2026-06-30

They’re letting me keynote at this conference for some reason

The @aiDotEngineer World’s Fair kicks off today in San Francisco. A lot of AI-pilled humans and human-controlled robots in action already! (btw this is @ricmac & I’ll be tweeting and blogging from AIEWF on Latent Space over the next several days, on behalf of @swyx)

A candid hallway track interview with the man himself - @swyx
Mini Podcast if you will, he's told me about a huge milestone tomorrow, things that have surprised him about @aiDotEngineer and ... a few things on the personal side!
0:00 — Daily Context AIE Newsletter
0:47 — @aiDotEngineer Growth
1:23 — Workshop Lines & Skills Track
2:28 — Jam Corner & Music
3:52 — AI Engineer Essay Anniversary
4:26 — Intensity, Integrity, Intentionality, Insanity
4:52 — AGI Pills & Insane Perks
5:56 — Livestream & YouTube CTA
Hope you enjoy this "behind the scenes" chat with the guy who made this whole thing happen.
Also chatting with us, @kentcdodds @liadyosef @WolframRvnwlf @RayFernando1337 @Gangadhar_P
Hallway track remains unmatched!

Introducing Memory Stargraph, an AI memory Visualizer for GBrain 👏
Memory Stargraph is a local web service for exploring a GBrain knowledge base through an interactive star-cloud entity graph.
I built it because my local agent clusters now share the same GBrain-backed memory infrastructure across multiple hosts and agent runtimes. Instead of each agent carrying more context forever, they can rely on a shared, evolving knowledge graph that stays compact, searchable, and useful over time.
Memory Stargraph makes that shared memory visible.
You can see what agents have created in GBrain, inspect entities, explore relationships, create new nodes, connect them, modify links, attach media, and follow the map as it loads new neighborhoods on demand.
It is part debugging tool, part knowledge browser, and part “wait, the agents remembered all this?” moment.
It’s also surprisingly mesmerizing.
Many thanks to @garrytan for his GBrain work, and I am very happy to be among the serious users of #GBrain!
If you’re interested in setting it up for your own agents:
https://github.com/techtony2018/memory-stargraph
#AI #AIAgents #KnowledgeGraph #LocalAI #OpenSource #BuildInPublic #Codex #OpenClaw

"Every one of these children had a name. They had dreams, loves, and futures. Their lives were as valuable and special as your child’s or mine. They were loved, and somewhere, their families grieve for them."
https://lnk.thebulwark.com/4xRra08

Supervisor Jackie Fielder spent 104 days on PAID leave — the longest any SF Supervisor has ever taken — collecting over $50k while hiding from an ethics investigation into her crypto PUMP-and-DUMP and a probe into her office leaking a memo to sabotage the RESET fentanyl center.
While she was gone, drug markets spread and overdose deaths continued in District 9…
We are demanding she resign by 5pm Friday. If she refuses, we serve the Notice of Intention to recall Monday.
She is not fit to return.
Jackie please do what’s best for yourself and the district and resign.

Thrilled to announce the Wearable AI Workshop at ECCV 2026 🎉 that we're organizing with an awesome group of folks across Meta Reality Labs, AMI Labs, HKUST, Georgia Tech, UCF, and U. of Edinburgh. If you work on egocentric perception, multimodal assistants, or long-form video understanding, definitely consider participating. Details below 👇
🔬 Wearable AI Workshop @ ECCV 2026 | Malmö, Sweden
Join us to advance the science of real-time, multimodal, context-aware assistants — with access to the largest egocentric wearables dataset ever released, $21K in challenge prizes, and a venue to shape the research agenda alongside leading groups in the field.
🏆 Grand Challenge — $21K in Prizes
Compete on the largest egocentric wearables dataset ever released (5,000+ hours) across three tracks:
Proactive AI — Predict when and how to proactively engage users
Multi-Turn Conversations — Streaming answer prediction from egocentric video
Long Video Q&A — Temporal grounding and recall over 30+ minute videos
🎤 Confirmed Keynote Speakers
Kristen Grauman (UT Austin) — Professor, Computer Vision Group
Dima Damen (U of Bristol, Google DeepMind) — Professor of Computer Vision
Raffay Hamid (Meta Reality Labs) — Distinguished Engineer, GenAI
If you work on egocentric vision, long-form video understanding, conversational AI, or on-device inference — this is your venue.
📅 September 2026 · 📍 Malmö, Sweden
🔗 Workshop URL: https://wearable-ai-workshop.github.io/
#Meta #MetaRealityLabs #MetaAI #AMIlabs #HKUST #UTAustin #GeorgiaTech #UniversityOfEdinburgh #UniversityOfBristol #ECCV #ECCV2026 #WearableAI #EgocentricVision #ComputerVision #MultimodalAI #AIResearch

great vibes at SFMOMA Speakers Dinner today!
if you’re at @aiDotEngineer, come join my talk on Thursday at 11:40. see you there!
Kate Deyneka: let’s goo!

okay but did you see @charles_irl and @swyx bringing the vibes today 💫

grats to @swyx 99% of my timeline is @aiDotEngineer now.

