Yu Su (NeoCognition) argues that intelligence and expertise are distinct and roughly orthogonal: intelligence is reasoning through an unfamiliar problem from handed context (where frontier models keep improving), while expertise is accumulated, situated competence that almost nobody is scaling. He explains why coding agents succeed — code is a symbolic, language-native world with tests as rewards — while the rest of digital work consists of millions of heterogeneous "micro worlds" that resist compression by a static model, framing this as a modern Moravec's paradox. He contrasts intelligence expanding search with expertise compressing it, warns that scaling intelligence alone produces the "world's smartest novice," links expertise to continual learning (parametric and nonparametric), and closes with the provocation of achieving unbounded expertise from bounded intelligence. Topics per timestamps include agents before language models, why code was the ideal first market, how experts see differently, conditionality and judgment, reliability vs. plasticity, and specialization as the next data opportunity. # Channel Presentation Observations ## Explicit Self-Description (Creator Claims) - **Dual identity: academic + founder.** The speaker states: *"I'm a professor at Ohio State... I also have another job which is a CEO at a company called Neocognition and we focus on agents and continual learning."* - **Company scope claim.** The channel/creator claims Neocognition "focus[es] on agents and continual learning," reiterated at the close: *"this is the future we're building towards at Neocognition."* - **Talk framing claim.** The speaker describes the content as *"won't be too technical... mainly a conceptual one"* addressing a *"conceptual distinction... between what is intelligence and what is expertise."* ## Observable Presentation Patterns ### Intended Audience - **Technical/industry AI audience.** Evidence: opens with *"I understand that I'm standing between you and the launch,"* indicating a conference/launch-event setting; references to *"Jack just gave some definition earlier,"* "Karpathy," and *"as Satya said two weeks ago"* presume familiarity with AI industry figures and discourse. - **Practitioners concerned with agent deployment and cost.** Evidence: discussion of enterprises *"curb[ing] their token maxing efforts"* and challenges *"deploying these agents in enterprise settings... and in personal settings."* ### Claimed Scope of Content - **AI agents and continual learning as core topics.** The entire talk is structured around: agent history (expert systems → deep RL → language agents), limits of coding agents, an intelligence-vs-expertise framework, and a four-part definition of continual learning (*"adaptive compression of experience into reusable structures"*). - **Research-agenda framing.** Explicit open questions presented: measuring expertise, the *"trade-off between reliability and plasticity,"* combining *"parametric or nonparametric"* learning, and specialization-to-generalization (*"the next internet scale data opportunity"*). ### Demonstrated Presentation Style - **Conceptual framing with coined terms.** Introduces original terminology: *"world's smartest novice,"* *"unbounded expertise from bounded intelligence,"* *"escape intelligence,"* and a *"modern version of the Moravec paradox."* - **Persuasive, thesis-driven structure.** States a clear thesis: *"how can something be so smart but also so brittle at the same time"*, then systematically defends it. - **Call-to-action endings.** Concludes with a vision: *"let's start scaling expertise... intelligence is already becoming abundant... but expertise is still scarce,"* framing societal benefit (personal health care, financial advisors, tutors; company "learning loops"). ## Notable Distinctions - Claims about the channel itself are limited to the speaker's professorship and Neocognition's agent/continual-learning focus; all other content is demonstrated presentation (argumentation style, audience targeting) rather than stated channel identity. # Delivery Observations — Yu Su, "Intelligence + Continual Learning = Expertise" ## Overall Tone & Register - **Conversational, low-key academic style**: The talk is explicitly framed as non-technical — *"it won't be too technical but… it will be mainly a conceptual one."* - **Self-aware, audience-conscious opening**: *"I understand that I'm standing between you and the launch, so I'll try to be quick."* - Frequent casual asides and hedging: *"I don't know how much uh thought has changed uh since like uh last time."* ## Verbal Fillers & Disfluencies - Heavy use of **"uh" and "um"** throughout, e.g., *"Um, you give it the provin statements, the context, the tools…"* - Frequent **"like"** as a filler: *"try to spin up like 100 different uh like parallel ways to to try to solve the problem."* - Occasional self-corrections: *"it's uh the 20110s"*, *"so is extremely idiosyncratic"* (dropped subject), *"to provide more food for thoughts."* ## Recurring Rhetorical Devices - **Rhetorical questions as structure**: - *"So then what happens when we leave the privileged world of code?"* - *"So how can something be so small but also so brittle at the same time?"* (Note: likely a transcription artifact — "smart" not "small" — stated with uncertainty.) - **"Food for thought" repeated twice**: - *"hopefully this will provide some food for thought before launch"* - *"to provide more food for thoughts"* - **"Right?" as a checking/tag particle**, very frequent: *"Right. First a bit of a history."