Latent.Space 🔜 @aiDotEngineer
@latentspacepod
Followers
17K
Following
1K
Media
214
Statuses
1K
The #1 AI Engineering podcast & newsletter. Technical insights and news today you will use at work tomorrow! Hosted by @swyx and @fanahova
Joined June 2023
🆕 Presenting: . The 2025 AI Engineering Reading List . . 1 paper/blog/model family per week for every week of 2025, for you to run paper clubs or binge over the break.
13
248
1K
🆕 NeurIPS 2023 Recap — Best Papers!. Our selection for AI Engineers this year includes:. - Word2Vec (@JeffDean).- DPO (@rm_rafailov).- Datablations (@Muennighoff).- QLora (@Tim_Dettmers).- DataComp (@sy_gadre).- LLaVA (@imhaotian).- Tree of Thoughts.
4
67
295
🆕 post: OpenAI Realtime API: The Missing Manual. Everything we learned, and everything we think you need to know, from technical details on 24khz/G.711 audio, RTMP, HLS, WebRTC, to Interruption/VAD, to Cost, Latency, Tool Calls, and Context Mgmt. Enjoy this first guest post from
Great demo by @swyx 🤯 of creating a game using voice commands + agents using . @ericsimons40 you should see this!. @stackblitz . @OpenAI Dev Day Singapore
6
43
278
🆕 post: Why MCP Won. lessons from @dsp_, @alexalbert__, @sebmarkbage, @paulgauthier, and many more. 1. MCP is “AI-Native” version of old idea.2. MCP is an “open standard” with a big backer.3. Anthropic has the best developer AI brand.4. MCP based off
The highly-anticipated MCP workshop with @MaheshMurag is now live!. Join @AnthropicAI Applied AI Engineer (+ creator of MCP) @MaheshMurag in this full 2-hour workshop covering the Model Context Protocol, where he covers:. 00:00 Introduction.00:35 What is MCP?.09:39 Building with
14
46
256
What's the #1 trick to massively stand out in your career?. @jeremyphoward drops his #1 tip. The bar has been set
6
38
188
🆕 short essay: Why GPT Wrappers Are Good, Actually.
5
22
175
🆕 The New Kings of Open Source AI. Recapping how @MistralAI took over Open Source AI, how the definition of Open is evolving, and why engineers freaked out about Copilot losses at @aidotengineer. Memes of the month: @itsmingjie @nearcyan. ft @TheNoahHein
3
21
167
🆕 Production AI Engineering starts with Evals. A 2 hour deep dive on the state of the LLM Ops industry with @ankrgyl following the @braintrustdata Series A! . We discuss:. why @HamelHusain was right: Evals are at the center of the production AI
6
17
136
🆕 The Great Debate: OpenAI vs LangGraph. The silent war in Agent Engineering gets loud.
OpenAI recently released a guide on building agents which contains some misguided takes. There's a lot of FUD, confusion, hype, and noise around agents. I wrote a blog on how to think about agent frameworks. Includes:. Background Info.- What is an agent?.- What is hard about
2
22
135
🆕 the one and only @jxnlco on:. • Why Pydantic is -still- All You Need.• Why most AI frameworks aren't VC backable.• Why you'd be happier being bootstrapped.• Why you should NOT hire ML Engineers.• the definitive origin of Instructor. (and bonus.
3
16
130
🆕 pod: The Yolo Researcher Metagame with @YiTayML!. OpenAI (ca. GPT4): ~600 people.Google Gemini: ~950 coauthors.@RekaAILabs: 20 people. @sama once speculated on the qualities of "10,000x AI researchers", and more recently @_jasonwei described the "Yolo
6
20
123
first run of "What is an AI Engineer?". Experimental new short form video format we are trying out - didn't publish on yt bc need to rehearse more but feedback welcome. (thanks to @lizziepika for the lightboard!!)