We can finally say AI isn't killing jobs.
A new paper from me, @tryramp, and @RevelioLabs uses firm-level spend and workforce data across 21K U.S. businesses to measure AI's impact on jobs.
Firms that adopt AI heavily grow headcount 10% over two years following adoption. Low adopters see no statistically significant change.

Today we're launching @SpellbookLegal's biggest thing yet: Autonomous Contract Management
It’s the first end-to-end AI infrastructure for contracts
The world is speeding up. We are in one of the biggest investment cycles in decades.
Behind every rocket launch, FIFA game, and datacenter, lies a web of hundreds of agreements.
Agreements are the invisible threads that allow us to work together.
We have infrastructure for finance (Stripe, Ramp), eCommerce (Shopify) and many business functions.
But agreement infrastructure is lacking. This creates a painful bottleneck on our ability to work together.
Online purchases take milliseconds. But agreements still take weeks.
CLMs were supposed to be the answer, but were designed in a pre-AI era.
AI fundamentally changed how computers can accelerate agreements.
Spellbook is the most used AI contract review tool in the world, with ~5,000 customers in 80 countries
Now we are expanding to deliver the first end-to-end, AI native stack for contracts.
From the moment a deal lands in your inbox, to the day it renews years later, Spellbook’s AI supports teams every step of the way, across all business teams.
It runs 24/7. While you sleep, it's reading the deals that came in overnight, flagging the parts that actually need a lawyer, and clearing the busywork that used to kill mornings.
Nothing gets handed off between systems or slips after signature, and the intelligence stays with you for the life of the contract.
AI for lawyers is great. There are 20 million lawyers in the world, and many are our users.
But there are billions who touch contracts. We're excited to help everyone move faster and do more of what they love, by building the best AI-powered contract infrastructure in the world.
Get early access: https://spellbook.com/acm

The discussion about @elonmusk's comments on humanitarian aid has mostly focused on our values, which seems right. But note that USAID served our interests as well as our values, by protecting us from diseases like the Ebola outbreak now ravaging Congo.
Atul Gawande: Speed of response is everything, and it's still lagging far behind. Essential explainer on the Congo Ebola outbreak from @JeremyKonyndyk. https://www.nytimes.com/interactive/2026/06/09/opinion/ebola-outbreak-africa-usaid.html?unlocked_article_code=1.o1A.Mkun.taxeXFX7bh8M&smid=url-share

Speaking at @aiDotEngineer today about long-horizon RL and the different tradeoffs involved. 2:25pm.
Also here in SF for a week if anyone wants to catchup 👋
Ross Taylor: Excited to be speaking at @aiDotEngineer World’s Fair alongside partner in crime @ChengxiTaylor.
It seems we have a lot to talk about:
🤼‍♂️ Actor-critic is hot again? How are current RL algorithms changing with long-horizon rollouts.
🌍 What does a good long-horizon environment

How companies should think about driving AI adoption from @jess__yan (Anthropic product lead):
“Alot of enterprises immediately jump to: How could I automate this crazy twenty-team workflow that would have required a lot of cross-cutting coordination?
But there is something really valuable about: how do we unlock the individual? How do we make any individual on any team feel exponentially more powerful?
There is a huge amount of value that’s unlocked simply by making every employee feel like they have their own power to develop products as a one-person startup.”
📌 Watch the full episode: https://www.youtube.com/watch?v=Xu5gz2qsaz8
Peter Yang: "We give [agents] tasks overnight and then we wake up and the backlog is resolved and bugs are squashed."
Here's my new episode with @jess__yan, product lead at Anthropic. Jess showed me how to build a long-running Claude agent from scratch and how Anthropic product teams use

We're coming out of stealth.
We've built our first racks after a successful A0 tapeout, $1B+ in customer contracts, and $800m raised.
Early customer tests show us achieving SOTA throughput, latency, and power efficiency on inference workloads.
Our first racks ship this summer.