*, *"Every episode is more or less like independent from each other here… Right?"* - **Rhetorical framing with "I think"** as a recurring hedge/claim marker (appears dozens of times): *"I think we are actually witnessing a modern version of the Maravx paradox."* ## Teaching Habits - **Explicit signposting of talk structure**: *"The second part of the talk I will try to establish the differences between intelligence and expertise."* - **Working definitions before contrasts**: *"Here are the working definitions. For intelligence, it's the capacity to reason through unfamiliar problems…"* - **Flagging importance**: *"this is maybe the most important figure in this talk."* - **Coined, quotable labels for concepts**: - *"the world's smartest novice"* - *"unbounded expertise from bounded intelligence"* - *"escape intelligence"* - **Concrete everyday examples to ground abstractions**: scheduling a meeting as *"a constraint optimization problem"*, an expert scanning *"a gigantic bug reports"* to locate failure points. ## Vocabulary & Framing Patterns - Contrast-pair vocabulary: **intelligence vs. expertise**, **reliability vs. plasticity**, **parametric vs. nonparametric**, *"hard things are easy, easy things are hard."* - Business/market framing mixed with research framing: revenue figures (*"grown 400 times"*, *"40 billion"*, corrected to *"maybe 60 billion"*), *"first mass markets"*, *"the next internet scale data opportunity."* - **External authority citations**: Andrew Kapathy (transcribed "Capathy"), Satya Nadella ("as Satia said two weeks ago"), Anthropic's revenue graph, cognitive science on expert perception. ## Uncertainty & Honesty Markers - Explicitly flags uncertain data: *"to uh 40 billion I think the newest number is maybe 60 billion"* — the *"I think the newest number is maybe"* construction signals hedging on facts. - Notes a boundary condition on his own chart: *"Of course, this is assuming like a given time horizon and experience horizon."* - Acknowledges terminological confusion in his field: *"continual learning… is such a confusing term… and Jack just uh gives some definition earlier uh with like 10 different names."* ## Closing Pattern - Ends with an aspirational **call to action** explicitly labeled: *"Okay. So finally a call to action… let's start scaling expertise"*, with a vision-of-the-future close and brief company mention: *"this is the future we're building towards new cognition."* ## Stable vs. Isolated Features - **Stable**: filler density, "right?" tagging, "I think" hedging, definitional teaching structure, contrast framing. - **Isolated**: the launch reference (opening only), the named-concept coinages (concentrated mid-late talk), the revenue anecdote (single instance) — these do not recur and should not be treated as habits. # Observation Document: "Intelligence + Continual Learning = Expertise" — Yu Su, NeoCognition ## Speaker - **Yu Su** (self-identified as "IU" in transcript; per title metadata: Yu Su), Professor at Ohio State and CEO of **NeoCognition**, a company focused on agents and continual learning. ## Central Thesis - Distinguish **intelligence** from **expertise**; argue that **continual learning** bridges the two. - Motivating puzzles Su poses: why are agents so successful at coding yet poor elsewhere, and why are they so token-inefficient that companies are curbing token usage? ## Historical Framing - Agents are not new, but earlier efforts (1960s–80s expert systems/logical agents; 2010s deep RL agents) captured only limited facets of intelligence (logic, single-modality perception). - Only recently have multimodal LLMs and language agents encoded **multisensory input into a unified neural representation conducive to symbolic reasoning and communication** — previously a uniquely human trait — marking "a new evolutionary stage of machine intelligence." ## Coding Success vs. Failure Elsewhere - Coding agents found the first mass market; illustrated by **Anthropic's revenue growing ~400x in under two years** to ~40 billion (Su notes the newest number may be ~60 billion annualized), largely driven by coding. - Coding is the **ideal market**: language-native, symbolic, structured, with rewards/tests built in. - Outside code ("the privileged world"), agents are brittle; Su cites **Andrew Karpathy's** claim it will be the "decade of agents" rather than "year of agents" because agents lack computer use and continual learning — difficulties Su says largely persist. ## Moravec's Paradox (Modern Version) - "Hard things are easy, easy things are hard": agents excel at symbolic reasoning (coding, math) but struggle with everyday digital work. - Reasoning: modern society is **millions of idiosyncratic micro-worlds** (domains, professions, companies configure the same software differently), each with unique local physics — structures, constraints, affordances, dynamics. Too heterogeneous and dynamic for a monolithic model to compress statically; agents must **continually learn on the job** to acquire specialized expertise per micro-world. ## Working Definitions - **Intelligence**: capacity to reason through unfamiliar problems from available context; episodes are largely independent; what frontier models do well. - **Expertise**: accumulated, situated competence — acting