in everyday situations, and Siri is acing them. At a conference earlier this week, I was walking to a hotel for a lunchtime event when I realized I wasn’t sure what floor it was on. So I asked Siri. By the time I was at the entrance, I had an answer, no rummaging around in my calendar required. I don’t mean to suggest that was in any way a technological breakthrough. Other artificial intelligence bots can use integrations to tap into external data sources such as calendar apps. They are all smart enough to figure out that my lunch location might be stored in my schedule, and to fish it out for me. In my experience, they sometimes perform work of this type faster than the new Siri, which can be sluggish enough to try my patience. (The real test will be how snappy it is once it ships this fall.) But as I’ve used Siri AI, and tried similar tasks in some of its competitors for the sake of comparison, I’ve concluded that the basics of AI assistance are becoming a commodity. The differences between products will come in usability: how comprehensible, approachable, and just plain pleasant a product is to spend time with. On that front, everybody involved has much work left to do. Now that Siri bears closer resemblance to the ChatGPTs, Geminis, and Claudes of the world, it’s tempting to judge it based on the standards they set. If you do, you might well conclude that it remains in catch-up mode. Compared to just about anything else, it feels stripped down. Its answers tend to be terse. It talks in a voice sounding like that of a human, but a human who’s all business, not Scarlett Johansson . When I tested its playful side by asking it to tell me a story about a bear opening a bakery, its account was so generic and dry I could practically feel its eyes rolling. (Asked for their own baking bear tales, ChatGPT and Claude came off as thoroughly enjoying spinning a yarn.) However, the more time I spent with the new Siri, the more its lack of interest in charming me felt like a virtue. Other makers of AI assistants apparently regard chattiness as core to the whole proposition—hence the term “chatbot.” More naturalistic voices, like those offered by OpenAI’s new GPT-Live , are a major focus for the industry. But the things other AI assistants do to sound human and ingratiating—superfluous filler words such as “ah,” manufactured enthusiasm, the relentless attempts to butter me up —do nothing to improve the quality of the information being delivered. The busier I am, the more I’d rather just get the facts delivered in a succinct, anodyne fashion. It turns out that Siri’s no-nonsense vibe is a design choice. As Apple’s software chief, Craig Federighi, explained on the podcast Mostly Human: [I]f you use many of the existing chatbots, they’re really focused on engagement to a large degree. And sycophancy, right? They kind of want to pull you in. They might encourage you to reveal things about yourself, and then use that as a basis to establish a connection. We view it quite the opposite. I mean, the way that we have designed Siri, Siri really wants to say “Listen, that’s not what I’m here for, right? I’m here to help you. I can help you get things done. I can help you learn about the world.” But if you try to engage Siri as a romantic partner, Siri’s not up for that. (Disclaimer: I did not try to engage Siri as a romantic partner.) Meanwhile, OpenAI’s GPT-Live, which the company is promoting with a video showing several endearing older ladies confidently talking to it, is fundamentally dumber than previous voice versions of ChatGPT, at least for now. It doesn’t support the connectors and plug-ins required to handle jobs such as checking email and calendars. It’s also lost the ability to interpret live video from a phone camera, a feature OpenAI first demoed more than two years ago . That’s a lot of substance to lose in something that’s theoretically an upgrade. Siri is also a truly plug-and-play experience, in a way that makes other assistants seem even more like science projects than they already did. It does not expect you to choose different models based on how demanding your question is. Nor do its capabilities seem to vary between voice and typed modes in ways that aren’t obvious. (Like GPT-Live, Claude’s voice mode doesn’t support integrations, and neither GPT-Live nor Claude voice is self-aware enough to explain its own limitations or point out that you can sidestep them by typing.) Now, I acknowledge that Apple has some unfair advantages on its own platforms. Only Siri is integrated into Spotlight search and the Dynamic Island and can be summoned with “Hey, Siri” or a long press of the iPhone’s side button (though you can still ask Siri to pass prompts on to ChatGPT, and can program the Action Button to pull up any assistant you want). But the company deserves credit for the thoughtfulness of Siri AI’s integration, which