… authorities of the folks, by the folks, for the folks …
— Abraham Lincoln, Gettysburg Deal with (1863)
The price of AI is dropping quickly. GPT-4-class capabilities value roughly $30 per million tokens in early 2023; immediately the identical runs below $1, and some providers are pushing costs below $0.10. Throughout benchmarks, inference prices have fallen between 9x and 900x per year, with a median decline close to 50x. Even frontier models are getting dramatically cheaper every technology, with open-source fashions following carefully behind. And crucially, even when “Nobel-Prize-winning genius-level” intelligence isn’t right here but, the intelligence that suffices for the overwhelming majority of information work is right here immediately, and getting cheaper by the month. At this fee, we’re quickly getting into the period of just about free intelligence—the sort that’s greater than sufficient for on a regular basis data work.
Disclosure: This publish is a perspective led by Aditya G. Parameswaran—an Affiliate Professor of EECS and co-director of the EPIC Knowledge Lab at UC Berkeley—collectively along with his collaborators. It’s half panorama survey and half perspective, and several other of the analysis instructions mentioned beneath (together with agentic hypothesis, structured reminiscence, and synthesizing customized knowledge methods from scratch) draw on the authors’ personal ongoing work.
So, what does this new period of near-free intelligence imply for knowledge methods? We consider three new challenges—and alternatives—stem from near-zero inference prices:
Knowledge Programs For Brokers. Brokers will quickly turn into the dominant workload for knowledge methods—with swarms of brokers spun up in response to every end-user request. Given variations in traits between brokers and people—or functions performing on their behalf—how ought to we redesign knowledge methods for such agentic customers?
Knowledge Programs Of Brokers. As brokers begin taking up the majority of information work, a brand new substrate is required for 1000’s of brokers to handle state over long-running duties, coordinate and attain consensus, and cope with failures. What do knowledge methods that reliably and effectively run and handle agent swarms appear like?
Knowledge Programs By Brokers. Brokers are quickly changing into able to synthesizing total knowledge methods in a single go—which means we are able to rebuild customized methods for every new workload. Verifying that such methods match meant habits is a problem. What does it take to let brokers synthesize knowledge methods we are able to truly belief?
Knowledge Programs For, Of, and By Brokers
Subsequent, we’ll focus on every in additional element, adopted by discussing the intertwined future of information methods and brokers, particularly because the three challenges intersect.
Knowledge Programs For Brokers
An agent querying a database doesn’t behave like an individual or a BI device. It performs what we name agentic speculation: a high-volume, heterogeneous stream of labor spanning schema introspection, columnar exploration, partial after which full question formulation. With a number of brokers every exploring parts of the speculation house, every person request may quantity to 1000s of particular person SQL queries. Now, customers can concern ‘high-level’ knowledge duties, e.g., root-cause evaluation—e.g., ‘why did espresso gross sales in Berkeley drop this 12 months’—or exploratory cohort evaluation—e.g., ‘which person segments are most probably to churn subsequent quarter’—every involving a combinatorial house of potential joins, aggregations, and filter mixtures.
Knowledge Programs Redesigned to Extra Successfully Help Agentic Hypothesis
The requests from these brokers have varied alternatives for optimization. As an example, on a text-to-SQL benchmark with a number of brokers trying every activity, solely 10-20% of the sub-plans are distinct. Thus, 80-90% of sub-queries carry out duplicate work. The identical experiments present activity success charges considerably growing with extra agentic makes an attempt—so the redundancy is definitely useful. However from the info system perspective it’s wasted work.
An agent-first knowledge system can exploit such properties to assist brokers make progress quicker. It might probably reuse outcomes throughout overlapping sub-plans, drawing on concepts from decades-old literature on multi-query optimization and shared scans. Or the info system can attempt to satisfice, returning approximate solutions which are ok for brokers to make progress, leveraging work from the AQP literature—or streaming the outcomes of the ultimate or intermediate operators to assist brokers determine if seeing the remaining is critical or useful.
One other alternative right here is to rethink the question interface fully: as an alternative of brokers issuing a single SQL question at a time, they may as an alternative concern a batch of queries, every with its personal approximation necessities. Since enumerating an exponential search house (as within the root trigger or cohort evaluation examples above) isn’t a superb use of agentic reasoning skill, maybe knowledge methods ought to help higher-level primitives fairly than requiring brokers to checklist every SQL question explicitly. One concept right here is to attract on DBT-style Jinja macros to offer looping-based primitives for brokers to work together with knowledge methods.
A Caffeinated Military of Brokers Able to Tirelessly Full Your Knowledge Duties
A remaining alternative right here is to cease considering of information methods as passive executors of queries; knowledge methods may very well be proactive, as they possess extra grounding in knowledge and system traits that brokers could lack a priori—they may steer brokers in numerous instructions, present outcomes for associated queries, and likewise present performance-level suggestions (e.g., as an alternative of executing an costly question, the system may first present the agent a latency estimate). The rationale we are able to do that now versus the previous is that an agent can settle for any type of textual suggestions and isn’t anticipating a strict SQL question outcome. The truth is, the info system may additionally put together each materialized and digital views for an agent prematurely, supplied to the agent as a part of context, as this can be cheaper or more practical than having an agent writer or use them.
