# Why Agent Engineering

swyx (Shawn Wang)

AI Engineer Summit 2025

[Talk page](https://ai.engineer/talks/5N33E9tC400-why-agent-engineering) · [Video](https://www.youtube.com/watch?v=5N33E9tC400)

## Why Agent Engineering

As models improve and inference gets cheaper, AI engineering is moving toward a more specific deliverable: agents that do useful work for users.

### What does an AI engineer build?

[0:34](https://www.youtube.com/watch?v=5N33E9tC400&t=34s)

How is AI engineering doing? There is now an O’Reilly book: Chip Huyen’s *AI Engineering*. There is also a hype-cycle chart placing the field near the peak—a setup for swyx’s joke that it can only go downhill from here. Recognition and inflated expectations have arrived together.

![Slide showing Chip Huyen holding the AI Engineering book beside a post with a hype-cycle chart marking AI Engineering near the peak.](https://ai.engineer/api/talk-essay-media/68a5e2902d8cbe871e3a6e5674c77fb8c084a1ca4194a2432ce33ddac403e38f)

How is AI Engineering doing? A book launch alongside a hype-cycle chart.

The earlier task was to establish that the discipline existed. [The Rise of the AI Engineer](https://www.latent.space/p/ai-engineer) named the emerging role; the first AI Engineer Summit explored three types of AI engineer; the following World’s Fair examined its maturation across disciplines. Those frameworks helped people organize teams, but a label that has gained acceptance still needs a distinct account of the work.

Two neighboring professions offer competing descriptions. From machine learning engineering, AI engineering can look like machine learning plus a few prompts. From software engineering, it can look like software plus a few LLM API calls. Neither description leaves much room for an independent discipline. Swyx’s own current heuristic is 90% software engineering and 10% AI, but he expects the AI portion to grow as the work develops its own practices.

Professional language already reveals different priorities. A machine learning engineer might say **test-time compute**, while an AI engineer says **inference-time compute** because inference is the product’s everyday work. A software engineer might simply call it reasoning. This is an anthropological view of engineering: people describe their work, form groups around shared concerns, and eventually establish industries. Agent engineering is the next proposed focus for that process.

### From frameworks to production

[2:48](https://www.youtube.com/watch?v=5N33E9tC400&t=168s)

Making the Summit an agent engineering conference required excluding RAG, open models, and GPUs as standalone topics. That was a choice about the program’s scope, not a rejection of those technologies. Narrowing the subject opened space for more specific kinds of agent work, and the previous year’s strongest-performing YouTube talks suggested that the audience wanted that focus.

Audience interest alone did not solve the selection problem. The available speakers were largely people who built agent frameworks for a living, while the unanswered question was who actually used agents in production. The new rule was no vendor pitches. Recruiting practitioners made curation harder: people running systems have less incentive to appear onstage than people selling the tools used to build them. Their experience was precisely what the conference needed.

**Agents can extend existing application categories.** Agent plus RAG, agent plus code generation, and agent plus search were swyx’s compact formula for commercial opportunity in 2025. The accompanying slide expands the combinations to vision, voice, wearables, finance, and money. An agent focus therefore need not isolate the field from the rest of AI engineering; it changes what builders try to accomplish with those components.

![Slide titled “have you tried rubbing an agent on it?” lists eight agent combinations and examples including Agentic RAG, Deep Research, Computer Use Agent, Bee, and Solana/Stripe.](https://ai.engineer/api/talk-essay-media/c2b517ff086afb7b41be4697000ae53f78b102adc1cdf62f1d97a9aec84cb8f8)

Agents combined with RAG, code generation, search, vision, voice, wearables, finance, and money.

### Reconsidering the year of agents

[4:18](https://www.youtube.com/watch?v=5N33E9tC400&t=258s)

Calling 2025 the year of agents does not make it so. Predictions often confuse what their authors want with what will happen. Swyx names Satya Nadella, Roman, Greg Brockman, and Sam Altman as advocates of that expectation, while acknowledging that he and his Latent.Space co-host Alessio had been skeptical themselves.

The audience had its own fatigue: in the previous day’s Family Feud exercise, agents ranked second among buzzwords people were tired of hearing. In a March 2024 conversation with David Luan, swyx and Alessio had advised people to remove agents from their branding. Now swyx was advising them to put it back. The reversal sets up two questions: what counts as an agent, and what has changed enough to make one useful?

### Several definitions, no settled boundary

[5:20](https://www.youtube.com/watch?v=5N33E9tC400&t=320s)

The definition depends partly on where the engineer starts. Machine learning brings reinforcement learning environments, actions, and goals. Software engineering offers the deliberately reductive joke: put it in a `for` loop. AI engineering has not yet settled on its own preferred boundary. These perspectives emphasize different parts of a system, even when they describe the same application.

Simon Willison’s crowdsourcing gathered hundreds of responses about what an agent is. Swyx organizes the recurring definitions around six properties:

| Emphasis | Defining concern |
| --- | --- |
| Goals | Autonomous pursuit of an objective |
| Tools | An LLM using tools |
| Control flow | An LLM directing what happens next |
| Long-running processes | Work that persists over time |
| Delegated authority | Permission to act on someone’s behalf |
| Multistep completion | Carrying a task through several steps |

These are different criteria, not a single checklist that every definition requires.

The full slide also includes metaphors, reductive descriptions, and specific examples. That last approach offers a practical test: collect systems people broadly agree are agents, then check whether a proposed definition includes them. It grounds a terminology debate in recognizable cases.

![Three-column slide beneath a Simon Willison post lists autonomous goal-oriented systems, tool-using LLMs, LLM-driven control flow, long-running processes, delegated authority, multi-step task completion, metaphors, reductiveness, and specific examples.](https://ai.engineer/api/talk-essay-media/648d9f01eb62336bac065662c6fa59cab4d054fdaf428c0235b6bef1337227a5)

Nine ways of describing agents, from goal-oriented systems to specific examples.

Definitions also matter because they shape platforms. OpenAI had introduced another agent definition in the previous day’s presentation, and swyx urges attention to it because the company will build around that understanding. The significance here is the connection between a definition and the systems it encourages, rather than a final resolution of the terminology.

### Why agents may work now

[7:08](https://www.youtube.com/watch?v=5N33E9tC400&t=428s)

Why should agents work now when they struggled a year or two earlier? The first answer is capability. Swyx presents trajectories from 2023 to 2025 approaching human baselines. Without named tasks and evaluation conditions, that is a directional argument about improving models, not a claim that agents have reached human performance across arbitrary work.