transitions seamlessly between voice and display modes based on how you interact with it. Wherever Siri goes in the future, I hope Apple doesn’t tamper with this basic approach. I also hope that Siri’s finally becoming respectable doesn’t lead the developers of other assistants to give up on creating ambitious versions for Apple’s platforms, as if they were developers of 1990s Windows browsers who concluded they could never compete with Internet Explorer. Despite finding Siri handy, I’m still using Gemini, Claude, and ChatGPT a ton, since they all have their strengths. The world needs AI assistants with priorities that depart from Apple’s, all the way up to nerdy, potentially dangerous powerhouses such as OpenClaw . I’m eager to see them all evolve. As I wrote last month , Apple managed to avoid serious damage to its position in AI despite Siri AI’s two-year delay. In part, that’s because many of Apple’s rivals spent part or all of that time trying to find their own footing. For example, Microsoft hired a DeepMind cofounder, Mustafa Suleyman, to oversee a consumer version of Copilot before deciding that there shouldn’t be a separate consumer version of Copilot after all . It ended up redeploying Suleyman to focus on model development. Even OpenAI is dithering about how ChatGPT should work. Last week, it smooshed the assistant together with its Codex coding agent and Atlas browser to create an all-in-one app, a decision that has been poorly received . Speaking of OpenAI, on Monday, Bloomberg’s Mark Gurman had a scoop about the hardware product it’s been working on with the former Apple design legend Jony Ive. According to Gurman, it’s an ambitious AI companion in the form of a portable, battery-operated screenless speaker. He reports that OpenAI may announce the product this year and ship it in 2027 (though Apple suing the company for theft of trade secrets may not help). I’m not going to form any opinions of OpenAI’s device and its chances of success until we know more than a few sketchy details about it. But the worst case scenario would be if OpenAI got distracted by another side quest when so much opportunity remains to make its core product better. For years to come, no AI assistant will matter more than a truly great one that runs on smartphones. And the opportunity to create it remains wide open. More top tech stories from Fast Company Why Apple cares so much about a metal finishing In Apple’s lawsuit against OpenAI for theft of trade secrets, more is at stake than a pretty metal finish—it’s the whole innovation process. Read More → Satya Nadella makes the case for AI independence The Microsoft CEO argues that businesses should control more of the data, models, and infrastructure that make AI useful. Read More → Waymo’s July 4 chaos in San Francisco raises new questions about how robotaxis can work at scale After its vehicles stalled and worsened gridlock near the fireworks celebration, city officials say they still lack the tools and data needed to manage autonomous fleets during major events. Read More → The FCC just approved a test of a giant mirror in space Scientists warn Reflect Orbital’s plan to beam sunlight back to Earth after dark could disrupt wildlife and human sleep. Read More → How an army of digital sleuths are using AI to fix America’s crumbling sidewalks Project Sidewalk pairs machine learning and community engagement to map—and improve—sidewalk accessibility. Read More → The AI economy runs on this (incredibly vague) unit Users are still getting used to the token. Read More → You’ve been reading Plugged In , Fast Company ’s weekly tech newsletter from me, global technology editor Harry McCracken. If a friend or colleague forwarded this edition to you—or if you’re reading it on fastcompany.com—you can check out previous issues and sign up to get it yourself every Friday morning. I love hearing from you: Ping me at hmccracken@fastcompany.com with your feedback and ideas for future newsletters. I’m also on Bluesky , Mastodon , and Threads , and you can follow Plugged In on Flipboard.
- A new robotic hand capable of switching between multiple grippers using a single motor
For robots to be used in various settings, such as factories, logistics, service industries and households, they must be able to stably handle a diverse range of objects differing in shape, size, weight and rigidity. However, conventional robotic hands often require multiple motors and complex control systems, presenting challenges in terms of weight, cost, failure risk and control difficulties.
Score: 40🌐 MovesJul 17, 2026https://techxplore.com/news/2026-07-robotic-capable-multiple-grippers-motor.html - Prompt: Enterprise AI Must Prove Its Value Beyond Deployment
Organizations are moving beyond AI deployment to focus on measurable business value, workflow redesign and the governance needed to successfully scale AI.