reliably, efficiently, with judgment to achieve reproducibly superior performance in a particular domain. ## Components of Expertise (from cognitive science) - Experts "see the world differently": superior pattern recognition (e.g., locating plausible failure points in a large bug report). - Deep problem structure (scheduling a meeting as constraint optimization over authority, priorities, urgency). - Conditionality of rules — knowing preconditions and when exceptions allow bending rules. - **Judgment and taste**: knowing what is high quality and **when to stop / when it's good enough**. - Experts effectively build a **world model** of their micro-world underpinning perception, reasoning, decision-making. ## Intelligence vs. Expertise Contrasts - Intelligence solves problems *given* context; expertise *brings the right context*. - Intelligence expands search (brute force, parallel attempts); expertise **compresses the search space** via learned shortcuts. ## Continual Learning Definition Su's definition: **adaptive compression of experience into reusable structures for future behavior**, with four dimensions: 1. **Experience**: episodes, semantic facts, procedures, human/environment feedback. 2. **Compression**: vectors, symbolic indices, distillation into parameters, RL. 3. **Adaptive**: prior learning shapes future compression. 4. **Reusable structures**: adapters/parameters, vectors, graphs, skills, world models — used for recall, prediction, planning, control, or as value functions. - The many instantiations of these dimensions explain why the field is "confusing"; Su references an earlier speaker ("Jack") giving ~10 different definitions. ## Key Figure: Orthogonality - Raw intelligence (x-axis) vs. expertise (y-axis) are **largely orthogonal**. - Scaling intelligence without continual learning yields **"the world's smartest novice"** — brute-forcing every problem. - Continual learning algorithms set the **slope** of the learning curve (simple in-context learning = shallow slope). - Preferred future: **unbounded expertise from bounded intelligence** — once intelligence crosses a threshold ("escape intelligence"), learning alone yields unbounded expertise. Implication: perhaps existing models suffice and better continual learning algorithms are the missing piece. ## Open Questions Posed - How to define/measure expertise (likely environment-specific). - **Reliability vs. plasticity** trade-off (stability resists change; plasticity embraces it); humans are cited as a "living existence proof" of both. - **Parametric vs. nonparametric learning**: Su's stated belief is both are needed; how to synergize them. - Specialization may lead to better **generalization**: public data is exhausted; private micro-worlds are the next "internet-scale data opportunity" if specialized agents learn in situ and channel learning back to general models. ## Call to Action - **Scale expertise** as a new dimension: intelligence is becoming abundant (frontier models "probably smarter than average humans"); expertise is still scarce. - Vision: abundant expertise — personal healthcare, financial advisors, tutors for everyone; per-company **human–AI learning loops** creating institutional memory and moats (attributed to "Satia" two weeks prior), companies keeping control of their "means of production." - Abundant expertise lowers friction, making new types of work economically viable ("across the threshold of worth doing"). ## Qualifications Preserved - Revenue figure hedged ("I think the newest number is maybe 60 billion"). - Orthogonality and threshold claims framed as Su's views ("I think," "if this is indeed true"). - Expertise-intelligence axis assumes "a given time horizon and experience horizon." # Speaker Attribution & Provenance Record ## Video Metadata - **Title:** Intelligence + Continual Learning = Expertise — Yu Su, NeoCognition - **Video ID:** `01a15ca7-fa91-77f0-94e4-2b7645c22ad2` - **Curated metadata:** Not supplied ## Speakers & Roles ### Primary (Sole) Speaker: "Yu Su" (self-introduced as "IU" in transcript; title names him Yu Su) - **Roles:** - **Professor** at Ohio State University (self-identified) - **CEO** of **NeoCognition**, a company focused on agents and continual learning (self-identified) - **Talk presenter / keynote speaker** (mentions speaking "before launch," suggesting a launch event context) - No other speakers appear in the transcript; no guest, narrator, or additional roles present. ## Promotional Segments (Flagged) - **Opening self-introduction** promoting NeoCognition as speaker's company ("I also have another job which is a CEO at a company called Neocognition and we focus on agents and container learning"). - **Closing call to action** promoting NeoCognition's mission: "this is the future we're building towards at Neocognition." ## Referenced Third-Party Claims & Named Sources (as cited by the speaker) - **Anthropic revenue figures:** "in just under two years their revenue has grown 400 times to 40 billion... maybe 60 billion annualized runway" — attributed by speaker, no source given. - **Andrew Karpathy** (transcribed "Andrew Capathy"): claim that it will be "the decade of agents" due to failures in computer use and continuous learning — third-party quote paraphrased by speaker. - **Satya Nadella** (transcribed "Satia"): quoted as saying two weeks prior about enabling "human AI learning loop at each company" — third-party