14
10
106
The Six Elements of Agent Engineering. compiling answers from @simonw's definitions thread (thanks @lilianweng @ruthefordml @hwchase17 @DanielMiessler @Dan_Jeffries1 @romechenko). IMPACT: .- Intent (Instructions + Evals).- Memory (@gneubig reusable workflows).- Planning (Editable
🆕 talk + essay: Agent Engineering. Why we went all in on Agents @aiDotEngineer.Defining Agents (thanks to @simonw).The Six Elements of Agent Engineering.Why Agents are ChatGPT's path to 1B MAU.Why work on Agent Engineering Now. live on @latentspacepod!.
5
12
110
🆕 The Busy Person's Intro to Finetuning & Open Source AI. with @winglian of Axolotl . Covering the SF AI meetup with @NousResearch, @Teknium1, and all the required knowledge to get started navigating Open Source AI and finetuning them. (special cohost.
2
23
106
🆕 post: Building AI for the Enterprise. The long-awaited third (!) guest post from @_anshulr on how to make $$$ with AI: Build with from the start with an Enterprise Infrastructure Native mindset!. Now. Not later. Premature optimization is not premature
Today we’re excited to launch the Windsurf Editor - the first agentic IDE, and then some 🏄. In Windsurf, we have given the AI a previously unseen combination of deep codebase understanding, powerful set of tools, and real time awareness of your in-editor actions. The result? A
3
5
105
🆕 2025's biggest surprise so far: . Reasoning is less of a moat than anyone thought.
1
19
89
Our last keynote:. Recapping 2024 in Agents. we save the most ambitious talk for last - we saked @gneubig, creator of the reigning #1 agent on SWE-Bench Full, to recap everything relevant to building agents in 2024, and boy did he deliver. Timestamps:. 00:00 Welcome to Latent
Thanks a bunch to @swyx and @latentspacepod for the invite!. Here are my (opinionated) slides on "Agents in December 2024": And for those who want evidence that my demo was real, here's the PR the agent wrote while I was talking 😃.
3
21
86
🆕 The Ultimate Guide to Prompting. with @sanderschulhoff of @LearnPrompting and The Prompt Report!. Timestamps.[00:00:00] Introductions.[00:07:32] Navigating arXiv for paper evaluation.[00:12:23] Taxonomy of prompting techniques.[00:15:46] Zero-shot
3
14
87
🆕 The new Claude 3.5 Sonnet, Computer Use, and Building SOTA Agents. with @ErikSchluntz, our first @AnthropicAI guest!. ---. The vibe shift we observed in July - in favor of Claude 3.5 Sonnet, first introduced in June — has been remarkably long lived and persistent, surviving
2
9
84
🆕 pod: The Inventors of Deep Research!. While everyone cloning Deep Research, we asked @AarushSelvan and Mukund Sridhar, the original PM and Tech Lead who created the newest killer use case of AI Agents now copied by @openai, @xai, @perplexity_ai and a.
6
15
81
THE RISE AND FALL OF THE VECTOR DB CATEGORY. 27min lightning pod, doubleclicking on this hit essaywith @jobergum. live now:. Chapters. 00:00 Introduction to Trondheim and Background.03:03 The Rise and Fall of Vector Databases.06:08 Convergence of Search
4
5
79
🆕 pod: From API to AGI: Structured Outputs, OpenAI API platform and O1 Q&A. Our @openai weekend special!. - 1hr convo with @michpokrass on Structured Outputs, ChatGPT-latest, gpt-4o, Assistants API, Voice Mode and ALL your other API questions and war
2
8
73
Prediction: "In the next 2 years, every LLM will become LMMs" (@chipro's term for Large Multimodal Models). 🆕: How to train your own Large Multimodal Model. with @HugoLaurencon and @LeoTronchon of @Huggingface!. We follow how HuggingFace's M4 team
1
18
71
🆕 pod with @ThomasScialom of @AIatMeta!. Llama 2, 3 & 4: Synthetic Data, RLHF, Agents on the path to Open Source AGI. shoutouts:.- Why @ylecun's Galactica Instruct would have solved @giffmana's Citations Generator.- Beyond Chinchilla-Optimal: 100x.
