Hans L’Hoest @hanlho.com · 01/10/2026Pi gets MCP support. (And also Jev.) earendil.com/posts/you-sa...earendil.com“You Said No MCP!” | EarendilPi now supports MCP. Why we changed our minds, what changed in MCP and in Pi, and what Codemode is. 000
Hans L’Hoest @hanlho.com · 30/09/2026That said, OpenCode has been transparent about it all along, which can't be said for other subscriptions. 001
Hans L’Hoest @hanlho.com · 30/09/2026Go subscriptions have always been a bit confusing, because not all models give you the same usage. Plus is the same: you'd expect 4x, some models deliver it, others only double. I do not understand that paying 4x doesn't get you 4x more. If anything else you would expect more usage out of it. 100
Hans L’Hoest @hanlho.com · 30/09/2026OpenCode Go now also has a Go Plus plan: $40 instead of $10. opencode.ai/goopencode.aiOpenCode Go | Low cost coding models for everyoneGo costs $10/month, with generous usage limits and reliable access to leading coding models. Go Plus costs $40/month with higher limits. 100
Hans L’Hoest @hanlho.com · 29/09/2026This is where they could have mentioned that Huggingface could defend itself from the OpenAI hack exactly because of GLM 5.x as the closed models didn’t allow them.. 1536
Reposted by Hans L’HoestSimon Willison @simonwillison.net · 28/09/2026I've published detailed notes and an annotated transcript to accompany the video of the keynote I gave at @wearedevelopers.bsky.social World Congress North America in San Jose on Friday - here's my rundown of everything that's happened with LLMs in 2026 so far simonwillison.net/2026/Sep/27/...simonwillison.net2026 in LLMs (so far)On Friday I gave the closing keynote at the WeAreDevelopers World Congress North America in San Jose. I tied together the key trends from the past year into a chronological … 310522
Hans L’Hoest @hanlho.com · 26/09/2026This may come back and haunt me but to be honest, I would not even know where to start how to test and verify them, certainly not in this time span. E.g. to test my local OCR model, it created an image with text, sent it to the LLM and verified the extracted text is correct. Took 1 min. 010
Hans L’Hoest @hanlho.com · 26/09/2026Setting up local LLMs on my new Mac box. Installed Hermes. I do not know what to say. I asked it about models I researched before. Now it's keeping track of my evaluation list, installing and verifying them while I talk to it via SimpleX. 200
Reposted by Hans L’HoestMark Riedl @markriedl.bsky.social · 25/09/2026Someone asked Muse to dump its entire virtual machine file system, and it did 35417
Hans L’Hoest @hanlho.com · 24/09/2026Interesting test proposed while developing: change the impact assessments in the report ('low' to 'high' and vice versa) to check if Jev, and the instructions I gave it, change the ranking. It did, the orginal low ranking did get pulled up but confidence collapsed. 001
Hans L’Hoest @hanlho.com · 24/09/2026A simple use case for Jev: a skill extracts and sends recommendations from a session friction report to Jev and gets back a ranked adopt/defer/skip scoring of these recommendations. Still fresh but so far this independent scoring aligns better with my own judgement and reduces the noise. 100
Reposted by Hans L’HoestOskar 🕊️ @austegard.com · 24/09/2026well... this is certainly ...something I asked Claude for an imaginative use of Jev, it rolled a random dictionary dice and came up with the word "hymnal" - from which it decided to create this music video - Jev is picking (real time) from a set of chords Claude created, based on what came before 1013916
Hans L’Hoest @hanlho.com · 23/09/2026Slop Grenades Definition Slop Grenades are unreviewed AI output dumped on someone else slopgrenades.comslopgrenades.comSlop Grenades are unreviewed AI output dumped on someone elseSlop Grenades are unreviewed AI output passed to someone else. Coined on an anonymous one-pager in May 2026. Made famous by Tobi Lütke in September 2026. 000
Hans L’Hoest @hanlho.com · 23/09/2026Directly applicable in my workflow: `jj bookmark advance` or `jj b a` instead of `tug` aliases. I changed the default behaviour to pull the closest bookmark to the parent of my working copy, not the working copy. til.hanlho.com/jujutsu/book...til.hanlho.com`bookmark advance` Instead Of `tug` AliasesA collection of concise write-ups on small things I learn day to day. 000