Knowledge Programs Of Brokers
Beforehand, we targeted on how brokers work together with knowledge methods. Now, we contemplate the whole lot else brokers have to maintain working: the place they reside, how they keep in mind, how they coordinate with one another, and the way they cope with failures of one another. This agentic substrate is separate from the inference stack powering uncooked intelligence. Nevertheless, the inference stack itself is being abstracted away via APIs (e.g., from OpenAI or Anthropic), or, for open-weight fashions, via serving frameworks that conceal low-level particulars. Up to now, the agentic substrate has been managed via harnesses like Claude Code and Codex, coupled with varied mechanisms to store and retrieve reminiscence.
First, on the reminiscence entrance, the present knowledge is that files are all you need; brokers write to unstructured markdown (MD) information, which may then be searched utilizing grep, or through embedding-based retrieval. The truth is, many argue that the answer to continuous studying is having brokers devour lots (e.g., a whole codebase, slack, firm wikis, …) after which write their learnings into MD information, that are then retrieved selectively on demand. Certainly, file methods, bash scripting, and MD information are and can nonetheless be essential for brokers. Nevertheless, at scale, when brokers are doing the overwhelming majority of information work, this method will not be efficient.
Given restricted context home windows, retrieving all MD file fragments which may be related and stuffing it into the context will break down in some unspecified time in the future. Even when context home windows proceed to develop, there are latency advantages to not put all data into context — and in lots of circumstances, e.g., when data work entails interacting with massive databases or code bases, it will likely be infeasible to serialize all related knowledge into context.
Knowledge Programs As A Substrate for Multi-Agent Swarms
One may use a knowledge graph representation, however data graphs undergo from the identical limitations as unstructured MD-based reminiscence because of their lack of structured search. What one wants is to have the ability to retrieve solely reminiscence that’s pertinent to the duty, throughout a number of attributes (or sides) of curiosity. For instance, an agent debugging a flaky take a look at ought to be capable of pull solely the recollections tagged with the related module, language, framework, and failure mode—fairly retrieving based mostly on key phrases or embedding similarity. A separate concern is what to really retrieve; uncooked agent traces with errors will not be very helpful as they may induce brokers to repeat the identical mistake—as an alternative, we would like the retrieved reminiscence to be corrective.
We not too long ago explored a associated notion of structured memory, the place we arrange reminiscence throughout varied attributes, every of which may very well be set as * to point common applicability, or set as an inventory of values to be matched. For a knowledge agent, the scale may embody the columns and tables, sort of operation, and at last, open-ended natural-language corrective directions. So, we may embody reminiscence that solely applies to a given sort of operation (e.g., ‘when performing date-time operations, use fiscal 12 months versus calendar 12 months conventions’), or a given desk (e.g., ‘column product_cleaned is most popular over column product when querying on product title’). One open query is defining an application-specific structured reminiscence—or what others have known as world models for memory. We consider that is akin to defining a schema for every software—and maybe brokers themselves may help us outline and refine it over time.
One Doable Method To Retailer and Retrieve Structured Data [From Here]
Structured reminiscence can be helpful additionally for evolutionary frameworks to successfully handle search areas. Certainly, storing, structuring, and mining massive volumes of single and multi-agent traces may help future brokers turn into rather more environment friendly—probably enabling efficient recursive self-improvement via structured memory-based mechanisms.
One other problem is to help concurrent edits to shared reminiscence, and concurrent edits typically, when there are numerous brokers performing transformations. Whereas there have been some helpful makes an attempt at supporting multiversioning and tech/agentfs/introduction”>copy-on-write semantics, it isn’t clear that such methods will suffice when 1000’s of brokers try to edit shared state on the similar time. As an example, when brokers are attempting varied potential transactions in response to a person request, the consequences of the overwhelming majority of those transactions must be rolled again—with solely the one ‘appropriate’ transaction’s outcome persisting. Work on supporting exactly-once semantics is related right here, as are underlying methods based mostly on CRDTs and operational transformation. For updates to fuzzy mechanisms corresponding to reminiscence, we might be able to sacrifice on consistency for excellent correctness within the curiosity of latency. Whereas brokers can cause about semantics to compensate or roll again their actions to finally finalize most duties, the first problem lies within the diploma to which they step on one another’s toes through the course of. An essential failure mode to be prevented is a type of “livelock,” the place incessant compensating actions forestall any significant progress.
Past shared state, different considerations emerge when attempting to help a military of brokers, together with what to do when brokers fail, how brokers ought to talk with one another (instantly or via intermediate shared state), and the way we should always cope with straggler brokers. There have been some developments in supporting sturdy multi-agent execution, corresponding to Temporal, nevertheless it stays to be seen if such options will apply at scale throughout 1000’s of brokers. On the subject of communication, we’d like mechanisms to allow brokers to barter with one another. Think about 4 developer brokers trying to succeed in consensus on a shared schema, with distinct however overlapping aims. In a human setting, this may contain iterative dialogue and compromise; for agentic swarms, we should outline the mechanisms that permit them to converge on a design that displays the underlying objectives of their respective principals. Or if brokers are all requiring entry to a restricted useful resource, once more communication can be mandatory. It stays to be seen if that is finest carried out through centralized coordination, or if a decentralized method is critical.