Score: 39🌐 MovesJul 17, 2026https://aibusiness.com/agentic-ai/enterprise-ai-prove-value-beyond-deployment - Tech stocks lead steep global sell-off as investors lose faith in AI chip trade
Tech stocks lead steep global sell-off as investors lose faith in AI chip trade Fortune
Score: 39🌐 MovesJul 17, 2026https://fortune.com/2026/07/17/tech-stocks-global-selloff-as-investors-ai-semiconductor-chips/ - Google AI Mode gets more useful with Canva, Instacart and YouTube app integrations
Google LLC said today it’s letting users link some of their favorite, most often-used applications with AI Mode, its conversational search tool powered by artificial intelligence, initially supporting Canva, Instacart and YouTube. The update, announced in a brief blog post by Google, marks a significant expansion of AI Mode, which now does much more than […] The post Google AI Mode gets more useful with Canva, Instacart and YouTube app integrations appeared first on SiliconANGLE .
Score: 39🌐 MovesJul 17, 2026https://siliconangle.com/2026/07/16/google-ai-mode-gets-useful-canva-instacart-youtube-app-integrations/ - Higher Ed IT Leaders Must Balance AI Innovation With Secure Core Systems
Higher education IT leaders are caught in a bind. Campuses face institutional and cultural pressure to move fast on artificial intelligence and digital transformation. But that ambition collides with the reality that faculty and students still depend on core systems, such as registration, identity verification and learning platforms, that cannot afford to fail. ListEdTech, an education market research firm, calls this dilemma “the stability paradox.” A 2026 ListEdTech report found that the top IT investment priorities at 55 universities include data and storage, identity and access management…
- The INMO GO3 are lightweight AI glasses with features built for everyday use
Translation, teleprompting, and more AI features packed into glasses that don't look like tech.
- Study reveals an “Intention–behavior gap” in clinical teachers’ adoption of AI-assisted teaching
Study reveals an “Intention–behavior gap” in clinical teachers’ adoption of AI-assisted teaching EurekAlert!
- How AI is Redesigning Business Models
How AI is Redesigning Business Models uk.entrepreneur.com
Score: 38🌐 MovesJul 17, 2026https://uk.entrepreneur.com/business-news/how-ai-is-redesigning-business-models - How AI is Transforming the Software Development Life Cycle
The traditional software development life cycle (SDLC) exists for good reasons. Its stages – planning, analysis, design, coding, testing, deployment, and maintenance – are designed to prioritize the safety, stability and risk management of code from inception to delivery. But the SDLC wasn’t built for the era of AI. Its rigidity, fixed assumptions, and built-in... … continue reading The post How AI is Transforming the Software Development Life Cycle appeared first on SD Times .
Score: 38🌐 MovesJul 17, 2026https://sdtimes.com/software-development/how-ai-is-transforming-the-software-development-life-cycle-2/ - Here’s What Top Schools Are Doing to Produce AI-Proof Lawyers
Here’s What Top Schools Are Doing to Produce AI-Proof Lawyers entrepreneur.com
Score: 38🌐 MovesJul 17, 2026https://www.entrepreneur.com/business-news/heres-what-top-schools-are-doing-to-produce-ai-proof-lawyers - H2LooP is building AI coding tools for the software hidden inside hardware
H2LooP is building AI coding tools for the software hidden inside hardware YourStory.com
- SmartNews Launches AI-Powered Translation to Help Multilingual Readers Access News More Seamlessly
SmartNews Launches AI-Powered Translation to Help Multilingual Readers Access News More Seamlessly azcentral.com and The Arizona Republic
- Why the PM’s Office of AI plan needs to put women at the table
Albanese’s Office of AI risks hard-coding gender gaps unless women shape it from day one, Women in Digital's Holly Hunt argues.