4
16
67
Announcing the AI Engineer Summit Online Track!. >30 talks. 9 hours of extra content. Free. Async-First. Keynotes from.- @klarnaseb .- @chipro.- @bernhardsson . Feb 22nd, 9am ET - 4pm ET, FREE on @aiDotEngineer youtube!. "Have You Tried Rubbing LLMs On It?". - Lessons Learned
2
12
68
🆕 The Creators of Model Context Protocol. with @dsp_ and @jspahrsummers!. We asked ALL your burning questions:.- The Origin Story of MCP.- MCP vs OpenAPI.- Building Agents with MCP.- How many MCPs is too many?.- Authorization and Trust in MCP Servers.-.
5
13
69
⚡️ Building Manus AI - the first ever talk from Manus AI cofounder Tao Zhang.
0
4
63
🆕 $2 H100s: How the GPU Rental Bubble Burst. A rare guest post, from returning guest @picocreator!. H100s used to be $8/hr (if you could get them). Now there's 7 different places sometimes selling them under $2. What happened?. @nvidia used to have
6
18
59
🆕 The Accidental AI Canvas. The full story of how @steveruizok took @gdb's GPT4V Demo to its logical conclusion, and accidentally invented the future of multimodal prompting with @tldraw. Now on YouTube + wherever fine podcasts are sold.
0
22
56
🆕 The Winds of AI Winter. The vibes have shifted. On @leopoldasch vs Sequoia, Goldman Sachs, @benedictevans, @cpaik's End of Software and why the AI Engineer rises above it all. The future is here, but it is not evenly distributed.
2
7
55
🆕 The new OpenAI Agents Platform. @OpenAIDevs are refreshing the entire suite of APIs, Tools, and SDKs for the year of Agents. We grab an exclusive interview with @romainhuet and @nikunjhanda to ask your burning questions on the Responses API, Web.
2
15
56
"Why does Production AI Engineering start with Evals?". the AI Engineer SDLC, silly. next in our @lizziepika lightboard series of short explainers (again, need to reshoot with better markers and idea of where the camera ends, but just trying out this new format). h/t to @doppenhe
🆕 Production AI Engineering starts with Evals. A 2 hour deep dive on the state of the LLM Ops industry with @ankrgyl following the @braintrustdata Series A! . We discuss:. why @HamelHusain was right: Evals are at the center of the production AI
0
8
57
🆕 Building AI for The People. Never has so much been shipped for so many by so few. @jeremyphoward is back on the pod! sharing the founding journey of @AnswerAI, predicting the @OpenAI governance crisis, hiring 1000x researchers and developers like.
1
12
51
🆕 pod: How to train a Million Context LLM!. @ylecun says we should publish, or perish. We asked @markatgradient to spill ALL the beans on how his team extended Llama-3 to 1M+ context with ~perfect @GregKamradt NIAH evals!. Also covering:. - A Brief.
2
11
48
🆕 The State of AI Startups. with @saranormous and @prnvrdy!. Proud to share the opening keynote of Latent Space LIVE! at @NeurIPSConf! . We'll be in your ears through the holidays helping to recap and celebrate the best of 2024 in each domain! It's been
fun to do a last minute “state of ai startups” talk (w/ @prnvrdy) this am at #neurips for old friend @swyx at his live paper-reading club. couple quick takeaways below on what’s working, where we disagree with the dominant narratives, and some of our latent priors in investing👇
3
10
48
🆕 Bret Taylor: The AI Architect. Presenting our conversations with @btaylor, the legendary CEO of @SierraPlatform, Chairman of @OpenAI, and creator of Google Maps/Facebook Likes on the future of Software Engineering, and building great products and teams.
4
11
46
🆕 How to Run a Paper Club!. Guest post from @eugeneyan celebrating >100 papers read over an >18 months unbroken streak for the Latent Space Paper Club! . - The First Rule of Paper Club.- Inside a Single Paper Club meeting.- How to Scale a Paper Club from >100 papers.- How to
1
3
47
🆕How to Hire AI Engineers. a rare guest post (and bonus pod) from @james_elicit and @_adamwiggins_! Covering:. - Defining the Hiring Process.- Defensive AI Engineering as a chaotic medium.- Tech Choices for Defensive AI Engineering.- How do you Interview for Defensive AI
1
6
42
🆕 pod: with @BEBischof of @_hex_tech!. On putting AI Magic into Notebooks, and how RAG is actually a recommendation systems problem. Also an @AAAzzam hot take: LLMOps is more like an "iron mine" than a "gold rush"!
