Latent.Space(@latentspacepod)

Chai Discovery's BioAI Phase Shift: Chai-3, Photoshop for Molecules, One-Shot Drug Design, & Enginee...

Latent.Space (@latentspacepod) — Tue, 11 Aug 2026 22:16:28 GMT

Chai Discovery's BioAI Phase Shift: Chai-3, Photoshop for Molecules, One-Shot Drug Design, & Engineering Biology latent.space/p/chai-discove…@chaidiscovery cofounder @Mattmcpartlon1 and product lead @realdealpatil explain why AI drug design has crossed the threshold from impressive demos to pharma-trusted tools, how Chai-2 and Chai-3 are pushing toward one-shot therapeutic-grade molecules, how Chai’s CAD-like design suite lets scientists precisely engineer antibodies and binding sites, why faster model-to-lab feedback loops could turn pharma’s waterfall discovery process into an iterative loop, and how better models, validation, compute, and infrastructure are turning biology from trial-and-error science into an engineering discipline.

Your browser does not support the video tag.

🔗 View on Twitter

💬3🔄3❤️31👀18068📊11

⚡ Powered by xgo.ing

Zawinski's Law of MultiAgents: Every agent attempts to expand until it can message other agents. Th...

Latent.Space (@latentspacepod) — Sat, 08 Aug 2026 01:13:00 GMT

Zawinski's Law of MultiAgents:

Every agent attempts to expand until it can message other agents. Those agents which cannot so expand are replaced by ones which can.

latent.space/p/ainews-zawin…

💬1🔄0❤️9👀1722📊3

⚡ Powered by xgo.ing

never ever fails

Latent.Space (@latentspacepod) — Fri, 07 Aug 2026 23:35:27 GMT

never
ever
fails

Boris Cherny@bcherny

turns out you can get indirect prompt injection to ~0 on unseen attacks if you stack enough layers (model training + input probes + a classifier checking intent). didn't expect that a year ago. auto mode is default in claude code as of next week

claude.com/blog/auto-mode…

🔗 View Quoted Tweet

💬1🔄1❤️1👀2430📊1

⚡ Powered by xgo.ing

Congrats @EngramLab & @harvey! Full Episode: https://t.co/hGVabn38zM

Latent.Space (@latentspacepod) — Fri, 07 Aug 2026 20:05:41 GMT

Congrats @EngramLab & @harvey!
Full Episode: https://t.co/hGVabn38zM

💬0🔄0❤️0👀976📊1

⚡ Powered by xgo.ing

The types of hard problems @EngramLab works on ft. co-founder @dan_biderman: "Clients do financing,...

Latent.Space (@latentspacepod) — Fri, 07 Aug 2026 20:05:40 GMT

The types of hard problems @EngramLab works on ft. co-founder @dan_biderman:

"Clients do financing, mergers, acquisitions and things like take loans and do deals. And there's many queries that agents might run into which are these kinds of ambient, hard questions that are not easily searchable with RAG.

For example, if you want to ask, which M&A deals haven't we completed this year? To actually solve this problem, you have to go client matter by client matter [and] read all the files. You can't read in any place that it was not completed."

Your browser does not support the video tag.

🔗 View on Twitter

Harvey@harvey

We're open sourcing a 100M+ token synthetic law firm we built with @EngramLab.

The firm contains work product from 250+ synthetic matters across 46 clients, spanning ~10k files.

We built this environment to evaluate an agents' ability to search and understand a firm's past practice to inform present work  - the same knowledge that a tenured associate or partner would have.

It's our first step towards building agents that deeply understand a firm's work and processes. More to come soon

Deep dive by @ItsJulioPereyra and @nikogrupen:

🔗 View Quoted Tweet

💬2🔄3❤️12👀8646📊6

⚡ Powered by xgo.ing

🆕 Unpacking ChatGPT Work https://t.co/wccYg8wY9K ChatGPT Work, an extension of the @openai Codex ...

Latent.Space (@latentspacepod) — Wed, 05 Aug 2026 18:37:46 GMT

🆕 Unpacking ChatGPT Work
latent.space/p/unpacking-ch…K

ChatGPT Work, an extension of the @openai Codex harness to cloud and general knowledge work, launched on July 9 and crossed 10M users 3 weeks later. 

@shloked guests with an A+ reconstruction of how Memory, Proactivity, Scheduling, Browser Use, Plugins, Skills and Tools work in this new harness that is bringing a full agentic experience to the almost 1B weekly active users of ChatGPT!