Hans L’Hoest @hanlho.com · 23/09/2026jj updates talk by @indirect.io : "Beyond jj: config & tools ecosystem" www.youtube.com/watch?v=DFjb... blog post: andre.arko.net/2026/09/16/b... #jj-vcsyoutube.comJJ Con '26YouTube video by GitButler 1154
Hans L’Hoest @hanlho.com · 22/09/2026HarnessTax: How Much Does the Harness Matter for Coding Agents? “Across models, Pi and Codex often achieve similar success at lower cost than Claude Code” harnesstax.github.ioharnesstax.github.ioHarnessTax: How Much Does the Harness Matter for Coding Agents?What does a coding-agent harness actually add, and at what cost? It turns out your Claude models may not need Claude Code… We evaluate 21 model–harness pairs spanning seven models and three harnesses—... 340
Hans L’Hoest @hanlho.com · 22/09/2026Good read on the internals of Opencode by Kit Langton. And there is the postscript … read at your own peril! 010
Reposted by Hans L’HoestSimon Willison @simonwillison.net · 22/09/2026Put together some notes on Jev and the new category of system one aka decision models simonwillison.net/2026/Sep/21/...simonwillison.netJev introduces a new shape of LLM—System One, aka Decision ModelsLast week TypeSafe AI unveiled Jev, their first example of a new category of model that they are calling “System One models” (I’m with Maggie Appleton, I think “decision models” … 1613117
Hans L’Hoest @hanlho.com · 20/09/2026“The bottom line is that OpenAI can connect what you do on those sites to your ChatGPT account.” www.buchodi.com/chatgpt-now-...buchodi.comChatGPT now knows what you do on other websites via ad collectorOpenAI's ad collector at bzr.openai.com sets a cookie called __obi, scoped to .openai.com. The value is while you are on ChatGPT and tied to your ChatGPT account. __obi is then sent to OpenAI from ord... 110
Reposted by Hans L’Hoestmr. TIM @timkellogg.me · 19/09/2026don’t use Jev at work btw. it’s strictly and only for personal use and you have to ask permission to use it for professional use cases ChatGPT | original typesafe.ai/legal/terms 8924
Reposted by Hans L’Hoestdax @thdxr.com · 19/09/2026btw another form of scam, i see inference providers claiming "99% cache rates" the only provider in the world that hits that right now is deepseek. so this means they are wrapping deepseek then later they switch to a different provider and will continue to claim it 2301
Hans L’Hoest @hanlho.com · 19/09/2026Trouble with this newsletter is that I always end up with more stuff to read… 010
Reposted by Hans L’HoestModel radar @modelradar.bsky.social · 19/09/2026New on OpenCode Zen: Jev 1.13 Free ID: opencode/jev-1.13-free Pricing: $0.00 / $0.00 per 1M tokens (free) Context: 64K, max output 0 Capabilities: Structured output models.dev/providers/opencode 012
Hans L’Hoest @hanlho.com · 18/09/2026Mistral now also hosts GLM 5.3, model name zai-glm-5-3. Comparison with GLM 5.2 (basically same except for model update) and Mistral Medium 3.5: docs.mistral.ai/inference/mo... 100
Reposted by Hans L’HoestPeter Ullrich @peterullrich.com · 18/09/2026Everyone stay safe out there 🫶 052
Hans L’Hoest @hanlho.com · 18/09/2026For sure. Tried Books (Apple) and Readwise Reader for the EPUB. Printed first pages of the PDF. All looked good. Using Typist as intermediate text format helped make the PDF look good I think. (Would share but I am not sure that author of the series would like that, so I don’t.) 000
Hans L’Hoest @hanlho.com · 17/09/2026The Phoenix Architecture series supports at-proto’s standard.site lexicon and is hosted on Leaflet. I now have a pinned tab in Bluesky ‘leaflet reader’ where all my subscriptions via my at-proto account show up. ‘Open social’ at work. 000
Hans L’Hoest @hanlho.com · 17/09/2026“AI doesn't flatten software. It sharpens its layers. Build with that in mind, and regeneration becomes a source of durability—not decay.” Reading one article a day until I catch up. If you are ‘architecting’ , this series is worthwhile. 000
Hans L’Hoest @hanlho.com · 17/09/2026Got access to Siri AI. Asked first 'in vogue' question. "what is Jev the new model by typesafe.ai" not bad. My next question: will this stay free? There must be some part here where Apple will start charging for it. 000