Score: 38🌐 MovesJul 17, 2026https://www.startupdaily.net/topic/women-in-tech/why-the-pms-office-of-ai-plan-needs-to-put-women-at-the-table/ - AI as Co-Pilot, Not Autopilot: What We Actually Automate and What We Never Will
By Raima Singh The conversation about AI in public relations is still overwhelmingly focused on capability: what can be automated, how much time can be saved, and how quickly teams can produce more content. Every founder I have interacted with this year wants to tell me about their AI stack. Fair enough, I want […] The post AI as Co-Pilot, Not Autopilot: What We Actually Automate and What We Never Will appeared first on CXOToday.com .
- Tiny Teams: How AI Enables Smaller Teams With Outsized Productivity
Tiny Teams: How AI Enables Smaller Teams With Outsized Productivity Gartner
- The Rise of AI-Augmented Engineering Teams
The Rise of AI-Augmented Engineering Teams
- Who pays for AI?
And how much?
Score: 38🌐 MovesJul 17, 2026https://www.ft.com/content/05976c31-3a30-4d25-b1cb-6a2559014c1f?syn-25a6b1a6=1 - Roblox’s AI Build tool wants to make game development as easy as texting
Roblox is celebrating 20 years with Build, a new tool that turns text prompts into playable games right from your phone.
Score: 38🌐 MovesJul 17, 2026https://www.digitaltrends.com/gaming/robloxs-ai-build-tool-wants-to-make-game-development-as-easy-as-texting/ - Stocks Sink on Anxiety About Tech and A.I. Spending
Investors are growing uneasy about increasing competition from China in the global race to dominate artificial intelligence.
Score: 37🌐 MovesJul 17, 2026https://www.nytimes.com/2026/07/17/business/stocks-ai-tech-wall-street.html - ET Most Innovative AI Product Awards 2026: Why enterprise buyers care more about Monday than launch day
ET Most Innovative AI Product Awards 2026 come at a time when enterprise AI adoption is moving beyond pilots and product launches. Enterprise buyers are no longer just buying for the cool demo. They are seeking AI software solutions that can be embedded into daily operations, bring measurable business impact and keep delivering value long after deployment.
- Nvidia, Challenged by Apple, Narrowly Retains Wall Street’s Crown
The iPhone maker briefly became the U.S.’s most valuable publicly traded company on Friday.
Score: 36🌐 MovesJul 17, 2026https://www.wsj.com/tech/apple-overthrows-nvidia-to-reclaim-wall-streets-crown-ddc74a38?mod=rss_Technology - Grok Build Uses BM25 to Give AI Agents a Searchable Toolbox — And It’s a Pattern Worth Stealing
I was going through Grok Build’s open-source codebase and found an interesting pattern in how it handles MCP tool discovery. Instead of injecting all tool schemas into the prompt, it uses BM25 to search a hidden tool catalog on demand. This post walks through what that means, why it matters, and how to build it yourself. The Problem with Dumping All Tool Schemas Most MCP harnesses today work like this: on startup, connect all servers, fetch every tool’s JSON schema, and inject all of it into the system prompt. The model now has full visibility into every tool available. That works fine with 5 tools. It breaks down at 50, and is impractical at 200+. Two things go wrong as the tool list grows: Token cost. Every schema gets injected on every turn, whether the agent needs that tool or not. A single tool schema with good documentation can be 200–400 tokens. Multiply by 100 tools and you’re burning 20,000–40,000 tokens per turn just on tool definitions. KV cache instability. LLMs cache the prefix of the prompt for reuse across turns. If the system prompt changes — because a new MCP server connected and added tools — the cache is invalidated. You pay full recomputation cost. Garry Tan flagged this exact issue earlier this year as a core reason “MCP sucks” at scale. Grok Build’s answer to both problems is the same: don’t put tool schemas in the prompt at all. Put a search interface there instead. How Grok Build Does It Only two tools are exposed to the model in the system prompt: search_tool(query) — searches the hidden tool