💬4🔄11❤️124👀76208📊67

⚡ Powered by xgo.ing

Full Episode: https://t.co/CYFNv3NlXY

Latent.Space (@latentspacepod) — Wed, 05 Aug 2026 01:40:12 GMT

Full Episode: youtube.com/watch?v=7PSXtr…

💬0🔄0❤️1👀972📊1

⚡ Powered by xgo.ing

"With megakernels, honestly I'm very bearish. [Megakernels] was a good research direction and it see...

Latent.Space (@latentspacepod) — Wed, 05 Aug 2026 01:40:11 GMT

"With megakernels, honestly I'm very bearish. [Megakernels] was a good research direction and it seems very... intuitively and theoretically, it's nice.

[However], the kernel complexity itself makes it very difficult to write a very optimized megakernel."

Your browser does not support the video tag.

🔗 View on Twitter

ali@waterloo_intern

two weeks ago i went on @swyx's pod and said some things that i... should not have said. 

a lot has happened since then, i owe you all an apology.

i'm sorry that i was right about every single thing. 

a) re megakernels are dead
why are megakernels useful? you spend two months writing a kernel to save time on launch overhead and poor inter-kernel overlap. you had PDL but then people said it wasn't perfect, that you could still get some marginal gains due to straggler CTAs and therefore- wait, sorry, I forgot, Rubin fixes that (kernel two needs 10 CTAs and kernel one has seven finished and three straggling, kernel two launches seven of its CTAs). given a long enough timeline, it all evens out. no serious inference provider is using a 67k loc hand-fused forward pass kernel in production, and the teams doing that are doing so out of pure research. dead. 

b) re ASICs are dead
i'm sorry. to be specific: data-center transformer-inference ASIC companies (not naming any) who etched the arch into silicon have bet on architectural convergence. read kimi's architecture. read deepseek. qwen. we did not converge, and probably will not. dead.

c) re gpu kernel dev is dead
this one kind of hurts because it is (was) my job. gpu kernel optimization is the single most RL-able task in existence
correct=check_correctness(kernel, shape) for shape in shapes 
if all(correct): time(kernel)
give an agent ncu cli and an mcp with nvidia's tribal knowledge and it's done. dead.  

d) re NVIDIA is scared of AMD 
humans hate programming AMD. i'm sorry. it's just true. fine taking a performance hit as long as i don't have to touch rocm or a programming paradigm that says a warp is 64 threads (wtf?)...but an agent does not...  so assuming software no longer moat, HBM capacity and bandwidth matter, and currently on perf / price they're goated. 'bUt NvIdIa iS gOaTeD oN hArDwArE sOfTwArE  cOdEsIgN' and that's the new moat. watch how much tooling they open source to get kernel devs on nvidia. 

apologies all.

🔗 View Quoted Tweet

💬2🔄0❤️5👀3266📊3

⚡ Powered by xgo.ing

Well @pmarca agrees: "I think in 10 years, I'm not sure there will even be a salient concept of a p...

Latent.Space (@latentspacepod) — Tue, 04 Aug 2026 17:09:29 GMT

Well @pmarca agrees:

"I think in 10 years, I'm not sure there will even be a salient concept of a programming language in the way that we understand it today.  And in fact, what we may be doing more and more is a form of interpretability."

Your browser does not support the video tag.

🔗 View on Twitter

Polymarket@Polymarket

NEW: Elon Musk predicts AI will soon make source code obsolete by generating efficient binaries directly.

🔗 View Quoted Tweet

💬4🔄2❤️13👀6063📊7

⚡ Powered by xgo.ing

Full Episode: https://t.co/gVtRX2lScM

Latent.Space (@latentspacepod) — Tue, 04 Aug 2026 17:09:29 GMT

Full Episode: youtube.com/watch?v=knx2wr…

💬0🔄0❤️1👀755📊1

⚡ Powered by xgo.ing

The Inference Engineering Masterclass: 10x faster models, quantization, speculative decoding, Rubin,...

Latent.Space (@latentspacepod) — Mon, 03 Aug 2026 21:49:11 GMT

The Inference Engineering Masterclass: 10x faster models, quantization, speculative decoding, Rubin, & self-optimizing AI latent.space/p/inference-eng@Baseten@philipkiely and @waterloo_intern explain what actually happens after a model is trained, why turning weights into a fast and reliable product creates an entirely new optimization problem, how quantization errors can cancel out to unlock more throughput, why inference teams are still finding 20–200% performance gains, how video generation runs into a quadratic attention wall, and how GLM-5.2 helped rewrite and optimize the GPU kernels serving GLM-5.2 itself.