Was thinking if I should highlight this tweet or not, but it’s a masterclass in the amount of vitriol people face when working on open source.
Is the app great yet? No. It’s a start.
It was built by the community. Getting the iOS and Android apps working with secure pairing and push notifications - and getting both through App Review -took a surprising amount of work.
OpenClaw wasn’t acquired by OpenAI and isn’t an OpenAI product. It’s an open, independent project under the OpenClaw Foundation. OpenAI sponsors the project’s token usage; I work there.
Cristian, your tweet was just one of ~30 I woke up to today. I’d genuinely love your help making it great.
Attention is still the scarcest resource.
I’d rather spend mine encouraging people who build.
cristian rus: imagine getting acquired by @OpenAI, get unlimited AI tokens and still drop this slop abomination

It's wild how quickly Etched designed and got the chips out, all within 2 years. They went deep, hardcoding attention into silicon and getting very high MFU. This kind of hardware tailored made for LLM inference is soon gonna bring cost of intelligence down 10x
Etched: We're coming out of stealth.
We've built our first racks after a successful A0 tapeout, $1B+ in customer contracts, and $800m raised.
Early customer tests show us achieving SOTA throughput, latency, and power efficiency on inference workloads.
Our first racks ship this summer.

Special new @latentspacepod live recording added to the Expo schedule on Thursday, with @swyx and @robertwachen alongside his Inference track appearance!
congrats to Etched on a massive milestone!
Etched: We're coming out of stealth.
We've built our first racks after a successful A0 tapeout, $1B+ in customer contracts, and $800m raised.
Early customer tests show us achieving SOTA throughput, latency, and power efficiency on inference workloads.
Our first racks ship this summer.

Looking for an early hire in SF for helping with model optimization: spec decoding, GPU kernels, pooling infra, and wants to go deep on vLLM/sglang. Hunger/interest over experience in this case. If you know anyone, DMs open. They’d work directly with me.

Giving a talk on agent-to-agent and AI network effects at @swyx 's AI Engineer World Fair today at 1:30p in Room 2010. Come say hi!
I think this talk will be a good one if I may say so myself.
https://www.ai.engineer/worldsfair/schedule?session=asn_slot_2026_06_30_breakout_track_01_1330_2026_06_11t09_55_41_463z

15 minutes before @swyx is on the main stage at @aiDotEngineer
The amount of people visiting the worlds fair this year is unbelievable 😲

Loops! @swyx

Re @swyx kicking it off! @aiDotEngineer

Re @swyx kicking off the @aiDotEngineer WF!

Re Opening welcome from @swyx.

Mr @swyx kicking of @aiDotEngineer #aiengineer
Slides at https://loopcraft.swyxio.workers.dev/

If there's one thing Americans can agree on, it's that there's too much money in politics.
But the Supreme Court just decided to demolish some of the last safeguards we had left - creating an even bigger role for special interest spending.
We must do whatever it takes - including a constitutional amendment if necessary - to combat the influence of money in our political system.

As a European, I don't understand how the sitting U.S. president—who has to pay $5 million to a woman he r@ped and defamed—can remain in office after everything he's done.

Word of the day so far at AIEWF is Loop. @swyx talked about “loopcraft” in his opening address, and the word was used constantly by the following speakers from Microsoft and OpenAI, and then “the clawfather” Peter Steinberger.

it's a sad day when we stop learning.
nothing quite replicates the pace and structure of college, but Independent Studies has become my version of continuing education post-grad.
some of my favorite classes back then were the ones i took on a whim, like cross-registering to take 'intro to philosophy' at Harvard or learning Chiense.
it's officially been 6 months of lessons across [Technical] Write and Learn (cohosted w/ @swyx) to learning Rust, from some of the best in the industry.
the generosity to teach, and the eagerness on the other side to learn, is one of the most beautiful things about sf.
we keep the classes small to stay engaging and personal. huge thank you to every teacher we've had along the way and please let me know if you'd ever like to teach or participate

Founders step on every rake in the garden trying to figure out the Valley. I know, because my forehead is full of rake marks. I decided to interview some of the best investors about where the rakes are hidden in a new series called First Principles, starting with @saranormous.