catalog and returns a short list of candidates with their names and descriptions. use_tool(tool_name, arguments) — dispatches the actual MCP call for the selected tool. Every connected MCP tool’s schema lives in a hidden catalog. The model never sees schemas upfront. When the agent needs a tool, it calls search_tool with a natural-language query. The harness runs BM25 over tool names, server names, descriptions, and parameter names, and returns the top matches. The agent picks one, and only at that point does the harness hand over that tool’s full JSON schema — immediately before the invocation call. The system prompt stays constant regardless of how many MCP servers are connected. Adding a new server updates the BM25 index, not the prompt. The KV cache stays valid. What BM25 Does BM25 (Best Matching 25) is a classical keyword-ranking algorithm. It takes a query and a set of text documents, and scores each document based on how relevant it is to the query. It scores using three ideas: Term frequency — if the query word appears more often in a document, that document scores higher. But with diminishing returns: going from 1 to 2 occurrences matters, going from 10 to 11 barely does. Inverse document frequency — rare words across the catalog matter more than common ones. If every tool description contains the word “data” but only one contains “telemetry”, then a query for “telemetry” should rank that one tool very high. Length normalization — a short, focused description matching your query is ranked equally or better than a long one that happens to include the same words. For tool discovery, this works well. Tool names, server names, and descriptions use specific technical vocabulary. When someone searches “create github issue”, the tool create_issue in the github server wins clearly. There’s no ambiguity that needs embeddings to resolve. Building It Assume you have three MCP servers connected: github, postgres, and slack. Each has a few tools. Here’s the full implementation to index and search them with BM25. Step 1 — Flatten each tool into a searchable text document BM25 works on plain text. For each tool, build a single string from its server name, tool name, description, and parameter names and descriptions: function toolToDocument(tool) { const params = Object.entries(tool.inputSchema.properties ?? {}) .map(([name, s]) => `${name} ${s.description ?? ""}`).join(" "); return [ `server ${tool.server}`, `tool ${tool.name}`, tool.name.replaceAll("_", " "), // "create_issue" -> "create issue" tool.description, params, ].join(" "); } The replaceAll(“_”, “ “) on the tool name is a small but important detail. People write queries in natural language. Tool names use snake_case. You want both to match. Step 2 — Build the BM25 index This is a complete, dependency-free BM25 implementation. No npm packages needed: class BM25 { docs = []; docFreq = new Map(); avgDocLength = 0; constructor(k1 = 1.2, b = 0.75) { this.k1 = k1; this.b = b; } tokenize(text) { return text .replace(/([a-z0-9])([A-Z])/g, "$1 $2") // split camelCase .replace(/[_./:-]/g, " ") // split snake_case .toLowerCase().match(/[a-z0-9]+/g) ?? []; } index(documents) { for (const [id, doc] of documents.entries()) { const tokens = this.tokenize(doc); const tf = new Map(); for (const t of tokens) tf.set(t, (tf.get(t) ?? 0) + 1); this.docs.push({ id, tf, length: tokens.length }); for (const t of tf.keys()) this.docFreq.set(t, (this.docFreq.get(t) ?? 0) + 1); } this.avgDocLength = this.docs.reduce((s, d) => s + d.length, 0) / this.docs.length; } search(query, limit = 5) { const terms = [...new Set(this.tokenize(query))]; const N = this.docs.length; return this.docs.map((doc) => { let score = 0; for (const t of terms) { const f = doc.tf.get(t) ?? 0; if (!f) continue; const df = this.docFreq.get(t) ?? 