Your browser does not support the video tag.

🔗 View on Twitter

💬3🔄15❤️257👀145072📊121

⚡ Powered by xgo.ing

AI engineers are rediscovering ontologies as a way to keep probabilistic agents inside deterministic...

Latent.Space (@latentspacepod) — Thu, 30 Jul 2026 13:43:18 GMT

AI engineers are rediscovering ontologies as a way to keep probabilistic agents inside deterministic boundaries. We delve into presentations from @coyle_frankp and @emileifrem from the @aiDotEngineer World's Fair, to see how technologies from the old web are being used to keep LLMs honest and agents on track. latent.space/p/ontologies-a…

💬5🔄7❤️11👀5334📊13

⚡ Powered by xgo.ing

OpenAI’s Everything Agent: ChatGPT Work, Codex, Sites, Memory, & the AI Generalist https://t.co/HfJH...

Latent.Space (@latentspacepod) — Tue, 28 Jul 2026 21:30:00 GMT

OpenAI’s Everything Agent: ChatGPT Work, Codex, Sites, Memory, & the AI Generalist latent.space/p/chatgpt-work@OpenAI Core Product Engineering lead @akshaynathan_ explains why Codex unexpectedly became a superpower for non-developers, how ChatGPT Work brings the same agent harness to knowledge work, why Sites may replace decks and spreadsheets, how memory and persistent computers are turning ChatGPT into a personal operating system, and why ideas, taste, and measuring real progress are becoming the new bottlenecks.

Your browser does not support the video tag.

🔗 View on Twitter

💬0🔄1❤️8👀1923📊2

⚡ Powered by xgo.ing

Poolside’s Race to AGI: Model Factory, Laguna S, Open Models, & Beyond MCP https://t.co/V9cICBXMmA ...

Latent.Space (@latentspacepod) — Thu, 23 Jul 2026 19:01:32 GMT

Poolside’s Race to AGI: Model Factory, Laguna S, Open Models, & Beyond MCP latent.space/p/poolside@poolsideai co-founder and CEO @eisokant explains why coding is the path to AGI, how Poolside’s Model Factory runs 10,000–20,000 experiments a month and turns new models around in as little as eight weeks, why Laguna S suggests persistence and verification may matter more than raw intelligence, why agents are already helping build the next generation of models, and why the future should have 100 foundation model companies not five.

Your browser does not support the video tag.

🔗 View on Twitter

💬1🔄11❤️57👀33084📊21

⚡ Powered by xgo.ing

err did codex stall out at 10M? or are we due a super big announcement today...

Latent.Space (@latentspacepod) — Mon, 20 Jul 2026 15:54:55 GMT

err did codex stall out at 10M?  or are we due a super big announcement today...

💬0🔄0❤️1👀537

⚡ Powered by xgo.ing

🔬 The Lab of the Future Should Feel Like a Data Center with Andy Beam & Rafa Gómez-Bombarelli, Lil...

Latent.Space (@latentspacepod) — Thu, 16 Jul 2026 20:32:24 GMT

🔬 The Lab of the Future Should Feel Like a Data Center  with Andy Beam & Rafa Gómez-Bombarelli, Lila Scienceslatent.space/p/the-lab-of-t…A

@LilaSciences CTO @AndrewLBeam and CSO Rafa Gómez-Bombarelli explain why science is the last untapped source of internet-scale training data, how AI Science Factories turn wet labs into 24/7 token generators, what’s inside 10 trillion experimentally validated reasoning tokens, why breadth across biology, chemistry, and materials can produce better general reasoners, and what it will take to move from superhuman test-taking to genuinely creative scientific discovery.

💬1🔄2❤️11👀3156📊5

⚡ Powered by xgo.ing

9M now https://t.co/4geDlFbdYC

Latent.Space (@latentspacepod) — Wed, 15 Jul 2026 16:12:58 GMT

9M now x.com/petergostev/st…

Peter Gostev@petergostev

Ok my previous chart wasn't quite right, this is the right trajectory

🔗 View Quoted Tweet

💬1🔄0❤️6👀2186📊2

⚡ Powered by xgo.ing

For those in SF - we're hosting the epic return of our Personal AI meetup this week, hosted by @swyx...