quote paraphrased. - **Moravec's paradox** (transcribed "Maravx/par morave paradox"): cited concept — "hard things are easy, easy things are hard." - **"Jack"** (unidentified person, apparently a prior speaker at the same event): referenced as having earlier given a definition of continual learning "with like 10 different names." ## Provenance & Uncertainty Notes - Speaker's framing (intelligence vs. expertise, "world's smartest novice," "escape intelligence," "unbounded expertise from bounded intelligence") is presented as **speaker's own thesis/opinion**, not sourced fact. - All quantitative figures (Anthropic revenue) and quotes are **speaker's paraphrases**; original sources not cited in transcript. - Transcript is automated and contains apparent transcription errors (e.g., "IU" for Yu Su, "container learning" for continual learning, "Maravx" for Moravec, "Capathy" for Karpathy, "Satia" for Satya, "open clause" for OpenClaude/possibly "open Claude agents" — ambiguous). - **Comment availability:** Not specified — no metadata supplied. - **No timestamps** available in transcript. All right. Uh, I understand that I'm standing between you and the launch, so I'll try to be quick. Uh, my name is IU. I'm a professor at Ohio State, the Ohio State and uh I also have another job which is a CEO at a company called Neocognition and we focus on agents and container learning. So today's talk um it won't be too technical but I will it will be mainly a a conceptual one but I think it's a very important uh conceptual distinction that I will try to make between what is intelligence and what is expertise and through this I will try to answer some of the very bothering questions for me that um like why we are so successful at the coding agent but they're so terrible at anything else, right? Why uh the current agents are so token inefficient uh that to the degree that every company right now is like coming out and try to curb their uh their token maxing efforts in the company. Um so hopefully this will provide some food for thought before launch. Right. First a bit of a history. Um so AI agents are not a new thing, right? it's uh we have been trying to develop agents throughout the whole history of AI. Um but the problem is that in the early stages uh let's say um in the n in the 1960s to uh 80s when we developed these expert systems or logical agents or like in the uh 20110s when we developed these deep RL based uh neural agents we were only able to capture some very limited facets of human intelligence right whether it's like logical reasoning or it's like uh perception in single modalities to decision. Um only recently with the mult model LRMS and the language agent built on top of them for the first time we have a neural model that is able to encode multiensory inputs into a uh unified neural representation that is also conducive to symbolic reasoning and communication. Right? So that was a trait unique to humans. Now uh AI agents finally have the same thing. So that drastically improved their expressiveness, their reasoning ability and adaptivity. So that's why I think we have really entered a new evolutionary stage of machine intelligence. And uh it didn't take long for these language agents to find their first mass markets which is coding. And the best way to illustrate this is probably through the uh revenue graph of anthropics right in just under two years their revenue has grown 400 times uh to uh 40 billion I think the newest number is maybe 60 billion uh annualizer runway and it's largely driven by coding and coding related productivity uh capabilities but if we think about it right coding is the really the ideal market for this language anguage agents because code is already a language native world. Everything is already represented symbolically and like uh recorded uh in a very structured way and you get your rewards, you get your uh like tests all in place in symbolic ways. So then what happens when we leave the privileged world of code? Well, not so well. Um we are running into a lot of challenges deploying these agents in enterprise settings and the uh also in personal settings like the uh like open clause uh constantly make this uh like uh quite brittle and silly uh errors and then to the extent that Andrew Capathy uh said that it's not going to be the year of agents it's going to be the decade of agents because they cannot do computer use they don't have continuous learning um I don't know how much uh thought has changed uh since like uh last time because of the coding agent everything but I think the difficulties with computer use with container learning still largely the same right now so how can something be so small but also so brittle at the same time here's my thesis around it I think we are actually witnessing a modern version of the Maravx paradox, right? So the par morave paradox essentially says that for AI uh hard things are easy, easy things are hard. So the modern version here is that we are very good at this symbolic reasoning task like coding and math which were considered crown jewel of uh of intelligence uh earlier. But then we still struggle with this everyday digital work because they really require quite different set of cognitive competencies to excel as them and more specifically I think modern society is really not just one unified world it's millions of these micro worlds like every domain every every profession is