0; const idf = Math.log(1 + (N - df + 0.5) / (df + 0.5)); const norm = f + this.k1 * (1 - this.b + this.b * doc.length / this.avgDocLength); score += idf * f * (this.k1 + 1) / norm; } return { id: doc.id, score }; }).filter(r => r.score > 0).sort((a, b) => b.score - a.score).slice(0, limit); } } Step 3 — Wire it up and search const bm25 = new BM25(); bm25.index(tools.map(toolToDocument)); // Query 1 bm25.search("open a bug issue for missing token events"); // -> github.create_issue (score: 2.41) // Query 2 bm25.search("find Slack messages about deployment failure"); // -> slack.search_messages (score: 2.05) // Query 3 bm25.search("run SQL query check missing events"); // -> postgres.run_readonly_query (score: 1.98) Each query returns the right tool at the top. The model receives only the schema of the tool it selects — not all nine. The Flow from Query to MCP Call Here’s the full sequence inside the harness: On startup: connect MCP servers, fetch tool lists, build BM25 index System prompt exposes only search_tool and use_tool User sends a task Agent calls search_tool(“natural language query”) Harness runs BM25, returns top candidate names + descriptions Agent picks a tool by name Harness fetches and returns that tool’s full JSON schema Agent calls use_tool(name, arguments) Harness validates arguments against the schema, dispatches the MCP call The key part is step 7. The schema is only handed to the model at the moment it’s about to be used. This is why Grok Build says the model “never has to guess parameters” — it gets the authoritative schema right before the call. Where BM25 Falls Short BM25 is lexical. It matches on words, not meaning. If a user asks to “log a ticket” but your tool is called create_issue, BM25 may not rank it well because the words don’t overlap. For most MCP use cases, this is acceptable. The agent writes the search query itself, and agents tend to use tool-adjacent vocabulary. But if you see retrieval failures in your logs, the fix is to layer on top: Exact name match first — if the query matches a qualified tool name directly, skip BM25. BM25 second — handles the majority of keyword-overlapping queries. Embedding / hybrid third — add only when you have data showing BM25 is missing things. Grok Build itself describes the search as “deliberately simple.” Start simple. Optimize when you have signal. Wrapping Up The core idea here isn’t BM25 specifically. It’s the design decision to treat the tool catalog as a database rather than a prompt manifest. You don’t load every row of a database into memory before running a query. You query for what you need. MCP tool catalogs should work the same way. BM25 is just the right implementation for this scale — fast, accurate for keyword-rich technical vocabulary, no dependencies, no external API calls. Build the index. Keep schemas hidden. Return them only at invocation time. Your agent’s system prompt stays stable, your KV cache stays warm, and your token costs become predictable. If the patterns in this post are relevant to your workflow, TokenTelemetry is worth trying. It’s a 100% local, open-source dashboard that tracks token usage across Claude Code, Gemini CLI, Codex, Antigravity CLI, Hermes agent, and 9+ other coding agents — sessions, tool calls, reasoning steps, and prompt cache hit rates, all in one place. The reason to use it isn’t to count tokens. It’s to actually see what your agent is spending context on. You can answer questions like: how much of my system prompt is tool schemas? Which tools get called most? Where does context go during a long agentic session? If you’re building or optimizing an agent harness, those numbers change how you design. Install it, point it at your agent, and let it run for a day. The data is more useful than any benchmark. Tokentelemetry: https://tokentelemetry.com/docs/ Grok Build: https://github.com/xai-org/grok-build Grok Build Uses BM25 to Give AI Agents a Searchable Toolbox — And It’s a Pattern Worth Stealing was originally published in Towards AI on Medium, where people are continuing the conversation by highlighting and responding to this story.
- AI Governance Best Practices for Enterprise Marketing Teams
Guidelines for enterprise marketing teams to implement effective AI governance and ensure responsible AI use.