Latent.Space (@latentspacepod) — Wed, 15 Jul 2026 04:04:18 GMT

For those in SF - we're hosting the epic return of our Personal AI meetup this week, hosted by @swyx and @shloked:

luma.com/personai

come if you are building in Personal AI - people with demos prioritized! max 50 spots only, this is not one of those slopfests.

💬1🔄0❤️15👀20006📊5

⚡ Powered by xgo.ing

8M now https://t.co/HsAYnpxxj5

Latent.Space (@latentspacepod) — Wed, 15 Jul 2026 03:33:35 GMT

8M now
x.com/thsottiaux/sta…

Tibo@thsottiaux

Hello. We have reached 8M active users across Codex and ChatGPT Work.

We are once again resetting the usage limits for all. And we continue to not have the 5h rate limit as well, allowing everyone to explore the boundaries of GPT-5.6 Sol and discover how ambitious you can be.

See you tomorrow for more updates on our growth!

🔗 View Quoted Tweet

💬2🔄1❤️21👀5554📊5

⚡ Powered by xgo.ing

🆕 5 Trends That Defined AI Engineering at World’s Fair 2026 https://t.co/qcB7lHSBI1 @ricmac's big...

Latent.Space (@latentspacepod) — Tue, 14 Jul 2026 23:22:58 GMT

🆕 5 Trends That Defined AI Engineering at World’s Fair 2026
latent.space/p/aiewf26trends1

@ricmac's big recap of @aidotengineer:

1. The focus shifts from agents to systems
2. Loop engineering is the new control layer
3. Al engineering enters the enterprise
4. Coding agents replace IDEs
5. Every agent platform is building around skills

💬4🔄6❤️12👀3339📊10

⚡ Powered by xgo.ing

Did... Codex just overtake Claude Code? 24.5 hours ago Tibo announced 6M active users. this means ...

Latent.Space (@latentspacepod) — Tue, 14 Jul 2026 01:25:40 GMT

Did... Codex just overtake Claude Code?

24.5 hours ago Tibo announced 6M active users.

this means Codex usage jumped 1M in ~ONE DAY.

the last user number we heard from Claude Code was 2M in Feb: latent.space/p/ainews-codex…

more analysis within, but this is very big if true.

Tibo@thsottiaux

Thank you to the 7M active users who are now using Codex and ChatGPT Work.

We have added a banked reset to everyone's account to celebrate the milestone. You can apply the reset in the desktop app or on web and it will replenish the weekly usage for you.

Have fun out there.

🔗 View Quoted Tweet

💬28🔄17❤️210👀209155📊57

⚡ Powered by xgo.ing

Youtube: https://t.co/hGVabn38zM

Latent.Space (@latentspacepod) — Mon, 13 Jul 2026 18:22:01 GMT

Youtube: youtube.com/watch?v=jhpmMT…

💬0🔄0❤️2👀1388📊1

⚡ Powered by xgo.ing

In this episode, @EngramLab co-founder and CEO @dan_biderman joins @allenpark to cook Mediterranean ...

Latent.Space (@latentspacepod) — Mon, 13 Jul 2026 18:22:00 GMT

In this episode, @EngramLab co-founder and CEO @dan_biderman joins @allenpark to cook Mediterranean meatballs with yellow rice and talk about building AI that actually learns from you: why long context, RAG, and compaction eventually break down, how Engram compresses knowledge into cartridges and model weights, what continual learning could unlock for long-horizon agents, why token efficiency is inseparable from intelligence, how personal models could improve like Tamagotchis, and what it takes to build the research and infrastructure for millions of continuously updated AI memories.

Timestamps:
00 Intro
26 Engram’s $98M Launch and Meatballs
45 From Naval Special Operations to AI Research
32 Israeli Military Culture and Founder Maturity
12 Why Engram Is Betting on Context and Continual Learning
14 Knowledge Cartridges, Compression, and Model Intuition
10 Trillion-Token Company Knowledge and Context Rot
05 Long-Context Limits, Compaction, and Neural Memory
20 Test-Time Training and “Destroying Prefill”
31 Harvey and Holistic Enterprise Queries Beyond RAG
02 Personal AI Models and Tamagotchi Weights
00 What Belongs in Weights vs. Text
25 Autonomous Memory and User-Specific Feedback Loops
20 Token Efficiency, Model Routing, and Harder Tasks
03 Engram’s Research Team and Product Culture
02 Hiring Researchers and Infrastructure Engineers
25 Doing More With Less
41 Where to Find Engram
19 Final Taste Test

Your browser does not support the video tag.