There is a lot of pride among AI founders today around doing "996." 9 to 9, 6 days a week. SF is normalizing the 72-hour week to win the AI race.
I started Upside to enable a different way of winning. The whole promise of AI is that people should work LESS, and only on WHAT MATTERS not get chained to their desks grinding.
@alexdbauer wrote more on how we did it, I just made the images :-) and @swyx and @vibhuuuus helped us print them at @aiDotEngineer yesterday.
Alex Bauer: http://x.com/i/article/2071855004240269312

Yo dawg, I heard you like loops...
(from @swyx's AI Eng keynote this morning)

What an honor to emcee the first day of @aiDotEngineer and introduce the Software Factories Track
Thank you @swyx & team, and @KeycardLabs for the support.
“A year ago @GeoffreyHuntley released the Ralph loop. It captured our attention and sparked our imagination as we watched Ralph loops work autonomously overnight and forge entire products on its own.
However, it wasn't perfect and in the early days it came recommended for greenfield work only and it came with the expectation that it would get you about 90% of the way there.
Since then we've learned quite a lot about loop engineering and in the last year there's been considerable advancements that make now that largest inflection point for software factories.
Model vision has improved allowing them to see and verify work they couldn't before. Agents now have access to far richer tool ecosystems and live data, enabling them to work across real production environments. AI security practices have matured, giving agents the security, autonomy, and capability they need to perform meaningful work. Context windows have gotten larger and agent memory has improved allowing agents in a loop to track what's been done before. Lastly, reasoning models have improved helping the model think through more sophisticated tasks.
All of this comes together now.
Software factories are no longer a vision for the future. They've become a practical reality.
Today's speakers are on the frontier of this shift and helping define this inflection point. They'll help us separate what's real from what's hype, share what actually works, and how to build software factories that produce reliable results, not slop.”

Introducing Claude Sonnet 5, our most agentic Sonnet yet.
It makes plans, uses tools like browsers and terminals, and runs autonomously at a level that just a few months ago required larger and more expensive models.

California passed some of the biggest pro-housing reforms in decades. But out of 6.1 million parcels made eligible under SB 9, just 266 projects were permitted by end of 2023.
A @Yimby_Law report documents how local governments systematically buried state housing law under emergency ordinances, fee spikes, design traps, and delay tactics.
https://www.yimbylaw.org/law-journal/californias-streamlining-laws-dlf8x

Certified token billionaire. Having a blast at AIE, fantastic conf @swyx
Ryan Lopopolo: Fixed

Re @charlieholtz preach!
“Factories” is a depressing vision of the future, metaphors matter

Don't fall for Elon Musk's gaslighting.
There is a massive, life-or-death difference between well-planned budget cuts and a sudden, reckless halt to promised medical aid.
Musk is actively lying about the fatal consequences of his USAID cuts. The US is not obligated to fund healthcare for the entire planet, but if a nation decides to pull back its global aid, it must be done responsibly. Ripping away funding mid-treatment means you are directly responsible for the preventable deaths that follow.
Instead of owning up to the fallout, Musk is flooding X with dishonest spin. He first challenged critics to name even one person whose death he caused. When people actually did, he immediately moved the goalposts, claiming that any spending cut inevitably causes deaths.
He is a dishonest charlatan trying to hide the fact that his chaotic decisions got thousands of vulnerable children and adults killed. That will be his lasting legacy

Marco Rubio and Elon Musk claim that no one died from shutting down USAID. Yet the best estimates are between 8–20 million by 2030. We can debate the exact number—but it's clearly in the millions.
When confronted by actual names by @NickKristof, who has gone to the villages and talked to the families, Musk's only response was to call him "an utter piece of shit and a liar." Elon Musk is not just out of touch with reality. He is causing real, devastating harm to the world's poorest people.

Had a lot of fun at @aiDotEngineer talking about BM25!
Ruggero Gargiulo: Cool Search paradigm by @jobergum: “BM25 + Grep is all you need”:
1. Use BM25 to narrow down a large universe of documents to a few candidates
2. Expose the candidates to the agent in a virtual file system
3. Let the agent grep through them and find information like it normally

Re Room 2016 for those attending @aiDotEngineer 2:25pm.
Will also cover Galactica, early Llama reasoning efforts and more - think this is the first time I’ve ever covered this in a public talk 👀.
@swyx

It's not just the Lutnick sons cashing in on the Trump family's White House self-dealing.
World Liberty Financial — the crypto venture Don Jr. & Eric launched with Steve Witkoff's sons, Zach & Alex — has netted a fortune for both families (at least $1.1 billion).