different every company is different even if you're using the same software every company company configure it differently so is extremely idiosyncratic especially in the digital world. It has its unique local physics like different structures, constraints, affordances and dynamics that you have to learn. It's just like too hoggenous and dynamic for any uh monolithic model to try to compress it into one static representation. So agents must continually learn on the job to acquire what I call specialized expertise for each specific micro world. The second part of the talk I will try to establish the differences between intelligence and expertise. Here are the working definitions. For intelligence, it's the capacity to reason through unfamiliar problems from available context. Right? This is what the frontier models are increasingly good at. Um, you give it the provin statements, the context, the tools, and they can reason through this even if it's seeing them for the first time. Uh, and they can do a great job. Every episode is more or less like independent from each other here. But expertise is different. Expertise is really accumulated and situated competence. It's the ability to act reliably, efficiently and with judgment to achieve reproducibly super real uh performance in a particular domain. Right? So this is in stark contrast with intelligence. And to uh show what does expertise actually contain. I think uh the key idea from cognitive science is that experts don't just know more facts, we actually see the world differently, right? So um the uh expertise allows you to do different pattern recognition. So you see through the surface patterns like if you're looking an expert is looking at like a gigantic bug reports they can immediately locate like the most plausible places where things could go wrong. Um and you think about the problem with like a very deep structure, right? When you are scheduling a meeting, you know that it's not just like finding the shared slots on everyone's calendar is actually a constraint optimization problem over everyone's authority, the priorities, the urgency and everything. Um and we don't expert don't just operate with a set of rules, a set of facts. We know that every single thing is conditional, right? Every rule has like the preconditions where it applies. But we also know when we can bend the reality, we can bend the rules when exceptions happen, right? And finally that also give us judgment and taste is importantly what's high quality and uh very importantly when to stop when is good enough. Um so all this together I think experts effectively has have built a world model of their environments right that it's a generalized notion of world model that captures how that micro world works and that becomes the basis for all of our perception uh reasoning decision- making uh and judgment. So intelligence and expertise are really quite different across many dimensions. Uh but some of the import uh interesting ones here are like intelligence is about hey when we have the context uh how to solve the problems through the context but expertise actually will bring you the right context right given any problem we know what context bring into are important for this problem and bring it in to solve the problem and because of that uh intelligence tend to expand our search like every problem solving is a search problem. So, intelligence tend to brute force it, try to uh try to spin up like 100 different uh like parallel ways to to try to solve the problem. Well, expertise will actually try to compress the search space because expertise has constructed this has learned this essential shortcuts for the problem space so that whenever you have a problem, you know the most plausible ways to solve it. Um and then I also think the final part here is that I think continuous learning is the important bridge from intelligence to expertise. But first let me try to define continuous learning because it's such a confusing term um and and Jack just uh gives some definition earlier uh with like 10 different names. Um but here's my the definition I work with. I think continue learning is adaptive compression of the experience into reusable structures for future behavior. So all of these four elements here are very important. For experience, we need to uh answer the question like what kind of experience we're talking about is more like episodes of experience or it's like these semantic facts or procedures or feedback from human or environments. And how do we compress that? Uh so in we embed them into vectors or we index them into some symbolic structure. Uh we uh distill them into model parameters or do some kind of reinforcement learning. And it's not just like one time compression. It needs to be adaptive compression like what you have learned what you have compressed so far should largely uh influence how you compress further. And what kind of structure we're looking at? It's just like parameters like adapters of your language models or it's vectors, graphs, skills or even word models. And then how do you use these reusable structures? It's like you use it just to recall these facts or use it for prediction of like future states. is use for uh for better planning for or even for the control like actuation layer of the agent or as a value function for potential states right so it's because of this uh the continent learning problem