He separates three familiar contributors:

- **Better reasoning:** stronger model capability for working through a task.
- **Better tool use:** improved ability to use external capabilities.
- **Better tools:** a stronger ecosystem for connecting models to useful functions and information, including the [Model Context Protocol](https://www.anthropic.com/news/model-context-protocol), or MCP.

The distinction matters: a model’s ability to use a tool and the availability of a useful tool are separate parts of the engineering problem.

Model diversity adds another source of opportunity. Swyx estimates that OpenAI’s share fell from roughly 95% two years earlier to 50%, without specifying the market denominator. He also points to two possible frontier-lab challengers emerging in the preceding week. The argument is that builders face a more varied competitive landscape; which providers will prevail remains uncertain.

### Cheaper intelligence, more operating options

[8:09](https://www.youtube.com/watch?v=5N33E9tC400&t=489s)

Capability is only part of viability; the cost of using it matters too. Swyx calls the decline in intelligence costs “super Moore’s law.” He claims that the cost of **GPT-4**-level intelligence fell 1,000-fold over 18 months, without specifying the benchmark or pricing basis. He sees a similar curve beginning for **o1**-level intelligence. The proposed enabling change is that an application can afford intelligence that would previously have been too expensive.

Reinforcement learning fine-tuning introduces another option for builders. Swyx explicitly says he has no personal experience with it and points to Will’s later presentation for a practitioner’s account. It belongs in the list of emerging capabilities, rather than as a technique demonstrated in this keynote.

The remaining changes concern both delivery and economics. A conversation with Brett Taylor raised charging for outcomes instead of costs, shifting the commercial unit toward the result a customer wants. Work on multiple agents expands how systems can organize tasks, while better hardware enables faster inference. Swyx points to Latent.Space conversations for the supporting discussions; here, these are additional contributors to agent viability, not detailed implementation recipes.

### Useful delegation and unwanted delegation

[9:02](https://www.youtube.com/watch?v=5N33E9tC400&t=542s)

Where does that combination already create value? Referring to Barry Zhang’s [Building effective agents](https://www.anthropic.com/engineering/building-effective-agents), swyx identifies coding and support agents as having product-market fit. He adds deep research to that category and anticipates further use cases in the conference program. These are his assessments of adoption and usefulness, not quantified results for every product in those categories.

His anti-use cases are more personal: stop repeating flight-booking agent demos. He wants to choose his own flights and place his own Instacart orders. That objection exposes a limit on indiscriminate delegation: a task can be technically automatable without being something a user wants to hand over. He also explicitly objects to astroturfing. The opportunity is to find valuable work users want agents to do, not merely another action an agent can perform.

### Adoption evidence and a growth forecast

[9:38](https://www.youtube.com/watch?v=5N33E9tC400&t=578s)

The final example is ChatGPT’s growth. Swyx brings in a newly published user milestone and describes asking deep research to research OpenAI and draw the growth chart. [Reuters reported](https://www.investing.com/news/stock-market-news/openais-weekly-active-users-surpass-400-million-3881021) more than 400 million weekly active users in February 2025, compared with 300 million in December 2024—approximately 33% growth over roughly two months. That is a shorter interval than the three months described onstage. Swyx frames the larger trajectory as going from zero to that scale in approximately two and a half years.

He reads the chart as a history of model-driven demand: a period he characterizes as a year without growth, followed by acceleration when agentic models arrived. Swyx credits the o1 models with doubling ChatGPT usage. That is his causal interpretation of the adoption curve; the reported user counts do not establish which releases caused the increase.

Extending the curve, swyx predicts one billion ChatGPT weekly active users by the end of 2025, five times the September 2024 level. The slide makes that extrapolation explicit with a December 2025 billion-user callout alongside model-release annotations. It is a forecast made in the talk, not an observed milestone.

![Slide titled “1B users by 2026 from Reasoning/Agents” shows a rising blue curve, model-release annotations, a “you are here” marker, and a “1B ChatGPT Users Dec 2025” callout.](https://ai.engineer/api/talk-essay-media/ddf3137a5bed73bbf4e978c31f0222754874860255ff5bd23c69b87f71df9cf6)

A projected path to one billion ChatGPT weekly active users, annotated with model releases.

The product thesis follows from that interpretation: stronger reasoning and more agents shipped to users can drive demand for AI products. Swyx frames the projected billion users as roughly one-eighth of the world’s population and sees substantial commercial room for other builders. The forecast supplies the scale of his ambition; the engineering question is which capabilities will become useful products.

### The AI engineer’s evolving deliverable

[10:56](https://www.youtube.com/watch?v=5N33E9tC400&t=656s)

**Machine learning engineers build models; software engineers build software; AI engineers increasingly build agents.** This is swyx’s proposed direction for the profession, not an exclusive boundary between jobs. It gives the emerging discipline a concrete deliverable: systems that put model capabilities to work on tasks users are willing to delegate.

## Resources

- [Agent Engineering](https://www.latent.space/p/agent) — Swyx’s expanded companion to the keynote, covering agent definitions, enabling trends and conference examples.
- [The Rise of the AI Engineer](https://www.latent.space/p/ai-engineer) — The original essay proposing AI engineering as a discipline built around applying foundation models.
- [Building effective agents](https://www.anthropic.com/engineering/building-effective-agents) — Anthropic’s guide to workflows, autonomous agents, tool design and practical coding and support applications.
- [Introducing the Model Context Protocol](https://www.anthropic.com/news/model-context-protocol) — The original MCP announcement explains how clients and servers connect AI applications to external tools and data.
- [OpenAI’s weekly active users surpass 400 million](https://www.investing.com/news/stock-market-news/openais-weekly-active-users-surpass-400-million-3881021) — Contemporaneous reporting on the February 2025 weekly-user milestone and its December baseline.

## Slides

### How is AI Engineering doing?

Chip Huyen’s AI Engineering book signals a growing field, while Gartner’s June 2024 hype cycle places AI Engineering near the peak. The speaker jokes that it’s all downhill from here.

[0:48](https://www.youtube.com/watch?v=5N33E9tC400&t=48s) [A post about the AI Engineering book appears beside a Gartner hype-cycle chart placing AI engineering near the peak.](https://assets.ai.engineer/sites/aiecode2025/sha256/35a1dab4d58298e112e0bc1e80d4d57badec8ab5557f4abf4dce05b48d1cd1dd.webp)

How is AI Engineering doing?

Chip Huyen @chipro · Jan 9
Finally got my copy! “AI Engineering” is officially out 🙏 🎉

AI Engineering
Building Applications with Foundation Models
Chip Huyen

TIL gartner, having missed the entire wave in ai engineering, is now calling the top in ai engineering

pack it up folks its entirely over only downhill from here

Time To Plateau Will Be Reached: < 2 yrs. | 2–5 yrs. | 5–10 yrs. | > 10 yrs.