3
9
43
Are GPU clouds intentionally losing money on last month's Mixtral inference price war?. Probably. Depends on batching and aggressiveness of accounting. As @NaveenGRao puts it - a lot of people are about to learn the meaning of "TCO".
🆕 The Four Wars of the AI Stack. Our Dec 2023 recap also includes a framework for looking at the key business battlegrounds of all of 2023:. In Data: . with OpenAI announcing a partnership with Axel Springer (see also its deal with the AP and its Data
6
4
45
🆕 ICLR 2024: Best Papers (Part 1). We present our selections of outstanding papers and talks thematically introducing topics for AI Engineers to track:. Section A: ImageGen, Compression, Adversarial Attacks. - Auto-encoding Variational Bayes (@dpkingma).- Würstchen: An Efficient
2
9
45
"You shouldn't even be hiring MLEs". @jxnlco on how companies are unnecessarily hiring MLEs when they should be looking to the under-utilized but motivated engineers that have a desire to pursue AI Engineering.
2
9
45
🆕 "What if you could have 1000s of Claudes spinning up at once?". We asked @_catwu and @bcherny ALL the questions you ever had about @anthropicai's Claude Code - including its origin story, cost ("It's like a wood chipper fueled by dollars" -.
1
13
44
The 6 "modes" of LLM inference, over time:. - 2021: "large" models.- 2023: "turbo"/"mini" models.- Apr 2024: Batch API.- Sep 2024: Reasoning models.- Oct 2024: Realtime API.- Nov 2024: Speculative Decoding APIs. seems pretty comprehensive. what else will be coming?
[4 Nov 2024]. OpenAI beats Anthropic and Fireworks to releasing Speculative Decoding.
2
4
43
## Recapping 2024 in Synthetic Data/Smol Models!. We are very honored to have @LoubnaBenAllal1 recap and pick all the best papers in:. Synthetic Data. and Smol Models. this year!. Timestamps. [00:00:05] Loubna Intro.[00:00:33] The Rise of Synthetic Data Everywhere.[00:02:57]
2
7
43
🆕ChatGPT Codex: The Missing Manual. ChatGPT Codex is here - the first cloud hosted Autonomous Software Engineer (A-SWE) from OpenAI. Josh Ma and @embirico tell us how to WHAM every codebase like a power user!
3
5
41
🆕 pod on Transformer Math 101!. with @QuentinAnthon15 of @AiEleuther. everything you ever wanted to know about the nitty gritty of training and making the gpu go brrrrrrrrrrrr
The one thing more valuable than GPUs is knowing how to use them (aka make them go brrrr 🔥). @QuentinAnthon15 of @AiEleuther, co-author of "Transformers Math 101", joined @swyx and I on @latentspacepod 🎙️ Take your notepads out before pressing play! We.
0
7
34
🆕 ⚡️ pod: Post-training Distillation for LLMs!. We asked Rishi to walk us thru his popular lecture given at @GoogleDeepMind. Now public for the first time ever on the LS YouTube!!
I recently gave a tutorial on knowledge distillation for LLMs, explaining the mathematical derivations behind the commonly used methods. Sharing the slides here given the recent interest in this topic.
2
8
37
Here is an 18min podcast about our podcast with @hwchase17 and @ShunyuYao12 from NotebookLM Deep Dive. let us know if you like this better than the. yknow. actual podcast. because then we dont have a job anymore ig
How do we make AI agents think and act? 🤖. Today's episode with @ShunyuYao12 (and special cohost @hwchase17!) is probably our best agents episode so far:.- Origins of ReAct and how it inspired @LangChainAI .- Agent-Computer Interfaces and tool design.- How to benchmark AI agents
4
9
36
🆕 Thank you for a great 2024!. Presenting our 2024 year in review: NeurIPS takeaways, Biggest Themes of the Year, the Four Wars of AI, and a month-by-month recap of the Year in AI. 00:00 Welcome to the 100th Episode!.00:19 Reflecting on the Journey.00:47 AI Engineering: The
5
6
35
🆕 SF Compute: Commoditizing Compute. We're excited for our latest deep dive into the compute market with @evanjconrad of @sfcompute!. It should not be normal for the prices of one of the world’s most important resources right now to swing from $8 to $1.
