In the last 2 years we\u2019ve seen eye watering robotics funding - e.g., Figure with a ~$39B valuation and $1B+ rounds, Skild AI raising ~$1.4B, Physical Intelligence raising hundreds of millions, and autonomous systems like Wayve\u2019s ~$1.5B robotaxi funding, not to forget Musk with his Optimus bots.
That\u2019s an insane capital wave, but from a core bottleneck POV, what\u2019s actually changed since 2016-2020?
We\u2019ve heard about vision models, RL advances, diffusion policies, better sim, and multimodal embodied models but have any of these really cracked generalization, reliable manipulation, or true sim2real at scale?
Some questions:
1. Are we meaningfully closer to generalist policies that work in messy, real environments?
2. Do \u201crobot foundation models\u201d solve the data bottleneck the way LLMs did for NLP?
3. Has manipulation gone beyond incremental improvements?
4. Are humanoids a technical leap or just a narrative that attracts capital?
5. What are the real research papers/benchmarks showing step-change progress?
Genuinely curious whether we are at a technological inflection point or are we going to hit hard physics/data/hardware problems again."},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Billions of dollars in funding, but what's changed for robotics?"}},"_tags":["story","author_ajax33","story_47216355","ask_hn"],"author":"ajax33","children":[47218518,47222360,47378471],"created_at":"2026-03-02T10:58:10Z","created_at_i":1772449090,"num_comments":3,"objectID":"47216355","points":9,"story_id":47216355,"story_text":"HN folks - school me if this is an uninformed take:
In the last 2 years we\u2019ve seen eye watering robotics funding - e.g., Figure with a ~$39B valuation and $1B+ rounds, Skild AI raising ~$1.4B, Physical Intelligence raising hundreds of millions, and autonomous systems like Wayve\u2019s ~$1.5B robotaxi funding, not to forget Musk with his Optimus bots.
That\u2019s an insane capital wave, but from a core bottleneck POV, what\u2019s actually changed since 2016-2020?
We\u2019ve heard about vision models, RL advances, diffusion policies, better sim, and multimodal embodied models but have any of these really cracked generalization, reliable manipulation, or true sim2real at scale?
Some questions:
1. Are we meaningfully closer to generalist policies that work in messy, real environments?
2. Do \u201crobot foundation models\u201d solve the data bottleneck the way LLMs did for NLP?
3. Has manipulation gone beyond incremental improvements?
4. Are humanoids a technical leap or just a narrative that attracts capital?
5. What are the real research papers/benchmarks showing step-change progress?
Genuinely curious whether we are at a technological inflection point or are we going to hit hard physics/data/hardware problems again.","title":"Ask HN: Billions of dollars in funding, but what's changed for robotics?","updated_at":"2026-03-28T05:36:41Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"whiplash451"},"story_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"There seems to be a consensus that the self-driving game is now over \u2014 in a way or another. But EU has a single startup in the field (Wayve) and nobody\u2019s doing long-haul self-trucking in EU. Would it be such a terrible idea to start a company now?
There are reasons to believe that it will be solved over the next two decades given AI progress. The business case seems incredibly obvious. And if one job can be socially acceptable to automate, that one likely is."},"title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: Is a new EU self-trucking startup a bad idea in 2025?"}},"_tags":["story","author_whiplash451","story_46013614","ask_hn"],"author":"whiplash451","created_at":"2025-11-22T10:14:58Z","created_at_i":1763806498,"num_comments":0,"objectID":"46013614","points":2,"story_id":46013614,"story_text":"There seems to be a consensus that the self-driving game is now over \u2014 in a way or another. But EU has a single startup in the field (Wayve) and nobody\u2019s doing long-haul self-trucking in EU. Would it be such a terrible idea to start a company now?