🔗 View on Twitter

💬4🔄5❤️37👀34580📊16

⚡ Powered by xgo.ing

@modal @akshat_b thanks for the love!! https://t.co/nYJN5ykbyw

Latent.Space (@latentspacepod) — Mon, 13 Jul 2026 16:28:18 GMT

@modal@akshat_b thanks for the love!! x.com/mstockton/stat…

Matt Stockton@mstockton

The recent @latentspacepod talking about Modal w/ @akshat_b is fantastic. Modal is an amazing technology to use. It just works and it is extremely well designed. They 100% get dev ergonomics correct.

I use it for a lot of different things, and honestly use it for things that go 'against the grain' of their core positioning / offering. Turns out that works well too :)

Wrote a bit about how I am using their services here:
mattstockton.com/2026/07/12/how…

🔗 View Quoted Tweet

💬0🔄0❤️1👀787

⚡ Powered by xgo.ing

Modal's Agent-Native Cloud: DX→AX, sandboxes, elastic inference, and 100,000 rollouts https://t.co/M...

Latent.Space (@latentspacepod) — Thu, 09 Jul 2026 00:00:01 GMT

Modal's Agent-Native Cloud: DX→AX, sandboxes, elastic inference, and 100,000 rollouts latent.space/p/modal2026@modal CTO @akshat_b explains why developer experience is becoming agent experience, why agents need infra they can operate instead of YAML they have to reason through, how sandboxes turn the agent loop into something real, why elastic inference and GPU snapshotting matter for production AI, how RL rollouts can require 100,000 sandboxes, and why Modal’s $355M Series C marks a new phase for AI-native cloud infrastructure.

Your browser does not support the video tag.

🔗 View on Twitter

💬5🔄5❤️34👀15946📊14

⚡ Powered by xgo.ing

tldr of thariq's AIEWF keynote https://t.co/YagrwF1a71 (more in today's ainews issue)

Latent.Space (@latentspacepod) — Tue, 07 Jul 2026 04:45:56 GMT

tldr of thariq's AIEWF keynote

latent.space/p/ainews-the-f…

(more in today's ainews issue)

Thariq@trq212

my keynote at AI Engineer World Fair: “A Field Guide to Fable” is live on YouTube!

youtu.be/9fubhllmsBU?is…

🔗 View Quoted Tweet

💬3🔄3❤️44👀16935📊8

⚡ Powered by xgo.ing

The AI Engineer World’s Fair ended with a debate about loops, a report on the state of AI engineerin...

Latent.Space (@latentspacepod) — Fri, 03 Jul 2026 05:13:53 GMT

The AI Engineer World’s Fair ended with a debate about loops, a report on the state of AI engineering, and closing keynotes focused on what to build next.

Also, locomotives! (hat-tip @GeoffreyHuntley) 

latent.space/p/aiewf-daily-…

💬1🔄9❤️25👀3343📊9

⚡ Powered by xgo.ing

In this interview at @aiDotEngineer World's Fair, @vercel chief of software @andrewqu explains why a...

Latent.Space (@latentspacepod) — Fri, 03 Jul 2026 00:41:02 GMT

In this interview at @aiDotEngineer World's Fair, @vercel chief of software @andrewqu explains why agents represent a new form of software, what Vercel learned from building its own, and why Vercel itself is turning into an agent! latent.space/p/vercel-agent…

💬2🔄2❤️25👀8124📊9

⚡ Powered by xgo.ing

Adobe is experimenting with “agentic sites” that generate pages around an individual user’s intent. ...

Latent.Space (@latentspacepod) — Thu, 02 Jul 2026 21:26:36 GMT

Adobe is experimenting with “agentic sites” that generate pages around an individual user’s intent. At @aiDotEngineer World's Fair, we talked to Carlos Sanchez about the Web's future. latent.space/p/the-website-…

💬0🔄1❤️4👀1622📊2

⚡ Powered by xgo.ing

Paul Bakaus (@pbakaus) from @impeccable_ai talks to us about skill engineering, human judgment in a ...

Latent.Space (@latentspacepod) — Thu, 02 Jul 2026 14:37:58 GMT

Paul Bakaus (@pbakaus) from @impeccable_ai talks to us about skill engineering, human judgment in a 'loopmaxxing' era, and why agents still need people to steer them. latent.space/p/skill-engine…

💬3🔄3❤️11👀1926📊7

⚡ Powered by xgo.ing

If you're into loopmaxxing, there were plenty of pro and con arguments about it at the @aiDotEnginee...