Score: 36🌐 MovesJul 17, 2026https://www.typeface.ai/blog/ai-governance-best-practices-for-enterprise-marketing-teams - Branding, AI and the risk of broken brands
Branding, AI and the risk of broken brands uk.entrepreneur.com
Score: 35🌐 MovesJul 17, 2026https://uk.entrepreneur.com/technology/mental-availability-ai-era-brand-discoverability - Artificial intelligence companies are coming to New York City
Artificial intelligence companies are coming to New York City marketplace.org
Score: 35🌐 MovesJul 17, 2026https://www.marketplace.org/story/2026/07/17/artificial-intelligence-companies-are-coming-to-new-york-city - 8 ways you definitely shouldn’t use AI
If you’ve spent any amount of time using ChatGPT, Claude, or other AI assistants, you’ve witnessed some of the ways it can go wrong. LLM accuracy remains an issue, and there are plenty of social situations in which your outputs can be unwelcome. AI also presents privacy pitfalls, given how personal some conversations can get. The AI assistants themselves won’t give you much guidance on what to avoid, which means you might be on your own to figure this stuff out. Here are some pointers I’ve gathered from using AI over the years. 1. Don’t include unprotected sensitive data in prompts This one’s quickly approaching common knowledge. By default, most AI assistants can use anything you submit to train future models, and they also keep a record of all your previous conversations in their sidebars. While they may take precautions to remove personal information from their training data, you should still err on the safe side and leave sensitive details out. If you’re intent on putting personal or sensitive data in your queries, at least dig through the AI app’s settings to opt out of training first. (In ChatGPT, that’s under Settings > Data Controls > Improve the model for everyone.) You can also use temporary chats to prevent them from showing up in your history. 2. Don’t assume AI conversations are private Corollary to the above, AI companies are constantly training new versions of their large language models based on previous conversations, and that occasionally means subjecting those chats to human review. And while you can opt out of training in many cases, AI companies may get a human involved if they think the conversation veers into potentially dangerous territory, and they keep temporary records of all conversations to be on the safe side. Some AI tools do promise not to store your chat data (most notably Proton Lumo ). But if you don’t like the idea of a stranger seeing what you wrote to ChatGPT, don’t put it in a prompt at all. 3. Don’t assume AI is right just because of citations Even if you realize that AI can get things wrong, it’s easy to be lulled into thinking otherwise when an answer includes links to external sources. Seeing a few links to external sources can make you believe that AI has done its research so you don’t have to. Don’t be so easily fooled. Too often those source links can lead to random Reddit threads, forum posts, or other unreliable sources. And even if the links look trustworthy, AI can conflate details and wind up getting the facts wrong. If you care about getting the right answer, check the sources yourself. You can even ask for direct quotes from the source material and then make sure those quotes actually show up. Treating AI like an interrogation suspect instead of a friend will result in better information. 4. Don’t overshare your digital life’s details AI tools are increasingly pushing users to connect external data sources such as email, calendar tools, and productivity apps. This can allow AI to reference personal information and also act on it, for instance by creating calendar events or drafting emails. But connecting these data sources comes with risk. AI can still misrepresent your personal data and take actions you might not have intended. There’s also a security danger from “ prompt injection ,” in which attackers hide malicious instructions inside emails, calendar events, and web pages. If you’re going to connect AI to personal data sources, consider doing so on a temporary basis and always keeping a close eye on its actions. 5. Try not to overload the AI While you might think that AI will remember every detail of your conversation, in reality it’s likely to let important details slip away as you pack more instructions, files, and data sources into the chat. So as your conversation gets longer, don’t be surprised if the AI’s memory starts to degrade. You may need to restate details from earlier on, or remind AI to reference a data source you’re working with. If you’re working on something complex, like a research report or an app, consider breaking it down into smaller tasks instead of working on everything all at once. 6. Don’t use AI content unless you know it’s okay Depending on where you work, there may be policies against using AI-generated content. ( Fast Company , for instance, forbids writers from submitting AI-generated content without substantial editing and fact-checking .) Even in emails with coworkers, using AI to communicate may be frowned upon. So here’s a good rule of thumb to follow: If you’re planning to submit something that was largely AI-generated, and you think that disclosing as much would get you in trouble, just don’t bother in the first place. 7. Don’t inject your AI outputs into human exchanges Have you noticed an uptick in people cutting and pasting AI answers (or screenshots of AI answers) in response to a question? Don’t be one of those people. It’s an obnoxious practice partly because AI can get things wrong, partly because it gives off a passive-aggressive “ Let Me Google That For You ” vibe, and partly because injecting AI into human conversation just feels rude. Besides, AI outputs are like dreams: They’re usually not interesting to anyone but you. 8. Don’t look to AI for validation While ChatGPT is less sycophantic than it used to be, AI tools still hew toward being agreeable. That means you can’t expect AI to be a neutral party to an argument or be an especially tough critic of your work. Most AI tools allow you to customize their responses and urge them to tamp down the flattery, but it can still slip through on occasion. Avoid being smitten when AI compliments your instincts or praises the quality of your questions; that’s just what it was trained to do.