There are reasons to believe that it will be solved over the next two decades given AI progress. The business case seems incredibly obvious. And if one job can be socially acceptable to automate, that one likely is.","title":"Ask HN: Is a new EU self-trucking startup a bad idea in 2025?","updated_at":"2026-03-05T23:05:04Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"programd"},"title":{"matchLevel":"none","matchedWords":[],"value":"Gaia-1 a 9B parameter generative world model for autonomous driving"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/thinking/scaling-gaia-1/"}},"_tags":["story","author_programd","story_37782670"],"author":"programd","children":[37785105,37785158,37785195,37785793,37785948,37786118,37786368,37786794,37787208,37787379,37788056],"created_at":"2023-10-05T18:59:42Z","created_at_i":1696532382,"num_comments":62,"objectID":"37782670","points":118,"story_id":37782670,"title":"Gaia-1 a 9B parameter generative world model for autonomous driving","updated_at":"2025-11-10T20:42:11Z","url":"https://wayve.ai/thinking/scaling-gaia-1/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"Chirono"},"title":{"matchLevel":"none","matchedWords":[],"value":"Photorealistic reconstruction in static and dynamic scenes"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/thinking/prism-1/"}},"_tags":["story","author_Chirono","story_40707592"],"author":"Chirono","children":[40730181],"created_at":"2024-06-17T16:32:19Z","created_at_i":1718641939,"num_comments":1,"objectID":"40707592","points":9,"story_id":40707592,"title":"Photorealistic reconstruction in static and dynamic scenes","updated_at":"2024-09-20T17:11:57Z","url":"https://wayve.ai/thinking/prism-1/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"roryirvine"},"title":{"matchLevel":"none","matchedWords":[],"value":"Self-driving taxis cleared for London streets 'later this summer'"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://www.theguardian.com/technology/2026/aug/05/self-driving-taxis-approved-london-summer-2026-wayve-uber"}},"_tags":["story","author_roryirvine","story_49193915"],"author":"roryirvine","children":[49193980,49194029],"created_at":"2026-08-06T08:11:36Z","created_at_i":1786003896,"num_comments":3,"objectID":"49193915","points":5,"story_id":49193915,"title":"Self-driving taxis cleared for London streets 'later this summer'","updated_at":"2026-08-06T10:55:37Z","url":"https://www.theguardian.com/technology/2026/aug/05/self-driving-taxis-approved-london-summer-2026-wayve-uber"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"EvgeniyZh"},"title":{"matchLevel":"none","matchedWords":[],"value":"Lingo-1: Exploring Natural Language for Autonomous Driving"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/thinking/lingo-natural-language-autonomous-driving/"}},"_tags":["story","author_EvgeniyZh","story_37507411"],"author":"EvgeniyZh","created_at":"2023-09-14T11:09:02Z","created_at_i":1694689742,"num_comments":0,"objectID":"37507411","points":3,"story_id":37507411,"title":"Lingo-1: Exploring Natural Language for Autonomous Driving","updated_at":"2024-09-20T15:07:15Z","url":"https://wayve.ai/thinking/lingo-natural-language-autonomous-driving/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rbanffy"},"title":{"matchLevel":"none","matchedWords":[],"value":"UK startup thinks it can win the self-driving car race w better machine learning"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://techcrunch.com/2018/05/22/wayve/"}},"_tags":["story","author_rbanffy","story_17125152"],"author":"rbanffy","created_at":"2018-05-22T11:31:57Z","created_at_i":1526988717,"num_comments":0,"objectID":"17125152","points":3,"story_id":17125152,"title":"UK startup thinks it can win the self-driving car race w better machine learning","updated_at":"2024-09-20T02:29:12Z","url":"https://techcrunch.com/2018/05/22/wayve/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"bookofjoe"},"title":{"matchLevel":"none","matchedWords":[],"value":"London's first self-driving taxis for hire hit the streets"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://www.theguardian.com/technology/2026/sep/03/london-first-self-driving-taxis-for-hire-wayve-uber"}},"_tags":["story","author_bookofjoe","story_49552161"],"author":"bookofjoe","created_at":"2026-09-03T15:58:03Z","created_at_i":1788451083,"num_comments":0,"objectID":"49552161","points":2,"story_id":49552161,"title":"London's first self-driving taxis for hire hit the streets","updated_at":"2026-09-08T16:53:48Z","url":"https://www.theguardian.com/technology/2026/sep/03/london-first-self-driving-taxis-for-hire-wayve-uber"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"TMWNN"},"title":{"matchLevel":"none","matchedWords":[],"value":"I