1
7
37
🆕 We Are Running Out of Low-Background Tokens. Was 2022 the last year we had "clean" data uncontaminated by ChatGPT?. (h/t @jeffreyhuber for the concept!). also:.- @martin_casado's Voice Cloning Red Alert.- that @pika_labs video.- incredible LCM demos
2
13
33
“Cursor is the best product I've used in a while” - @MacCaw. “It's so elegant and easy.” - @AndrewMcCalip. “Coding with AI is getting insane.” - @MckayWrigley. The Latent Space pod is proud to present: the first podcast with @amanrsanger of @anysphere!.
4
9
35
🆕 Our first NSFW AI episode, with @itsjessehi!. Covering:.- the Waifu vs Fan Platform market.- 2-5x'ing the income of OnlyFans creators.- Building and Finetuning with DSPy.- Model graded evals.- Guarding against prompt injections.- Memory for.
8
2
33
🆕 LLM Gateway: The One Decision That Removes 100 AI Engineering Decisions. quick pod/post ft @PortkeyAI @jumbld !
"Why does Production AI Engineering start with Evals?". the AI Engineer SDLC, silly. next in our @lizziepika lightboard series of short explainers (again, need to reshoot with better markers and idea of where the camera ends, but just trying out this new format). h/t to @doppenhe
3
13
52
🆕 pod: Efficiency is Coming: 3000x Faster, Cheaper, Better AI Inference. with @nyla_worker of @nvidia, @convaitech, @googleai!. The commoditization of intelligence takes on a few dimensions:. Time to Open Model Equivalent: 15 months between GPT-4 and Llama 3.1 405B (h/t
0
6
32
## Prompting, ICL & Chain of Thought. We initially started writing about the AI Engineer as a reaction against "Prompt Engineer" being a fulltime role in 2023; now the pendulum has swung the other way and not enough employers appreciate the importance of good prompting in AI
1
9
32
🆕 Agent Engineering with Pydantic + Graphs . with @samuel_colvin!. We've heard from @jxnlco 3 years running that @Pydantic Is All We Need, and this is even more true today with the launch of Pydantic AI! With a little @marimo_io love at the end :).
2
8
32
🆕 Llama 2: The New Open LLM SOTA. >2400 AI Engineers have now tuned into yesterday's recap of Llama 2, ft:.- @natolambert of HuggingFace.- @BornsteinMatt of a16z.- @atroyn of Chroma.- @russelljkaplan of Scale. and our returning faves @simonw + @altryne!.
3
13
29
🆕 short pod - how @shawnup got SOTA SWE-Bench Verified. with @openai o1. by building his own tools (on @weights_biases Weave) to look at data!
Our SWE-Bench submission has been accepted and is officially SOTA! Thanks SWE-Bench team for making such an important benchmark.
3
8
28
🆕 Bolt, Flow Engineering for Code Agents, and >$8m ARR in 2 months as a Claude Wrapper. with @ericsimons40 and @itamar_mar!. We are excited to catch up with @QodoAI and debut @stackblitz on the pod, who have both captured the hearts and minds of everyone from @levelsio to
3
8
29
OpenAI's RLHF evaluation metrics show us that RLHF can improve some metrics, but it doesn't always mean people are going to be happy with the changes - with @natolambert
2
11
23
🆕 The AI-First Graphics Editor. We talked with @suhail on:. - his journey from Mixpanel to Mighty.- How a conversation with @model_mechanic inspired @playground_ai.- raising $40m to reimagine the graphics editor.- how to self-teach AI with.
4
6
25
🗽 News: @aiDotEngineer is headed to NYC!!. Announcing the 2025 AI Engineer Summit. Applications open now!.