Knowledge Programs By Brokers
Lastly, if intelligence is successfully free, then we are able to make use of this intelligence to synthesize new knowledge methods from scratch. Certainly, in lots of settings, general-purpose knowledge methods could also be overkill, as they must help each schema, question, and {hardware} goal. Given a workload, latest work, together with Bespoke OLAP and GenDB, has proven that one can use an agentic pipeline to synthesize a whole, workload-specific analytical engine—in minutes to some hours, at a price of some {dollars}. The engines are disposable: when the workload shifts, one can merely regenerate them. Analogously, our work has proven that one can synthesize customized key-value stores from scratch, focused to the workload. The truth is, trendy IDEs, corresponding to Kiro, elevate specs for methods growth to be a first-class citizen.
Brokers Can Synthesize Customized Knowledge Programs From Scratch
The primary concern, nevertheless, is that specs are sometimes imperfect, and don’t cowl all nook circumstances. Current-day brokers will exploit the lacking specs to reward-hack their solution to a excessive efficiency metric. In our customized key-value retailer work, we discovered that one solution to alleviate that is to have auxiliary verification brokers attempting to generate take a look at circumstances that catch the exploitation of nook circumstances, primarily increasing the specification. One more method is to each generate a system and a proof for its correctness collectively, for which we now have discovered some early success, however extra must be carried out to solidify the method. Additional, it stays to be seen what’s one of the best ways to solicit human-written specs for a system—can this be carried out in an iterative, human-in-the-loop method, versus a one-shot, incomplete one. Certainly, human-written specs are incomplete even for manually authored software program, so one would count on that future brokers which are extra aligned will more and more train higher judgement when making design selections.
One Doable Knowledge System Synthesis Pipeline [From Here]
Different questions right here contain testing whether or not ranging from a mature system (e.g., Postgres) and eradicating elements/performance can result in greater efficiency or extra person belief. Individually, is there a possibility to make the design composable, comprising varied verified elements which are blended and matched given a workload? For instance, maybe the workload hasn’t modified sufficient for the storage layer to be up to date, however maybe the question optimizer requires adjustments. A maybe extra viable proposition entails using brokers coupled with proof methods to focus on vital components of the code related to formal proofs, fairly than doing so for your complete system.
A remaining alternative right here is to maneuver away from the normal knowledge methods stack with clearly-defined interfaces (e.g., parser, question optimizer, storage supervisor, …) — that have been every largely the prerogative of a single human group to handle. As an alternative, brokers can discover new methods to “mix” these elements collectively, maybe figuring out new optimization alternatives in consequence. Brokers may fill in lacking gaps in performance to make present methods rather more feature-complete, or attain feature-parity with different competing methods—or analogously, constantly refining open-source methods in response to function requests or points (maybe filed by different brokers!) Doing so in a approach that prioritizes correctness, long-term upkeep, and human interpretability can be a problem.
Trying Additional Forward
Within the period of near-free intelligence, knowledge methods matter greater than ever. As brokers tackle the majority of information work, the workload for knowledge methods will change, the substrate they should run on must be constructed, and more and more, they may take part in designing knowledge methods themselves. Every of those shifts opens up a brand new, thrilling analysis agenda.
Co-Evolution of Knowledge Programs and Brokers
Trying additional out, the boundaries between brokers and knowledge methods will probably begin to blur. As an example, brokers could design the info methods they themselves run on, defining each the interfaces in addition to the system elements beneath. Each the interfaces and internals may be advanced over time by brokers in a type of recursive self-improvement. There may be additionally a possibility to rethink knowledge methods as a holistic supply of fact for the whole lot of related state: together with uncooked knowledge, reminiscence, and coordination state, additional erasing the distinctions between the info that’s being queried by brokers and knowledge generated because of agentic exercise. Lastly, knowledge methods could themselves incorporate agentic elements, basically evolving from passive computation engines into clever, proactive, self-optimizing architectures. It’s arduous to foretell what the long run could maintain. We’re in for a wild experience!
Acknowledgments
The attitude and ongoing work described on this publish are the product of joint analysis and plenty of discussions with great collaborators on the EPIC Data Lab, Data Systems & Foundations group, and the broader Berkeley AI-Programs group. Thanks all!
BibTex for this publish:
@misc{intelligence-is-free-blog,
title={Intelligence is Free, Now What? Knowledge Programs for, of, and by Brokers},
writer={Aditya G. Parameswaran and Shubham Agarwal and Kerem Akillioglu and Shreya Shankar
and Sepanta Zeighami and Rishabh Iyer and Matei Zaharia and Alvin Cheung
and Natacha Crooks and Joseph Gonzalez and Joseph Hellerstein and Ion Stoica},
howpublished={url{https://bair.berkeley.edu/weblog/2026/07/07/intelligence-is-free-now-what/}},
12 months={2026}
}