Latent.Space (@latentspacepod) — Thu, 02 Jul 2026 13:50:39 GMT

If you're into loopmaxxing, there were plenty of pro and con arguments about it at the @aiDotEngineer World's Fair yesterday. Here's my roundup of the talking points, including comments from @pbakaus & @addyosmani, as we head into day 4 of the conference. https://t.co/t7uVVcVEqr

💬0🔄1❤️5👀2075📊3

⚡ Powered by xgo.ing

Today at the AI Engineer World's Fair in San Francisco: the 'software factory' vision met resistance...

Latent.Space (@latentspacepod) — Thu, 02 Jul 2026 06:17:06 GMT

Today at the AI Engineer World's Fair in San Francisco: the 'software factory' vision met resistance from speakers defending human understanding and control. latent.space/p/aiewf-daily-…

💬2🔄3❤️8👀3448📊5

⚡ Powered by xgo.ing

Your humble Latent Space contributor in the front row, to the right in this pic. What a crowd today!

Latent.Space (@latentspacepod) — Thu, 02 Jul 2026 04:44:15 GMT

Your humble Latent Space contributor in the front row, to the right in this pic. What a crowd today!

Thariq@trq212

hello from AI engineer!

🔗 View Quoted Tweet

💬0🔄0❤️17👀2767📊2

⚡ Powered by xgo.ing

In our latest interview from AIEWF, @IntrospectionAI co-founder @rolandgvc explains autoresearch, ag...

Latent.Space (@latentspacepod) — Wed, 01 Jul 2026 23:54:33 GMT

In our latest interview from AIEWF, @IntrospectionAI co-founder @rolandgvc explains autoresearch, agent “recipes,” self-improving loops, and why humans remain central to the software factory. latent.space/p/autoresearch…

💬1🔄2❤️8👀1494📊4

⚡ Powered by xgo.ing

More loops in this morning’s AIEWF keynotes. Anthropic and Sonar getting in on the buzzword bingo.

Latent.Space (@latentspacepod) — Wed, 01 Jul 2026 20:08:51 GMT

More loops in this morning’s AIEWF keynotes. Anthropic and Sonar getting in on the buzzword bingo.

💬0🔄1❤️4👀1329📊1

⚡ Powered by xgo.ing

Cursor's Pauline Brunet explains how her team of Forward Deployed Engineers help organizations imple...

Latent.Space (@latentspacepod) — Wed, 01 Jul 2026 19:05:08 GMT

Cursor's Pauline Brunet explains how her team of Forward Deployed Engineers help organizations implement agents — essentially setting up software factories. latent.space/p/cursor-forwa…

💬0🔄1❤️6👀1339📊3

⚡ Powered by xgo.ing

I'm preparing for another mind-expanding day 🤯 at the AI Engineer World's Fair. While we wait for d...

Latent.Space (@latentspacepod) — Wed, 01 Jul 2026 14:42:55 GMT

I'm preparing for another mind-expanding day 🤯 at the AI Engineer World's Fair. While we wait for day 3 keynotes, here's my roundup of all the action yesterday — including loops, software factories, forward deployed engineers, open source AI. (@ricmac)latent.space/p/aiewf-daily-…I

💬0🔄0❤️0👀1191📊1

⚡ Powered by xgo.ing

Warp's founder @zachlloydtweets thinks every major software project will soon run on an automated "f...

Latent.Space (@latentspacepod) — Wed, 01 Jul 2026 14:30:59 GMT

Warp's founder @zachlloydtweets thinks every major software project will soon run on an automated "factory." I caught up with Lloyd on the expo floor of the @aiDotEngineer World's Fair, to discuss why & how engineers should prepare for this shift. https://t.co/1cfqSsxR1E

💬2🔄7❤️17👀3342📊8

⚡ Powered by xgo.ing

My report from day 2 of the @aiDotEngineer World's Fair. There was lots of talk about loops, agent e...

Latent.Space (@latentspacepod) — Wed, 01 Jul 2026 04:49:42 GMT

My report from day 2 of the @aiDotEngineer World's Fair. There was lots of talk about loops, agent engineering, and the emergence of software factories. Also a hot topic: open source AI. (by @ricmac) latent.space/p/aiewf-daily-…

💬1🔄3❤️13👀3322📊5

⚡ Powered by xgo.ing

Sierra's Natalie Meurer on why product engineers and forward deployed engineers are starting to conv...