0
4
28
i should call her.
1
2
27
## The Best of 2024 in Post-Transformers Arch. We're very honored to feature a special joint talk between rival teams working on subquadratic attention:. - @realDanFu of @togethercompute (SSMs).- @picocreator of @recursal_AI (RWKV) @RWKV_AI . If you're looking to catch up on the
1
5
27
🆕 Open Operator, Serverless 🅱️rowsers and the Future of Computer-Using Agents. ft. @pk_iv of @browserbasehq !. Time for the OTHER big agent theme of the year so far, Operator and the category of Computer Use Agents (we last discussed with on our.
2
5
26
🆕 pod: Bee AI: The Wearable Agent. We sat down with @mariadlzollo and @EthanSutin, cofounders of @bee__computer, to learn about how they survived the brutal 2024 winter of AI Wearables (RIP Humane, Rabbit) and built the ultimate personal AI wearable.
5
6
25
Section 1: Frontier LLMs. the @openai canon, @AnthropicAI Claude and @GoogleDeepMind Gemini, @AIatMeta Llama 1/2/3, @MistralAI, @deepseek_ai 1/2/3, @apple intelligence. special mention: @soldni/@natolambert AI2, @xai grok, @awsai Nova, @LoubnaBenAllal1 SmolLM, et al.
1
1
27
Presenting: . The Most Important Vision Papers of 2024. as picked by @roboflow and @vikhyatk . The biggest trends are:. 1. Vision and Video are converging .-> Sora and OpenSora were built with Diffusion Transformers and MagViT.-> @nikhilaravi extended SAM from images to video. 2.
0
4
26
some AWESOME presentations at the @latentspacepod Survey Paper Club today:. - @honicky's guide to ColBERT (@lateinteraction): - @dimfeld's guide to The Landscape of Emerging AI Agent Architectures (@alexchaomander): -
0
5
26
🆕 NeurIPS 2023 Recap — Top Startups!. part 2 of our NeurIPS coverage is the biggest ensemble of founders and company leaders we’ve ever featured:. • @jefrankle - Chief Scientist, MosaicML.• @lqiao - CEO, Fireworks AI.• @amanrsanger - CEO, Anysphere.
1
11
24
🆕 Why you should write your own LLM benchmarks . w/ Nicholas Carlini of @GoogleDeepMind. Covering his greatest hits:.- How I Use AI.- My benchmark for large language models.- Extracting Training Data from Large Language Models (RIP @openai logprobs). Full episode below!
2
11
25
Thanks to @eugeneyan for starting the LS Discord #paperclub and @youngphlo_ our fancy new logo!. If you want a welcoming, unrecorded space to dive in to LLM papers every week, we're still open to new members on the LS Discord/Paper club. join in!
2
4
23
Must Know Benchmarks and Evals:. Knowledge: @DanHendrycks' MMLU and MATH, @idavidrein's GPQA and BIG-Bench and their polyunsaturated 2025 variants. Ditto Math lvl 5, AIME, @tamaybes's FrontierMath, etc. Long Context: @ZayneSprague's MuSR, @realYushiBai's LongBench,
3
1
24
🚨 EMERGENCY PAPER CLUB. The @latentspacepod discord is meeting in 2hrs to talk thru @lvdmaaten et al's The Llama 3 Herd of Models, early contender to win the POTY* Awards!. Join us (link below) with @swyx, @vibhuuuus, @picocreator, @eugeneyan, et al!. *Paper of The Year, totally
2
5
23
🆕 pod: RLHF 201. Our deep dive into Reinforcement Learning with Human Feedback, with @natolambert of @allen_ai + @interconnectsai!. Covering:. - History of RL and its applications.- Intellectual history of RLHF dating to Aristotle (!).- RLHF for
1
6
23
🆕 Building AGI in Real Time. Our @OpenAI DevDay Recap is now live!. A comprehensive audio experience of DevDay, with the people who made it happen:. - @oliviergodement.- @romainhuet returns!.- @michpokrass returns!.- @AlistairPullen returns!.- @simonw.