AI Engineering
Responsible AI
Prompt Engineering
AI TRiSM
Sovereign AI
Artificial General Intelligence
Composite AI
Neuro-Symbolic AI
Decision Intelligence
AI-Ready Data
Causal AI
AI Simulation
Multiagent Systems
Embodied AI
First-Principles AI
Quantum AI
Autonomic Systems
Edge AI
Foundation Models
Synthetic Data
ModelOps
Generative AI
Neuromorphic Computing
Smart Robots
Cloud AI Services
Intelligent Applications
Knowledge Graphs
Autonomous Vehicles
Computer Vision

Y-axis: EXPECTATIONS
X-axis: TIME
Stages: Innovation Trigger | Peak of Inflated Expectations | Trough of Disillusionment | Slope of Enlightenment | Plateau of Productivity
As of June 2024

Interface text:
224
461
6.4K
408K

### The Evolution of the AI Engineer (old)

The original diagram places AI engineers between model research and product engineering, with a permeable API line. It assigns them product-specific data and evals; the speaker now considers the framing stale.

[1:27](https://www.youtube.com/watch?v=5N33E9tC400&t=87s) [Diagram places AI engineering between research and full-stack engineering, noting that the API boundary is permeable and product-specific data and evaluations matter.](https://assets.ai.engineer/sites/aiecode2025/sha256/23fd69af79b9bb3967f1148afc7dab4a0e1c6c9111be5e3d40d82e7bd117f2c6.webp)

Data/Research constrained
API
Product/User constrained

ML Researcher / Research Scientist
- training
- evals

ML Engineer/ Data Scientist/ Research Engineer
- inference
- data

AI Engineer
- chains/agents
- tooling & infra

Fullstack Engineer
- product
- platform

Important: the API line is permeable - AI Engineers can go left to finetune/host models and Research Engineers go right to build atop APIs too! This diagram has also been criticized for the placement of evals and data; we certainly agree evals are an important part of the job! MLR/MLEs handle foundation model concerns - aka pretrain scale data and general benchmark evals; but AI Engineers should certainly view product-specific data and evals as their job.

### The Evolution of the AI Engineer (now)

Two nearly overlapping circle pairs show competing perspectives: AI engineering as mostly MLE plus a few prompts, or mostly software engineering plus a few LLM API calls.

[1:34](https://www.youtube.com/watch?v=5N33E9tC400&t=94s) [MLE and software engineering perspectives shown as pairs of largely overlapping circles.](https://assets.ai.engineer/sites/aiecode2025/sha256/243bd0fc2786225b3032d3b0998a44a59a4c9134c98a936fef968c4e9ab85dbe.webp)

The Evolution of the AI Engineer (now)

MLE POV
SWE POV

### The Evolution of the AI Engineer (ideal)

The speaker contrasts the language of MLEs, AI engineers, and software engineers: test-time compute, inference-time compute, and reasoning. He ties the AI engineer’s term to a focus on inference.

[2:15](https://www.youtube.com/watch?v=5N33E9tC400&t=135s) [Columns labeled MLE, AIE, and SWE, with test-time compute shown under MLE.](https://assets.ai.engineer/sites/aiecode2025/sha256/e50c3843c586f7dddab0aec35768ecd1d47e545f84db8d6f58d63fffd8d2fd50.webp)

The Evolution of the AI Engineer (ideal)

MLE
AIE
SWE

MLE: test time compute

[2:29](https://www.youtube.com/watch?v=5N33E9tC400&t=149s) [MLE, AIE and SWE are associated with test-time compute, inference-time compute and reasoning, respectively.](https://assets.ai.engineer/sites/aiecode2025/sha256/6976ce5bf2e3e3a08465b4fd00a17dbd6c0044fb5646db6cbed84f0e67c3b130.webp)

The Evolution of the AI Engineer (ideal)

MLE: test time compute
AIE: inference time compute
SWE: reasoning

Interface text:
AIE
AGENT ENGINEERING
HTTPS://AI.ENGINEER

### Why We Are All-In on Agents

The AI Engineer Summit pivots to agent engineering, narrowing its focus from a broader AI engineering program.

[2:42](https://www.youtube.com/watch?v=5N33E9tC400&t=162s) [A slide about focusing on agents alongside an AI Engineer Summit announcement.](https://assets.ai.engineer/sites/aiecode2025/sha256/e91b23f9f7bfa183f5d8f1a63cc145ae71803f81a2814b606be34dbe6f56e68a.webp)

why we are
all-in on agents

Latent Space

Announcing AI Engineer Summit NYC: All in on Agent Engineering + Leadership

Announcing the theme of the second ever AI Engineer Summit (Feb 19-22 in NYC!), and releasing talks from the World's Fair AI Engineering Leadership track. Apply now!

BENJAMIN DUNPHY AND SWYX & ALESSIO
JAN 02, 2025

We’ll keep this short and sweet:
• Our first invite-only AI Engineer Summit was sold out, 10:1 applicant ratio
• Our multitrack AI Engineer World’s Fair sold out 3k seats, got >1m views

World’s Fair
OpenAI
aws
Google Cloud
MongoDB
neo4j
Microsoft
ANTHROPIC
Google DeepMind

still amazed at how we got all the top names in AI eng in one room

• The invite-only Summit is back! and

Interface text:
27
5
3
Share

### The Evolution of the AI Engineer Conference

The conference pivots from a broad slate of tracks, including RAG, open models and GPUs, to focus on agent engineering. The earlier agenda highlights AI Agent Applications as one of eight tracks.

[2:51](https://www.youtube.com/watch?v=5N33E9tC400&t=171s) [Two conference-day track lists, with AI Agent Applications highlighted.](https://assets.ai.engineer/sites/aiecode2025/sha256/c8e5d175192bc7662d41922db8de12e8f0d048a7a5962242c990dd1d83aac7fc.webp)

The Evolution of the AI Engineer conf

JUN 26 – CONFERENCE DAY 1
01 RAG (LLM Frameworks and Vector Databases)
02 Open Models
03 CodeGen and Dev Tools
04 AI in the Fortune 500

JUN 27 – CONFERENCE DAY 2
05 Multimodality and Omnimodality
06 GPUs and Inference
07 Evals & LLM Ops
08 AI Agent Applications

### How We Chose Your Speakers

The call for speakers turned an agent-focused day into a broad menu of disciplines, from memory and evaluations to finance and agent architecture.