is so rich like it has this four different aspects and if different aspects can be instantiated in different ways that makes this field so confusing but hopefully this is a definition that uh encompasses uh most of the versions of continuous learning. Then I think that uh this is maybe the most important figure in this talk. Um if we put raw intelligence as the x-axis and uh expertise as the yaxis I think we'll find that they are largely orthogonal to each other. If you don't have continuous learning uh all you do is scaling your model to to get better like raw intelligence then what we will get is what I call the world's smartest novice right super smart it can try to uh try to attack at any problem uh provide given to it but it doesn't accumulate expertise so it end up just like brute forcing its way at every problem then if you have continued learning uh like Different continuous learning algorithms will essentially set the slope of your learning uh curve here, right? If you have a sloppy CL algorithm, maybe some kind of simple incontext learning then uh with like uh increasing intelligence then your expertise will increase like a little bit but you have a really strong uh contin um increase like rapidly. Of course, this is assuming like a given time horizon and experience horizon. And then among all of these potential futures that good continued learning will bring us. I think this is pro probably the the one I like the most or I think it is the most interesting which I call the unbounded expertise from bounded intelligence. Right? What if we can uh come up with a computer learning algorithm such that um given up once the raw intelligence has crossed a certain threshold we don't need stronger intelligence anymore like learning will bring us like unbounded expertise once we have like a reasonable level of intelligence right then we can call this the escape intelligence and if this is indeed true then it will have a a lot of imp implications for the whole ecosystem, right? Do we need to continually train to train these larger and larger models or like this models like uh methods uh maybe they're already good enough? What we're missing is just like better continuous learning algorithms. So to be a little bit more concrete I think uh to provide more food for thoughts uh here are some open questions I think uh in this space the overarching question is like given any domain or environments right how can an agent continually learn to specialize and reach expert level competency but to do that you need to answer many other questions like how do you even measure uh define and measure expertise and this is probably environment specific and how to handle the trade-off between reliability and plasticity, right? Um we want these agents to be both reliable and plastic, but they are inherently conflicting with each other, right? Reliable systems or stable systems, they resist the change, but the plastic systems likes change. So how do we reconcile that? Um but fortunately we do have a living existence proof which is us ourselves humans uh that we are incredibly plastic but also manage to be dependable most of the time. Um then uh from a technical perspective like when we talk about learning largely there are like two forms of learning parametric or nonparametric. So how and my uh belief here is that both are really needed for uh this type of machine learning to to actually work. But how do we synergize the two and finally even though we are focusing on specialization I think there is a great potential for specialization to actually gener to lead to like better generalization. You know we are we have exhausted the public data for training LLMs but the next stage of training the next internet scale data opportunity is actually in all of these different uh like private worlds. If we can make this specialized agent work they can learn in sichu and uh channel back the learning to the general model I think that may be the next uh internet scale uh data opportunity. Okay. So finally a call to action um think let's start scaling expertise this will be a new dimension for us to scale because intelligence is already becoming abundance the frontier models they are probably smarter than average humans um but expertise is still scarce and we want to build a world where expertise becomes abundant where everyone can gets expert uh support because in an ideal world everyone can can have their personal health care, personal financial advisor and uh personal tutors and so on so forth and then every company can build their their own learning loop, right? They can, as Satia said two weeks ago, like we want to enable this human AI learning loop uh at each company that turns into institutional memory and for every company to build their uh own modes and to uh to still be in charge of their means of production. And finally um I think with abundance of expertise uh we will actually see more types of work become possible because there uh right now there are still a lot of opportunities that uh that are locked up because the friction is just so high to make them uh econ economically viable. But with abundance of expertise, I think that we will be able to lower the friction and uh make many of the new type of work across the threshold of worth doing. So this is the future we're building uh towards new cognition and happy uh to share this with you and uh thanks for the attention.