0
3
23
defining inner loop (in a git commit) vs outer loop (after the commit) AI agents. Do you think we'll have true outer loop agents? if so, when?
5
6
20
🆕 The Agent Reasoning Interface: o1/o3, Claude 3, ChatGPT Canvas, Tasks, and Operator. ft @karinanguyen_ who is teasing her keynote at @aiDotEngineer . Timestamps. •00:11 Introducing Karina Nguyen.•02:21 Karina's Journey to OpenAI.•04:45 Early Prototypes and Projects.•05:25
1
6
21
ok apparently there is TOO MUCH ALPHA in @sama's GPT 4.5 podcast today. we are hosting a live watch party in 5 mins. join.
wild this has <1k likes rn there’s a lot of stuff in here i did not know we were allowed to talk about lmao.
0
0
22
THIS SUNDAY IN SF. COME PLAY POKEMON. WITH AI AGENTS. AND CLAUDE. WITH SPECIAL GUEST @DavidSHershey .AND VMS FROM @JESSEMHAN. 100 SEATS ONLY.
3
5
23
actually Relu is the alexnet sysadmin's name
Today I learned: The term "Rectified" in ReLU comes from electronics, which means converting alternating current (AC) to direct current (DC) by blocking out the negative voltage (effectively making it zero).
2
5
22
⚡️ Raycast's new AI Extensions — with @thomaspaulmann. Anything a @raycastapp extension can do, can now be done in natural language!. We also dive into the behind the scenes of the new Ray-1 and Ray-1-mini models, and the at-scale usage of tool use.
0
1
22
🆕: Everything you need to run Mission Critical Inference (ft. DeepSeek v3 + SGLang). We chat with @amiruci and @yinengzhang about the Chinese Whale Bro drop of 2024:. - @deepseek_ai v3.- @lmsysorg's SGLang.- the Three Pillars of Mission Critical.
3
6
21
we are pr- . uh. we are happy to announce our brand refresh, reflecting our role in dimensionality reduction and translation of complex AI/ML topics to an AI engineering audience
ML twitter drinking game take a shot every time you see this stupid fucking shape
2
2
21
FYI: we are holding a snap @latentspacepod on Llama 2 at 4pm PT today!. ft. @natolambert @simonw @BornsteinMatt and ofc @fanahova and @swyx!.
1
5
21
🆕 State of the Art: Training >70B LLMs . We are excited to deep dive into @imbue_ai's incredible new releases with @joshalbrecht (CTO of Imbue). AND with the best return GUEST COHOST possible for the job: @jefrankle (Chief AI Scientist of @DbrxMosaicAI
Early this year, we trained a 70B model optimized for reasoning and coding. This model roughly matches LLAMA 3 70B despite being trained on 7x less data. Today, we’re releasing a toolkit to help others do the same, including:. • 11 sanitized and extended NLP reasoning
1
5
19
One Year of Latent Space 🔭. Friday is @latentspacepod birthday, so I wrote out some thoughts on our first year of hosting it with @swyx. Grateful for all the support, and a special s/o to @OfficialLoganK for helping us kick it off as the first guest 🚀
1
4
20
🆕Please stop forcing Clippy on those who want Anton. ChatGPT-4o's glazing lays open Clippy vs Anton: The two extremes of desires in AI post-training and product
2
6
23
The good news: Open models -exploded- in 2024!. We were honored to have @soldni and @sophiamyang break down the incredible year in open models in our latest published pod! . The bad news? Incentives, regulation, resource constraints, and many more challenges in 2025. If the
3
3
20
🆕 GPT 4.1 with @michpokrass and @j_mcgraph!. Part 1.00:00:00 – Introduction and Guest Welcome.00:00:57 – GPC 4.1 Launch Overview.00:01:54 – Developer Feedback and Model Names.00:02:53 – Model Naming and Starry Themes.00:03:49 – Confusion Over GPC 4.1 vs.
2
2
19
🆕 pod: Agents @ Work with @Altimor!. Discover why folks like @awilkinson are -obsessed- with Lindy, and why 2024 was the year AI agents went from "very exciting, but not quite working yet" to "making a real business impact / automating entire functions"!.
1
4
17