Latent.Space (@latentspacepod) — Wed, 01 Jul 2026 00:22:52 GMT

Sierra's Natalie Meurer on why product engineers and forward deployed engineers are starting to converge. latent.space/p/forward-depl…

💬0🔄1❤️3👀1448📊2

⚡ Powered by xgo.ing

After two packed AIEWF workshops, @TheAhmadOsman makes the case that local AI is catching up fast — ...

Latent.Space (@latentspacepod) — Tue, 30 Jun 2026 23:42:23 GMT

After two packed AIEWF workshops, @TheAhmadOsman makes the case that local AI is catching up fast — from laptops and phones to enterprise-grade infrastructure. latent.space/p/ahmad-osman-…

💬0🔄1❤️1👀992📊1

⚡ Powered by xgo.ing

Word of the day so far at AIEWF is Loop. @swyx talked about “loopcraft” in his opening address, and ...

Latent.Space (@latentspacepod) — Tue, 30 Jun 2026 17:05:01 GMT

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.

💬3🔄4❤️14👀2768📊6

⚡ Powered by xgo.ing

Nearly time for the @aiDotEngineer World’s Fair keynotes. Cavernous was the word that came to mind w...

Latent.Space (@latentspacepod) — Tue, 30 Jun 2026 15:36:50 GMT

Nearly time for the @aiDotEngineer World’s Fair keynotes. Cavernous was the word that came to mind when I walked into the keynote hall. It’s massive and the AIE signage is impressive. You can follow the keynotes online on YouTube: youtube.com/live/htM02KMNZ…

💬1🔄1❤️8👀2236📊3

⚡ Powered by xgo.ing

In 2023 I attended the very first AI Engineer Summit in San Francisco, with 500 attendees. I took th...

Latent.Space (@latentspacepod) — Tue, 30 Jun 2026 04:05:36 GMT

In 2023 I attended the very first AI Engineer Summit in San Francisco, with 500 attendees. I took this pic of @swyx back then. This week I'm at the AI Engineer World's Fair in SF, with 7,000 attendees! 🚀🚀🚀 I'll be blogging it all latent.space2kE (@ricmac)

💬1🔄0❤️29👀2892📊4

⚡ Powered by xgo.ing

The @aiDotEngineer World’s Fair kicks off today in San Francisco. A lot of AI-pilled humans and huma...

Latent.Space (@latentspacepod) — Tue, 30 Jun 2026 01:08:56 GMT

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)

💬5🔄3❤️17👀4979📊9

⚡ Powered by xgo.ing

How close is AGI? @OpenAI Chief Research Officer @markchen90 discusses what the future of model cap...

Latent.Space (@latentspacepod) — Fri, 26 Jun 2026 23:00:01 GMT

How close is AGI?

@OpenAI Chief Research Officer @markchen90 discusses what the future of model capabilities looks like: "We're getting closer and closer to a world where the models can come up with more of the innovation on their own."

Your browser does not support the video tag.

🔗 View on Twitter

Latent.Space@latentspacepod

In this episode, @OpenAI Chief Research Officer @markchen90 joins @allenpark to flambé shrimp, cook Korean stew, and chat about being at the frontier of AI research: why scaling laws and pre-training still matter, how OpenAI chooses research bets and allocates compute, what it means to develop research taste, why evals are in crisis, how to avoid benchmark-maxing, and what it will take for models to handle long-horizon real-world work, multimodal reasoning, and eventually end-to-end AI research.

Timestamps:
00 Intro
28 The Soup Story
52 From Trading to AI Research
21 How to Develop Research Taste
23 RL, Evals, and Superhuman Benchmarks
17 Cooking Begins on the Impulse Stove
53 Scaling Laws, Pre-Training, and Reasoning
33 OpenAI’s Research Roadmap and Compute Allocation
48 What Makes a Great Researcher
33 The Evals Crisis and Benchmark-Maxing
34 Jagged Intelligence, Context, and Long-Horizon Learning
14 Shrimp Flambé and New Research Bets
32 Multimodal Models and One Architecture
36 Vibe Researching and End-to-End AI Research
36 Failed Bets, Postmortems, and OpenAI’s Alpha
07 Final Taste Test
53 Overrated vs. Underrated AI Research
00 Closing

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬5🔄4❤️41👀10492📊11

⚡ Powered by xgo.ing

Is pre-training dead? @OpenAI Chief Research Officer @markchen90 doesn't think so: "We've always fo...