[3:03](https://www.youtube.com/watch?v=5N33E9tC400&t=183s) [A list of agent engineering topics, including memory, evaluations, coding, planning, safety and agent architecture.](https://assets.ai.engineer/sites/aiecode2025/sha256/134cf83371983b2d677372bfb963908ea725484d8e182f5c759eb1c77e8482f8.webp)

how we chose your speakers

Day 2 will be the day we focus the conference entirely on Agents. Many are saying “2025 will be the year of Agents”, but significant gaps remain between expectation and reality, and we will gather the world’s best agent builders to share the state of the art, commiserate over how much further we have to go, and to make concrete plans and connections to fix those gaps.

Our request for talks include these agent-oriented interest areas:
• Memory
• Personality
• Agent Evals
• Voice Agents
• LLMs for Agents
• Proactive Agents
• Coding/CodeGen
• Planning & Reasoning
• Multi-Agent Architectures
• Guardrails/Safety/Privacy for Agents
• Computer Use/Model Context Protocol Agents
• AI/UX for Agents aka Human-Agent Interaction design
• AI Agents in Finance (we will have 50% banking/hedge fund/RIA audience)
• Agent Architecture aka How To Make A SOTA Agent (demonstrate why it is SOTA)

“I made it up”

AIE
AGENT ENGINEERING
HTTPS://AI.ENGINEER

[3:08](https://www.youtube.com/watch?v=5N33E9tC400&t=188s) [Slide listing agent engineering topics from the conference call for speakers.](https://assets.ai.engineer/sites/aiecode2025/sha256/6d68cf997c6c7703bdbb94d2cefbbe2120d535d73646d5505a43e6f2567ddc9c.webp)

how we chose your speakers

Day 2 will be the day we focus the conference entirely on Agents. Many are saying “2025 will be the year of Agents”, but significant gaps remain between expectation and reality, and we will gather the world’s best agent builders to share the state of the art, commiserate over how much further we have to go, and to make concrete plans and connections to fix those gaps.

Our request for talks include these agent-oriented interest areas:
• Memory
• Personality
• Agent Evals
• Voice Agents
• LLMs for Agents
• Proactive Agents
• Coding/CodeGen
• Planning & Reasoning
• Multi-Agent Architectures
• Guardrails/Safety/Privacy for Agents
• Computer Use/Model Context Protocol Agents
• AI/UX for Agents aka Human-Agent Interaction design
• AI Agents in Finance (we will have 50% banking/hedge fund/RIA audience)
• Agent Architecture aka How To Make A SOTA Agent (demonstrate why it is SOTA)

“I made it up”

AIE
AGENT ENGINEERING
HTTPS://AI.ENGINEER

### Just make everything agentic?

Last year's top-performing YouTube talks showed demand for agentic topics, but the speakers largely built agent frameworks. The question became: who's putting them in production?

[3:16](https://www.youtube.com/watch?v=5N33E9tC400&t=196s) [Slide showing video thumbnails for talks on agentic RAG and building an agent company.](https://assets.ai.engineer/sites/aiecode2025/sha256/114383e3b54b1136b7d6185e6bc7017495595de21f8bc9d5b46dfbfbe5a260a4.webp)

just make everything agentic?

AI Engineer World's Fair
Agentic RAG & LLAMA AGENTS
Jerry Liu | LlamaIndex
16:55
The Future of Knowledge Assistants: Jerry Liu
[illegible] months ago

Agents
BUILDING AN Agent Company
João Moura | crewAI
13:46
Using agents to build an agent company: Joao Moura
35K views • 6 months ago

AIE
AGENT ENGINEERING
HTTPS://AI.ENGINEER

### Have you tried rubbing an agent on it?

Swyx jokes that adding agents to existing categories is a formula for making money in 2025, pairing applications from RAG and search to finance and payments with named examples.

[3:59](https://www.youtube.com/watch?v=5N33E9tC400&t=239s) [Slide pairs agents with applications including RAG, coding, search, vision, voice, wearables, finance and money.](https://assets.ai.engineer/sites/aiecode2025/sha256/5ae72dc4f5f49285fd24fa5ed010ad868e3cf50e304e2186cefe94482a31dcbe.webp)

have you tried rubbing an agent on it?

Agent + RAG = Agentic RAG!
Agent + CodeGen = Bolt! Windsurf Cascade!
Agent + Search = Deep Research!
Agent + Vision = Computer Use Agent!
Agent + Voice = SuperDial! Method Financial!
Agent + Wearable = Bee!
Agent + Finance = BrightWave/Bloomberg/BlackRock!
Agent + Money = Solana/Stripe!

### 2025 is the year of agents

The speaker questions the familiar prediction: people often confuse what they want to happen with what will actually happen.

[4:17](https://www.youtube.com/watch?v=5N33E9tC400&t=257s) [Slide displaying the phrase “2025 is the year of agents” in quotation marks.](https://assets.ai.engineer/sites/aiecode2025/sha256/fa4fac092e8b90d381f71a8bd83554429954acbe3dad4e802d26c3dc1dd5f667.webp)

“2025 is the year of agents”

AIE
AGENT ENGINEERING
HTTPS://AI.ENGINEER

### 2025: The year of agents

A Satya Nadella video and posts from Romain Huet, Greg Brockman, and Sam Altman tout 2025 as a turning point for agents. Swyx says he and Alessio were skeptical.

[4:36](https://www.youtube.com/watch?v=5N33E9tC400&t=276s) [A collage of a video listing and social posts discussing AI agents in 2025.](https://assets.ai.engineer/sites/aiecode2025/sha256/291e7ef900e4f433dd3bb6f877faa4c05540d97b2d837b096533f4ca8a0ceaaa.webp)

Satya Nadella on the Future of SaaS, How 2025 is the year of Agents, Advice for Indian Engineers

Thumbnail: Satya Nadella on Is SaaS DEAD?

Romain Huet @romainhuet · 2h
2025 is the year of agents! Check out this demo of @OpenAI Operator—an agent that uses its own browser to handle tasks for you, currently in research preview. It sees webpages and navigates by typing, clicking, and scrolling. youtube.com/watch?v=CSE77w...

Greg Brockman @gdb · 3h
Operator — research preview of an agent that can use its own browser to perform tasks for you.

2025 is the year of agents.

Sam Altman @sama
reflections:

We are now confident we know how to build AGI as we have traditionally understood it. We believe that, in 2025, we may see the first AI agents “join the workforce” and materially change the output of companies. We continue to believe that iteratively putting great tools in the hands of people leads to great, broadly-distributed outcomes.

Interface text:
687K views · 2 weeks ago
Varun Mayya
Just had an incredible conversation with Satya Nadella! From the future of SaaS to AI agents reshaping b...
4K
6 chapters Is SaaS Dead? | India's Competitiveness | Career Advice For 25-Year-Ol...
Satya Nadella On Evolution Of SaaS
6   2   35   1.7K

### Define agent

Swyx challenges the audience to define “agent,” calling it a monumental task he can tackle in one slide.