TryNearby's @FaZeApex went from making YouTube videos at 14 to co-building FaZe Clan, and now he's chasing his next big idea. Watch the full interview on YouTube to hear how his creator roots are shaping what he builds next.
https://youtu.be/0dDMmST3rQY

http://x.com/i/article/2072056179295965184

Cute illustrations made by @bclavie giving a talk at @aiDotEngineer @mixedbreadai
“Code is knowledge but knowledge is not code”

California's June primary showed that voters want competence over ideological theater. In SF, the moderate coalition consolidated around Mayor Lurie while the once-dominant progressive machine fractured into infighting. November's ballot, a one-time tax on the wealthiest, will show whether that momentum can scale statewide.
Read the full analysis, from Forrest Liu: https://garryslist.org/posts/california-s-primary-has-ushered-a-new-mandate-time-to-move-beyond-ideology-to-governing-competence

Thank you to everyone to came to the Claude managed agents workshop at @aiDotEngineer with @gcemaj and I. We had an absolute blast sharing our journey and walking you through building your first agent. And really enjoyed engaging with the community and answering your questions.
Thank you @swyx for this opportunity!

Claude can run your company's growth with these 5 skill libraries.
Bookmark this.
1. alirezarezvani/claude-skills — 345 skills across 17 domains, 46 of them are specifically for marketing, organized into pods for content, SEO + AEO, CRO, growth, and landing pages.
2. coreyhaines31/marketingskills by @coreyhainesco – Marketing skills for Claude Code and AI agents. CRO, copywriting, SEO, analytics, and growth engineering.
3. gooseworks-ai/goose-skills by @GooseworksAI: 100+ skills + data APIs for growth, social, ads, and gtm. We built these and we use these every day. You can use this to find leads, scrape social, check ad performance, find influencers, generate creatives and more.
4. GBrain by @garrytan - this is an incredible foundation you can use to create a company brain. It has a knowledge graph, a synthesis layer to give you real answers and (my personal favorite) dream cycles so it updates the knowledge base every day.
5. AgriciDaniel/claude-ads — audits a creative against 250+ checks before you spend.
Links to all below.

. @swyx @vibhuuuus @aiDotEngineer i'm ready for the poaster tomorrow. Is speculative decoding all we need? will glm 5.2 dflash/dspark be lossy on long tail task?
swyx @aiDotEngineer WF: very successful first poster sessions day at AIEWF, thanks to @heathercmiller and @ACM_President for the support!!
tomorrow: Poaster sessions! submit your hottest tweets to @vibhuuuus for printing!
spotted: left shark

Can regularization based JEPA (e.g. SIGReg) scale and compete with SOTA foundation models (DINO)? Here is the answer: yes and with 10x less data.
VISReg (slight variation of SIGReg) competes with DINOv2-LVD142M while only training on inet22k.
Try it out: https://huggingface.co/BooBooWu/visreg
Haiyu Wu: Working on world model or SSL? You definitely need to try our new work: VISReg!
What does it achieve?
💪 Strong collapse prevention: High gradient when embedding collapse
⚡ Friendly to scale training: Linear complexity to scaling factors
🧩 Easy to train: Similar to LeJEPA, it

it’s incredible what swyx has built with aie in just three years

He's an adjudicated rapist, and SCOTUS knows it. He fomented an insurrection, and Jack Smith knows it. He botched a pandemic, and Dr Fauci knows it. He has committed every type of fraud known to man, and all men know it. He protects pedophiles. He lies as he breathes. He monetizes the presidency and weaponizes justice. He's lawless. He obstructs at every turn. He's an idiot - useful and otherwise. He's narcissistic. He's psychopathic. He's paranoid. He's sadistic. I can go on.
YET - no matter how bad we know he is, and no matter how bad the things we know he did are, remember this:
Some of the things we don't yet know, are even worse.

Context engineering has its own track at the @aiDotEngineer World's Fair this year. 🎉 I've respected what @swyx and the @latentspacepod team have been building for years — and I'm pumped to be a part of it.
This is a conference about shipping AI, not just talking about it.
I'll be contributing to the aforementioned context engineering track with a breakdown on WTF is the context layer, and how teams are using it to improve agent accuracy in production.
If you'll be there, let's catch up!

We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5.
We'll begin restoring access tomorrow, and will share an update soon.
We’re grateful to our users for their patience, and to everyone who worked with us on redeploying the models.

Just SF things feat. the Orc-estrator & @swyx
#aiewf @aiDotEngineer

Got this at ai engineer world fair lol
@swyx