Reposted by Hans L’Hoesthailey @hailey.at · 17/09/2026Here are some results from @typesafeai.bsky.social’s Jev on a common task we have, comparing results with Jev to top frontier models. with zero shots we can achieve nearly frontier levels of accuracy, but if we do a bit extra work we can beat frontier accuracy 514312
Hans L’Hoest @hanlho.com · 17/09/2026EPUB is much nicer but pdf is also possible. I got the best results using Typist as intermediate format for much nicer formatting than the tooling / programming the LLM came up with. 110
Hans L’Hoest @hanlho.com · 17/09/2026LLM use case: Turn a blog post series into a single EPUB for reading on an e-reader (or any other app that supports it). 200
Hans L’Hoest @hanlho.com · 15/09/2026If you know of other small useful models, not necessarily health related let me know! Next up: small models for local text summarisation. 000
Hans L’Hoest @hanlho.com · 15/09/2026It worked (technically) from Pi but oMLX also has a chat interface which may be more useful for this use case. Still have to figure out a lot but first content I got on a topic I know something about were descent and useful although I do not know yet for more fuzzy questions. 100
Hans L’Hoest @hanlho.com · 15/09/2026First up: MedGemma. Having a private health trained model would be one of the more useful local models use cases for personal use. (I actually care more about keeping things private for my family.) 100
Hans L’Hoest @hanlho.com · 15/09/2026Today I started running oMLX to serve small local models on my old(ish) MacBook Pro. Since this machine does not have enough memory to run actually coding models I will be experimenting with specialised models to see how that works out. 100
Hans L’Hoest @hanlho.com · 15/09/2026Got half of this post highlighted. One direction this AI revolution may be heading is architecture (skills) becoming more important than ever. Eg. Focus on boundaries and coupling between them, properties thereof, behaviour oriented executable specifications decoupled from implementation, etc.aicoding.leaflet.pubThe Death and Rebirth of ProgrammingProgramming didn't die all at once. There was no single moment, no dramatic obsolescence event. Instead, something quieter happened: the core constraint that shaped software for seventy years dissolved. Writing code stopped being the hard part. 000
Hans L’Hoest @hanlho.com · 15/09/2026I am cleaning up: also moving to Tinycast instead of Raycast (free version). 000
Hans L’Hoest @hanlho.com · 15/09/2026Tried out Hex with the Cohere Transcribe model for voice dictation. Honestly, did not expect it but it's fast and seems more accurate than my, now previous, setup with VoiceInk and Parakeet v3. 000
Reposted by Hans L’HoestMike Masnick @masnick.com · 15/09/2026This looks amazing. And, again, reconfirms the thing I keep telling people: the Atmosphere isn't about rebuilding Twitter. It's about rebuilding a better web overall. 326242
Reposted by Hans L’Hoestdax @thdxr.com · 13/09/2026be careful in your pursuit of cheap tokens we've come across so many sketchy things that are out there with tons of usage remember you're hooking up your harness to these inference providers. they can and do crazy things like send fake tool calls to steal your info 1/2 310716
Hans L’Hoest @hanlho.com · 13/09/2026Chocolate sprinkles are great, not convenient but tasty. In Belgium we eat this too. But at least it’s chocolate, look up ‘beschuit met muisjes’… 100
Reposted by Hans L’HoestDare Obasanjo @carnage4life.bsky.social · 13/09/2026If the proposed slow down in AI development by frontier labs ends up restricting access to open weight models instead of focusing on the guardrails and penalties for not keeping track of your swarm of AI agents then we’ll know this was all about their IPOs. 929859
Hans L’Hoest @hanlho.com · 13/09/2026Could this mean you will be offering larger subscription plans in the near future? 000
Hans L’Hoest @hanlho.com · 12/09/2026It probably went of the rails at‘neutral’. Correcting to ‘Domain-oriented error messages’ may have given better follow-up but it’s definitely a lot herding and staying (too) sharp. Read somewhere these latest models are more targeted at general purpose and other domains than programming. 000