[5:19](https://www.youtube.com/watch?v=5N33E9tC400&t=319s) [A dark slide reads, “go ahead, define agent.”](https://assets.ai.engineer/sites/aiecode2025/sha256/9a61d6333ddb2fcffcbc0e4d0a0a81263c58181400997125d869e31d36ba22b1.webp)

go ahead
define “agent”

SlidesCarnival
August 19, 2024

### Definitions of Agent Based on POV

The machine learning view defines agents by perception and goal-directed action; the software engineering view reduces them to LLM calls in a for loop. The AI engineering definition is left open.

[5:46](https://www.youtube.com/watch?v=5N33E9tC400&t=346s) [The slide contrasts a machine learning definition with an unknown AIE definition and a software engineering description of LLM calls in a loop.](https://assets.ai.engineer/sites/aiecode2025/sha256/795841f71274668b09a335d1277103d2386685887c90db7fd59b5066ba52c7b4.webp)

Definitions of Agent based on POV

MLE
“A system that perceives its environment through sensors and acts upon that environment through actuators to achieve specific goals.”

AIE
???

SWE
“LLM calls in a for loop”

AIE
AGENT ENGINEERING
HTTPS://AI.ENGINEER

[5:55](https://www.youtube.com/watch?v=5N33E9tC400&t=355s) [Slide comparing machine learning, AI engineering, and software engineering perspectives on agents.](https://assets.ai.engineer/sites/aiecode2025/sha256/4a90e7b78cb03d046a81578c23b14640dd694871448f4a261edcf90323a577da.webp)

Definitions of Agent based on POV

MLE
“A system that perceives its environment through sensors and acts upon that environment through actuators to achieve specific goals.”

AIE
???

SWE
“LLM calls in a for loop”

AIE
AGENT ENGINEERING
HTTPS://AI.ENGINEER

### Crowdsourced Agent Definitions

Six definitions emphasize goals, tools, control flow, long-running processes, delegated authority, and multi-step tasks. The speaker suggests testing definitions against examples such as Deep Research and Operator.

[6:27](https://www.youtube.com/watch?v=5N33E9tC400&t=387s) [Slide presents six definitions of agents, plus metaphors, a reductive definition, and examples.](https://assets.ai.engineer/sites/aiecode2025/sha256/cb0410f516bdf8a7d5164fba5485a1a6b7bcb40166b0b9d2c6f6bd6bbadb7f00.webp)

Simon Willison
@simonw
Let’s see if we can crowdsource a robust definition of “agent” (with respect to AI and LLMs) that fits in a <=280 character tweet

Autonomous Goal-Oriented Systems
“An application that can autonomously learn, adapt, and act in pursuit of its assigned goal”

Tool-using LLMs
“agent = llm + memory + planning + tools + while loop” - swyx

LLM-Driven Control Flow
"The more agentic an application is, the more an LLM decides the control flow of the application" - @hwchase17

Long-Running Processes
“An AI system that's capable of carrying out and completing long running, open ended tasks in the real world.” - @Dan_Jeffries1

Delegated Authority
“a system that is trusted to make impactful decisions on behalf of me”

Multi-Step Task Completion
“An AI system component that performs non-trivial, multi-step operations that previously would have required a human.” - @DanielMiessler

Metaphors
“control system for future states” - @plinz

Reductiveness
"LLM calls in a for loop"

Specific Examples
Code Interpreter
Deep Research
Operator

### Defining agents

OpenAI defines an agent as a model guided by instructions, extended with tools, and encapsulated in a runtime with a dynamic lifecycle.

[6:58](https://www.youtube.com/watch?v=5N33E9tC400&t=418s) [An agent is described as a model with instructions, tools, and a runtime.](https://assets.ai.engineer/sites/aiecode2025/sha256/8e9a319eb55b13c8f9b5255e2fe5d65bdc9ce4504e20aeecc55d8f5846a1efda.webp)

Defining agents

An agent is an AI application consisting of a model equipped with instructions that guide its behavior, access to tools that extend its capabilities, encapsulated in a runtime with a dynamic lifecycle.

Agent =
+ Model
+ Instructions
+ Tools
+ Runtime

OpenAI
Confidential and proprietary.
8

### AI agent capability on GAIA

GAIA test scores rise through 2025. A dashed S-curve extends toward the 92% human baseline, prompting the question of whether agents will reach human-level performance on 10-minute tasks in 2026.

[7:18](https://www.youtube.com/watch?v=5N33E9tC400&t=438s) [Chart plots GAIA scores over time against a projected S-curve and a human baseline.](https://assets.ai.engineer/sites/aiecode2025/sha256/fda80b5934aa39e55819ca7f5a342fa294eeedc548e1c87f55fbea3d46c4aa2a.webp)

AI agents are quickly passing the S-curve of capability
% of correct answers on GAIA test set

Y-axis: Score
Y-axis ticks: 0, 20, 40, 60, 80, 100
X-axis: Year
Timeline: 2023, 2024, 2025, 2026, 2027, 2028

GPT-4-Turbo - FRIDAY
GPT-4o - smolagents
GPT-4o - Dynasaur
Claude-3.5 - Langfun
Claude-3.5 - h2oGPTe
92% : human baseline

Agents reach human-level performance for 10-minutes tasks in 2026?

### Why agents now?

The speaker groups familiar reasons for agent progress into better capabilities, tool use, and tools, from structured output to MCP, browsers, and memory/RAG frameworks.

[7:40](https://www.youtube.com/watch?v=5N33E9tC400&t=460s) [Slide lists better capabilities, tool use, and tools as reasons for agent progress.](https://assets.ai.engineer/sites/aiecode2025/sha256/c008ddeb7065ba6ebaa0a9a1af0be699bbd64e0f2a5dfa56bcdef8aa7d09e195.webp)

why agents now?

Better capabilities
Long inference/reasoning
Better/More data
MMLU/GPQA Saturation

Better tool use
100% Structured Output, BFCLv3

Better tools (AX)
MCP, Sandboxes, Browsers, Search
Memory/RAG frameworks/DBs

SlidesCarnival
August 19, 2024
AIE
AGENT ENGINEERING
HTTPS://AI.ENGINEER

[7:43](https://www.youtube.com/watch?v=5N33E9tC400&t=463s) [Initial slide listing better capabilities, tool use, and tools as reasons for the rise of agents.](https://assets.ai.engineer/sites/aiecode2025/sha256/866ad1a81bf2c93b615117da7462001289c5dc56a73c226e104f76019d4ed48a.webp)

why agents now?