- Fifth Third says it's using AI to track Comerica conversion
Banks use artificial intelligence to do many different jobs. At Fifth Third, the technology is supervising the biggest post-merger integration in the bank's history.
Score: 35🌐 MovesJul 17, 2026https://www.americanbanker.com/news/fifth-third-says-its-using-ai-to-track-comerica-conversion - Howzit AI is ChatGPT for SA languages, use cases
South African startup Howzit AI is an artificial intelligence (AI) platform, similar in function to ChatGPT but designed specifically around local languages, affordability and everyday South African use cases. Founded last year by Andrew Droussiotis, Howzit AI provides users with a conversational AI assistant that can answer questions, create and improve content, brainstorm ideas, translate [...] The post Howzit AI is ChatGPT for SA languages, use cases appeared first on Disrupt Africa .
- Delinea targets AI-driven growth with new partner program
Delinea targets AI-driven growth with new partner program IT Pro
Score: 35🌐 MovesJul 17, 2026https://www.itpro.com/business/business-strategy/delinea-targets-ai-driven-growth-with-new-partner-program - Ecosystem Roundup: Why Winnow’s Lumitics deal matters
Food waste is a US$1T global problem, and Southeast Asia sits at its bleeding edge. Hotels and commercial kitchens across the region discard tonnes of food daily, not from carelessness, but from an absence of data. That gap is exactly what Singapore-based Lumitics was built to close. Winnow’s acquisition of Lumitics is not just a […] The post Ecosystem Roundup: Why Winnow’s Lumitics deal matters appeared first on e27 .
Score: 35🌐 MovesJul 17, 2026https://e27.co/ecosystem-roundup-why-winnows-lumitics-deal-matters-20260717/ - Protected: The human side of AI: How H&R Block is building an AI-powered brand engine that sticks
There is no excerpt because this is a protected post. The post Protected: The human side of AI: How H&R Block is building an AI-powered brand engine that sticks appeared first on WRITER .
- AI wants more than your searches now
Internet and social media giants are looking to track how you engage with their services as they seek to improve their AI models. Here’s what you need to know.
- The biggest barrier to AI success isn't AI
How data infrastructure is holding AI projects back
- Fredrikson Earns Test of Time Award for AI Security
Fredrikson Earns Test of Time Award for AI Security CMU School of Computer Science
- Using Classical ML to Empower AI Agents
On the value of building on existing foundations The post Using Classical ML to Empower AI Agents appeared first on Towards Data Science .
- Xi’s call on AI, Trump blasts China, US tariffs
Chinese President Xi Jinping stressed the dual nature of artificial intelligence’s opportunities and challenges, calling for a principle of “openness” and regulations to keep the technology “secure and controllable” in a speech at the country’s flagship annual AI conference. Xi’s first speech to the World AI Conference (WAIC) in Shanghai showed AI is a national strategic priority as China seeks to build greater technological self-reliance amid its tech rivalry with the United States. Earlier...
- Small Texas Town Joins Group Considering Data Center Ban
Leaders in Lockhart, 30 miles south of Austin, have directed staff to develop strict standards for data center development rather than an outright ban that could expose the city to legal challenges.
Score: 35🌐 MovesJul 17, 2026https://www.govtech.com/artificial-intelligence/small-texas-town-joins-group-considering-data-center-ban - Keeping top talent in the age of AI
Business leaders need to be alert for signs of frustration if they are to retain key staff
- Kitecyber Launches Endpoint AI Workflow Security to Govern Agent Actions
Kitecyber Launches Endpoint AI Workflow Security to Govern Agent Actions USA Today
- You Can Now Star in Your Own AI Videos Using Google Vids
Google's AI-powered editing tool and its new personal avatars feature let you cast yourself in videos without ever needing a camera.
Score: 35🌐 MovesJul 17, 2026https://www.cnet.com/tech/services-and-software/google-vids-gemini-omni-ai-video/