Latent.Space (@latentspacepod) — Fri, 26 Jun 2026 19:00:35 GMT

Is pre-training dead?

@OpenAI Chief Research Officer @markchen90 doesn't think so: "We've always found some kind of technique whether it be better engineering or some new research insight that helps you break past the boundary."

Your browser does not support the video tag.

🔗 View on Twitter

Latent.Space@latentspacepod

In this episode, @OpenAI Chief Research Officer @markchen90 joins @allenpark to flambé shrimp, cook Korean stew, and chat about being at the frontier of AI research: why scaling laws and pre-training still matter, how OpenAI chooses research bets and allocates compute, what it means to develop research taste, why evals are in crisis, how to avoid benchmark-maxing, and what it will take for models to handle long-horizon real-world work, multimodal reasoning, and eventually end-to-end AI research.

Timestamps:
00 Intro
28 The Soup Story
52 From Trading to AI Research
21 How to Develop Research Taste
23 RL, Evals, and Superhuman Benchmarks
17 Cooking Begins on the Impulse Stove
53 Scaling Laws, Pre-Training, and Reasoning
33 OpenAI’s Research Roadmap and Compute Allocation
48 What Makes a Great Researcher
33 The Evals Crisis and Benchmark-Maxing
34 Jagged Intelligence, Context, and Long-Horizon Learning
14 Shrimp Flambé and New Research Bets
32 Multimodal Models and One Architecture
36 Vibe Researching and End-to-End AI Research
36 Failed Bets, Postmortems, and OpenAI’s Alpha
07 Final Taste Test
53 Overrated vs. Underrated AI Research
00 Closing

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬3🔄3❤️125👀29216📊24

⚡ Powered by xgo.ing

In this episode, @OpenAI Chief Research Officer @markchen90 joins @allenpark to flambé shrimp, cook ...

Latent.Space (@latentspacepod) — Thu, 25 Jun 2026 21:41:50 GMT

In this episode, @OpenAI Chief Research Officer @markchen90 joins @allenpark to flambé shrimp, cook Korean stew, and chat about being at the frontier of AI research: why scaling laws and pre-training still matter, how OpenAI chooses research bets and allocates compute, what it means to develop research taste, why evals are in crisis, how to avoid benchmark-maxing, and what it will take for models to handle long-horizon real-world work, multimodal reasoning, and eventually end-to-end AI research.

Timestamps:
00 Intro
28 The Soup Story
52 From Trading to AI Research
21 How to Develop Research Taste
23 RL, Evals, and Superhuman Benchmarks
17 Cooking Begins on the Impulse Stove
53 Scaling Laws, Pre-Training, and Reasoning
33 OpenAI’s Research Roadmap and Compute Allocation
48 What Makes a Great Researcher
33 The Evals Crisis and Benchmark-Maxing
34 Jagged Intelligence, Context, and Long-Horizon Learning
14 Shrimp Flambé and New Research Bets
32 Multimodal Models and One Architecture
36 Vibe Researching and End-to-End AI Research
36 Failed Bets, Postmortems, and OpenAI’s Alpha
07 Final Taste Test
53 Overrated vs. Underrated AI Research
00 Closing

Your browser does not support the video tag.

🔗 View on Twitter

💬11🔄13❤️209👀62110📊83

⚡ Powered by xgo.ing

Why the Frontier Ecosystem must be Open — Matei Zaharia and Reynold Xin, Databricks https://t.co/lSn...

Latent.Space (@latentspacepod) — Wed, 24 Jun 2026 18:59:09 GMT

Why the Frontier Ecosystem must be Open — Matei Zaharia and Reynold Xin, Databricks latent.space/p/databricks@databricks cofounders @matei_zaharia and @rxin  explain why Databricks is moving into the infrastructure layer for enterprise agents, how Omnigent creates a shared harness for coding agents and custom agents, why LTAP and Lakebase rethink the split between operational and analytical databases, why agent security needs contextual policies and spend controls, and why the future of software may be as simple as getting the right data in place and putting agents on top.

Your browser does not support the video tag.

🔗 View on Twitter

💬4🔄25❤️82👀61946📊36

⚡ Powered by xgo.ing

@claudeai https://t.co/keN9LYE519

Latent.Space (@latentspacepod) — Wed, 24 Jun 2026 07:16:07 GMT

@claudeailatent.space/p/ainews-claud…

💬0🔄0❤️0👀880