Better capabilities
Long inference/reasoning
Better/More data
MMLU/GPQA Saturation

Better tool use
100% Structured Output, BFCLv3

Better tools (AX)
MCP, Sandboxes, Browsers, Search
Memory/RAG frameworks/DBs

SlidesCarnival
August 19, 2024
AIE
AGENT ENGINEERING
HTTPS://AI.ENGINEER

### Model diversity

The speaker estimates OpenAI’s share fell from roughly 95% to 50% over two years. The slide pairs a model-spend chart with Grok 3 and Thinking Machines as signs of a more diverse field.

[7:49](https://www.youtube.com/watch?v=5N33E9tC400&t=469s) [Model diversity slide with a model market-share chart and references to Grok and Thinking Machines.](https://assets.ai.engineer/sites/aiecode2025/sha256/a8c7815e16abad7a858a5304e443e486e5b124fdd09b4a7203b52ce0b972f7a9.webp)

model diversity

1. It’s a much closer race

Y-axis: Share of spend
Y-axis ticks: 0%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100%
Timeline: Nov 23, Dec, Jan 24, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep
Visible bar labels, left to right: 39%, 49%, 7%; 47%, 46%; 53%, 38%; 66%, 33%; 71%, 26%; 53%, 41%; 70%, 18%; 17%, 58%, 10%; 22%, 55%, 13%; 24%, 38%, 21%; 44%, 23%, 9%, 12%
Legend: Claude 2, Claude 3 Haiku, Claude 3 Opus, Claude 3 Sonnet, Claude 3.5 Sonnet, Claude Instant, Claude Standard, GPT-3.5, GPT-4, GPT-4o
Ramp
CONVICTION

GROK
February 19, 2025
Grok 3 Beta — The Age of Reasoning Agents
We are thrilled to unveil an early preview of Grok 3, our most advanced model yet, blending superior reasoning with extensive pretraining knowledge.

THINKING MACHINES

Interface text:
The State of AI Startups in 2024 [LS Live @ NeurIPS]
Latent Space
13.4K subscribers
Subscribed
Share
Ask
Download
Save
2.7K views  2 months ago
Latent Space LIVE presents our first keynote from NeurIPS 2024 in Vancouver - ‘The State of AI Startups in 2024’. Sarah Guo (Conviction founder, No Priors host) and Pranav Reddy (Conviction [illegible]

## Full transcript

[0:00](https://www.youtube.com/watch?v=5N33E9tC400&t=0s) [on-hold music] Hi.

[0:17](https://www.youtube.com/watch?v=5N33E9tC400&t=17s) Good morning, everyone. [cheering] [clapping] Love that. Love that. Um, I'm gonna get right into it. One of the challenges we have with Summit is that we actually ask our spe- speakers to do very short talks.

[0:27](https://www.youtube.com/watch?v=5N33E9tC400&t=27s) So I, as the leader of Summit, I have to do even shorter talks. [laughing] So let's go. Uh, you can see a lot of these. There'll be a lot of show notes and homework.

[0:34](https://www.youtube.com/watch?v=5N33E9tC400&t=34s) You can see it on the live stream. How is AI engineering doing? Uh, it's pretty good. We have an O'Reilly book. That's pretty cool. Yeah. [laughing] Um, uh, Chip is actually a good friend, and she's actually speaking at, uh, she's giving our keynote for the workshop session, uh, tomorrow, which is pretty cool.

[0:48](https://www.youtube.com/watch?v=5N33E9tC400&t=48s) Uh, Gartner hates us. Gartner thinks we've, we've hit the peak. [laughing] So it's only downhill from here, guys. I'm sorry to inform you that AI engineering is over. Uh, there's no, there's nowhere else, else, uh, else to go but down.

[0:58](https://www.youtube.com/watch?v=5N33E9tC400&t=58s) Um, a lot of, uh, what I try to do with these eight, with these, uh, talks that I do at, at each conference is to try to landmark the se- the, the state of the art or the state of the industry.

[1:10](https://www.youtube.com/watch?v=5N33E9tC400&t=70s) Um, so with Latent.Space, I, I did the rise of the AI engineer. With the first AI Engineer Summit, we talked about the three types of AI engineer. And with last year's AI Engineer World's Fair, we talked about how the discipline of AI engineering was maturing and spreading across different disciplines.

[1:26](https://www.youtube.com/watch?v=5N33E9tC400&t=86s) Um, uh, I think this is starting to get a little stale. Uh, by now a few million people have seen this and, like, you know, uh, used this to form their teams, and I think that was the intended effect.

[1:33](https://www.youtube.com/watch?v=5N33E9tC400&t=93s) What I a- am encountering these days is the two resistance from two sides of the AI engineer spectrum. Uh, if you come from an MLE point of view, you think that the AI engineer is just, like, mostly an MLE plus a few prompts.

[1:46](https://www.youtube.com/watch?v=5N33E9tC400&t=106s) If you come from the software engineering point of view, you think that it's mostly software engineering and, uh, calling a few LLM APIs. Um, I think over time, it- the AI engineering is gonna basically emerge as its own discipline, and it's still not there yet.

[1:59](https://www.youtube.com/watch?v=5N33E9tC400&t=119s) It's still very, very early. I still say things like, "Oh yeah, AIE is ninety percent software engineering and ten percent AI." I think that will grow over time, and I think this is the year when it starts to spread out, and that's, that's what I'm here to talk about a little bit today.

[2:13](https://www.youtube.com/watch?v=5N33E9tC400&t=133s) Um, so for example, uh, I, I think, like, what I try to do with AIE is also, like, it's a, it's a work in anthropology, like how people describe themselves, form groups, form identities, and form industries.

[2:24](https://www.youtube.com/watch?v=5N33E9tC400&t=144s) Uh, so MLE, you know, it leaks out in your language. Um, they say test-time compute because the only reason to run inference is to test it. Uh, AIE will maybe say inference-time compute because we actually really care about inference.

[2:34](https://www.youtube.com/watch?v=5N33E9tC400&t=154s) Um, software engineers maybe reas- reasoning. Um, a- a- and I think you see these differences, and I'm trying to articulate them over time. Um, part of what I wanna do here to set context is to explain why we've kind of pivoted AI Engineer Summit to be the Agent Engineering Conference.

[2:48](https://www.youtube.com/watch?v=5N33E9tC400&t=168s) Um, [coughs] it's not a decision that we made lightly because, uh, we're saying no to all these things. We're saying no to RAG. We're saying no to open models, uh, GPUs, and we're just saying, uh, you know, th- this is the only thing that we're gonna do today.

[3:03](https://www.youtube.com/watch?v=5N33E9tC400&t=183s) Um, and, uh, but, like, closing all those doors actually opens up others. So when we put out the call for speakers, we, uh, made up all this list of, uh, you know, other a- agent engineering disciplines.

[3:13](https://www.youtube.com/watch?v=5N33E9tC400&t=193s) Uh, and I soon realized we didn't have to. I'll talk about this in a bit. Um, I also looked at last year's top-performing talks on YouTube, and you guys told us, uh, that, you know, you really wanted all the, all the agentic things.

[3:25](https://www.youtube.com/watch?v=5N33E9tC400&t=205s) Now, the only problem with this is that we only got speakers who basically made agent frameworks for a living, uh, and everyone's asking the, the, the real question, "Who's putting this in production?"

[3:35](https://www.youtube.com/watch?v=5N33E9tC400&t=215s) So we had a new rule this year of, all right, no more vendor pitches. Um, you know, you, you complain about-- Yeah. Let's-- [clapping] Oh, thank you. [laughs] [clapping] Uh, as a, as a curator, it makes it so mu- infinitely harder because, uh, basically the people that you're about to see have no incentive to come on stage and share what

[3:53](https://www.youtube.com/watch?v=5N33E9tC400&t=233s) they're sharing, uh, but somehow we talked them into it. So, uh, I hope you're looking forward to that.

[3:58](https://www.youtube.com/watch?v=5N33E9tC400&t=238s) Uh, the, the other thing also I realized that, like, everything plus agent works basically. So agent plus RAG works. Agent plus code gen works. Agent plus search works. Um, and this is kind of like the simple formula for, like, making money in twenty twenty-five.

[4:10](https://www.youtube.com/watch?v=5N33E9tC400&t=250s) Uh- [laughing] Most of these, most of these names you'll see in the talks that, uh, that will follow, uh, in, in the sessions.

[4:18](https://www.youtube.com/watch?v=5N33E9tC400&t=258s) Um, some of you have heard this one before, twenty twenty-five is the year of agents, right? If you say it often enough, it might be true. Uh, I think that when people make predictions, oftentimes they confuse what they want to happen for what will actually happen.

[4:31](https://www.youtube.com/watch?v=5N33E9tC400&t=271s) Um, so maybe you believe Satya Nadella, maybe you believe Roman, maybe you believe Greg Brockman, maybe you believe Sam Altman. All of them want you to believe that twenty twenty-five is the year of agents.

[4:41](https://www.youtube.com/watch?v=5N33E9tC400&t=281s) Uh, and I'll be very honest, uh, me and my co-host, Alessio, I think I saw you over there. Hey. [clapping] Um, uh, [laughs] we were pretty skeptical as well. We were on the record being skeptical.

[4:51](https://www.youtube.com/watch?v=5N33E9tC400&t=291s) Actually, actually, all of you are being o- on the record because last-- yesterday, uh, Barr played, uh, Family Feud with, with our, with the audience, and the number two, uh, buzzword that everyone is tired of hearing is agents.

[5:02](https://www.youtube.com/watch?v=5N33E9tC400&t=302s) Um, but fortunately, you guys are not tired enough 'cause you came to today. I have you for one more day of, uh, of agents talk. Uh, but we're on record March twenty twenty-four with David Luan, uh, the former VP of Eng of OpenAI, uh, saying that we, we tell people to take agents off of their branding.

[5:18](https://www.youtube.com/watch?v=5N33E9tC400&t=318s) Uh, now we tell them to put it back on. [laughing]

[5:20](https://www.youtube.com/watch?v=5N33E9tC400&t=320s) So okay. The, um, I, I, I think, I mean, I'm doing this as a public service. To start any agents conference, we have to define the word agents. Are you guys ready?

[5:29](https://www.youtube.com/watch?v=5N33E9tC400&t=329s) Yeah. [laughs] All right. I actually have one. I, I-- it's a monumental task. I could do it in one slide. Um, so if you talk-- Again, this, this is a very POV, sort of anthropological point of view.

[5:41](https://www.youtube.com/watch?v=5N33E9tC400&t=341s) The machine learning people will talk about some kind of reinforcement learning environments. They wanna talk about actions, achieving goals, and all that. Um, AIE, we don't know what they, what they want yet.

[5:49](https://www.youtube.com/watch?v=5N33E9tC400&t=349s) Uh, the, the software engineers are very reductive. They're just, you know, put it in a for loop.[audience applauding] [laughs]

[5:56](https://www.youtube.com/watch?v=5N33E9tC400&t=356s) Okay, you-- It seems like you agree. Um, so, uh, fortunately, you know, I think every AIE conference needs to invoke the name of Simon Willison. Uh, he is our, uh, patron saint.

[6:05](https://www.youtube.com/watch?v=5N33E9tC400&t=365s) Um, he's actually gone and crowdsourced three hundred, uh, definitions of what an agent is. So I didn't have to survey all of you. I, I was thinking about asking every single speaker to start with what is your definition?

[6:14](https://www.youtube.com/watch?v=5N33E9tC400&t=374s) Uh, it doesn't matter. Um, there's-- here's six of them, right? You-- It's either about goals, it's about tools, it's about control flow, it's about long-running processes, it's about delegated authority, uh, it's about multi-step task completion.

[6:26](https://www.youtube.com/watch?v=5N33E9tC400&t=386s) Yeah, I see all the phones coming out. Don't worry, it's on the live stream, right? There's, like, twenty thousand people, uh, watching along. Um, and then there's a, there's a bunch of other things.

[6:33](https://www.youtube.com/watch?v=5N33E9tC400&t=393s) Uh, I think, I think the last one on the bottom left, bottom right is, uh, is an interesting one. Like, just have some things that everyone defines, uh, agrees is an agent, and make sure that they're sort of-- your agent definition is passing those things.

[6:46](https://www.youtube.com/watch?v=5N33E9tC400&t=406s) Um, except, so that was my one slide. That was my slide, uh, of, like, what, what defining an agent. And then yesterday, OpenAI went and dropped a new agent's definition, uh, on the live stream, uh, that you can w- well, watch yesterday as well.

[6:58](https://www.youtube.com/watch?v=5N33E9tC400&t=418s) Um, so this is something that they're obviously going to work with. Um, and, uh, I, I think you should, you should definitely pay attention to, to this because they're, they're building on top of this, uh, new definition as well. [lips smack]

[7:08](https://www.youtube.com/watch?v=5N33E9tC400&t=428s) So that's defining agents. Why now? Why is, why are agents working now when they did not work a year ago, two years ago? Um, I have a rough idea.

[7:18](https://www.youtube.com/watch?v=5N33E9tC400&t=438s) So the people are ta-ta-talking about capabilities, and so, uh, you can see that capabilities, even, even on a trajectory of twenty twenty-three, twenty twenty-five, um, have been, have been really growing, and they start to ru- around to hit human baselines, uh, right about now. [lips smack]

[7:32](https://www.youtube.com/watch?v=5N33E9tC400&t=452s) Um, and I also have a map of other cap-- [phone beeping] uh, reasons as well. So I'll just bring you through each of them. Most people will say, "Oh yeah, we have better reasoning now, we have better tool use now, we have better tools," um, including MCP, which, which you're doing a workshop on, uh, tomorrow. [lips smack]

[7:47](https://www.youtube.com/watch?v=5N33E9tC400&t=467s) Uh, but I think there are some other less appreciated things which I'm gonna bring up to you right now. Model diversity, right? Uh, the OpenAI market share has gone from, like, let's say ninety-five percent two years ago, now down to fifty percent.

[7:56](https://www.youtube.com/watch?v=5N33E9tC400&t=476s) It's a much more diverse, uh, uh, landscape, including, like, this, this, this past week, um, two frontier model labs that are possible challenges to OpenAI have emerged, and which I think, which I think is, um, really exciting for twenty twenty-five.

[8:09](https://www.youtube.com/watch?v=5N33E9tC400&t=489s) We, we don't actually know what is gonna shake out to it by the end of the year. Uh, the second thing is, uh, that the cost of intelligence is super Moore's law, is what I call it.

[8:17](https://www.youtube.com/watch?v=5N33E9tC400&t=497s) Um, it's, it's gone, uh... The cost of GPT-4 level intelligence has gone down one thousand times in the last eighteen months. Um, and you can see the same curve starting for the o1 level intelligence. [lips smack]

[8:26](https://www.youtube.com/watch?v=5N33E9tC400&t=506s) Um, uh, and also we now start to have RL, RL fine-tuning options. Um, I have zero experience in this area. [laughs]

[8:33](https://www.youtube.com/watch?v=5N33E9tC400&t=513s) But fortunately, one of our speakers, Will, uh, is gonna tell us, talk to us, uh, later today about this, about this. Um, so we have all these reasons. We have-- Uh, I have a few more.

[8:42](https://www.youtube.com/watch?v=5N33E9tC400&t=522s) Uh, you know, in our conversation with Brett Taylor, um, he talked about, uh, co- charging for outcomes instead of c-- for, uh, instead of costs. Um, there's a lot of work on multi-agents as well as, uh, faster inference as well that's coming out from the, the better hardware that we have.

[8:53](https://www.youtube.com/watch?v=5N33E9tC400&t=533s) Um, there's more homework there if you want. Uh, this is all sourced and, uh, you know, co-- has, has, has some backing in our, in our Latent.Space conversations, uh, but I don't really have time for that.

[9:02](https://www.youtube.com/watch?v=5N33E9tC400&t=542s) Okay, so one last thing for you guys on agent use cases. So, uh, I think most people agree w-with, like, Barry, um, Barry's, uh, building effective agents talk. Um, he's, he's gonna talk about how coding agents and support agents have product-market fit.

[9:14](https://www.youtube.com/watch?v=5N33E9tC400&t=554s) I think now it's fair to say deep research has PMF. Um, but also I will say up and coming are some of these use cases, uh, some of which you'll, you're gonna see in the con-- the talks later.

[9:23](https://www.youtube.com/watch?v=5N33E9tC400&t=563s) But I also wanna s-- offer anti-use cases. Can we please stop demoing agents that book flights? Yeah?

[9:28](https://www.youtube.com/watch?v=5N33E9tC400&t=568s) Yeah.

[9:29](https://www.youtube.com/watch?v=5N33E9tC400&t=569s) No more flight booking agents. [audience applauding] Uh, [laughs] I wanna book my own flights, thank you very much. I wanna, I wanna book my own Instacart orders, and also please don't astroturf for it.

[9:38](https://www.youtube.com/watch?v=5N33E9tC400&t=578s) Right. Uh, okay. So, uh, one, uh, yeah, and I think the reason that the tell that, uh, you know... This is, this is a headline that I saw yesterday.

[9:45](https://www.youtube.com/watch?v=5N33E9tC400&t=585s) I had to put this in. Um, OpenAI reported four hundred million users, uh, which is a thirty-three percent growth from three months ago. Um, and then you can ask deep research to research OpenAI and draw this chart of ChatGPT growth, uh, going from, uh, zero to, uh, four hundred million users in two years-- in two and a

[10:01](https://www.youtube.com/watch?v=5N33E9tC400&t=601s) half years. Um, so, uh, I, I, I remember this chart very well because Open-- Uh, ChatGPT spent a year not growing. And why did it spend a year not growing?

[10:09](https://www.youtube.com/watch?v=5N33E9tC400&t=609s) Because they didn't ship any, any agentic models. Um, and if you actually just look at the, uh, the sort of weekly active user chart and stretch it out, you actually get this chart, uh, which is actually super interesting because it basically shows that one, one, um, the sort of o1 models have doubled ChatGPT usage.

[10:27](https://www.youtube.com/watch?v=5N33E9tC400&t=627s) And if you stretch it out, um, ChatGPT is gonna hit a billion users by the end of this year. This year. Uh, it's basically going to quintuple the number of users it had, uh, as of September of last year.

[10:37](https://www.youtube.com/watch?v=5N33E9tC400&t=637s) Um, and so, like, the, the, the, the, the growth of, of ChatGPT and the growth of any AI product is gonna be very, very tied to reasoning capabilities and the amount of agents that you can ship for your users.

[10:46](https://www.youtube.com/watch?v=5N33E9tC400&t=646s) Um, it is, it is real. It is, it is, uh, huge, huge numbers. This is one-eighth of the world population that's gonna be using ChatGPT by the end of this year, and I think there's a lot of money left on the table for everyone else.

[10:56](https://www.youtube.com/watch?v=5N33E9tC400&t=656s) So, um, I hope you enjoy doing that. Um, I'm well past time, so I'm gonna skip all this. But basically, I, I think that the job of AIE is now evolving towards building agents in the same way that MLEs build models, software engineers build software.

[11:07](https://www.youtube.com/watch?v=5N33E9tC400&t=667s) Um, so, uh, I'm gonna skip all that. You can see all, you can see all that on, on the, on the live stream. Uh, but we're actually, uh, you know, just here to welcome you to the show.

[11:15](https://www.youtube.com/watch?v=5N33E9tC400&t=675s) Um, and, uh, I'm really excited to introduce you to everyone. So, um, thank you, and I hope you enjoy. [upbeat music] [audience applauding]

[11:25](https://www.youtube.com/watch?v=5N33E9tC400&t=685s) Thank you.