rode in one of the UK's first self-driving cars"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://www.theverge.com/report/820324/wayve-driverless-robotaxi-london"}},"_tags":["story","author_TMWNN","story_45936689"],"author":"TMWNN","children":[45937396],"created_at":"2025-11-15T11:24:25Z","created_at_i":1763205865,"num_comments":0,"objectID":"45936689","points":2,"story_id":45936689,"title":"I rode in one of the UK's first self-driving cars","updated_at":"2026-03-05T22:59:57Z","url":"https://www.theverge.com/report/820324/wayve-driverless-robotaxi-london"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"BerislavLopac"},"title":{"matchLevel":"none","matchedWords":[],"value":"Driverless taxis are coming to London's streets in the spring"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://www.londoncentric.media/p/driverless-taxis-waymo-wayve-are-coming-to-london"}},"_tags":["story","author_BerislavLopac","story_45602849"],"author":"BerislavLopac","created_at":"2025-10-16T08:32:06Z","created_at_i":1760603526,"num_comments":0,"objectID":"45602849","points":2,"story_id":45602849,"title":"Driverless taxis are coming to London's streets in the spring","updated_at":"2026-03-05T22:48:02Z","url":"https://www.londoncentric.media/p/driverless-taxis-waymo-wayve-are-coming-to-london"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"sturza"},"title":{"matchLevel":"none","matchedWords":[],"value":"Fusing Object-Level Vector Modality for Explainable Autonomous Driving"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://github.com/wayveai/Driving-with-LLMs"}},"_tags":["story","author_sturza","story_43375009"],"author":"sturza","created_at":"2025-03-15T20:40:44Z","created_at_i":1742071244,"num_comments":0,"objectID":"43375009","points":2,"story_id":43375009,"title":"Fusing Object-Level Vector Modality for Explainable Autonomous Driving","updated_at":"2025-03-15T20:52:54Z","url":"https://github.com/wayveai/Driving-with-LLMs"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"Chirono"},"title":{"matchLevel":"none","matchedWords":[],"value":"Lingo-2: Driving with Natural Language"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/thinking/lingo-2-driving-with-language/"}},"_tags":["story","author_Chirono","story_40062679"],"author":"Chirono","created_at":"2024-04-17T10:27:41Z","created_at_i":1713349661,"num_comments":0,"objectID":"40062679","points":2,"story_id":40062679,"title":"Lingo-2: Driving with Natural Language","updated_at":"2024-09-20T16:50:47Z","url":"https://wayve.ai/thinking/lingo-2-driving-with-language/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"marban"},"title":{"matchLevel":"none","matchedWords":[],"value":"The big new idea for making self-driving cars that can go anywhere"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://www.technologyreview.com/2022/05/27/1052826/ai-reinforcement-learning-self-driving-cars-autonomous-vehicles-wayve-waabi-cruise/"}},"_tags":["story","author_marban","story_31557398"],"author":"marban","created_at":"2022-05-30T10:21:21Z","created_at_i":1653906081,"num_comments":0,"objectID":"31557398","points":2,"story_id":31557398,"title":"The big new idea for making self-driving cars that can go anywhere","updated_at":"2024-09-20T11:16:16Z","url":"https://www.technologyreview.com/2022/05/27/1052826/ai-reinforcement-learning-self-driving-cars-autonomous-vehicles-wayve-waabi-cruise/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"rdednl"},"title":{"matchLevel":"none","matchedWords":[],"value":"Simulation Training, Real Driving"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/blog/sim2real"}},"_tags":["story","author_rdednl","story_18664270"],"author":"rdednl","created_at":"2018-12-12T14:42:40Z","created_at_i":1544625760,"num_comments":0,"objectID":"18664270","points":2,"story_id":18664270,"title":"Simulation Training, Real Driving","updated_at":"2024-09-20T03:32:14Z","url":"https://wayve.ai/blog/sim2real"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"gyre007"},"title":{"matchLevel":"none","matchedWords":[],"value":"Dreaming about driving using Reinforcement Learning"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/blog/dreaming-about-driving-imagination-rl"}},"_tags":["story","author_gyre007","story_17917556"],"author":"gyre007","created_at":"2018-09-05T13:12:31Z","created_at_i":1536153151,"num_comments":0,"objectID":"17917556","points":2,"story_id":17917556,"title":"Dreaming about driving using Reinforcement Learning","updated_at":"2024-09-20T02:56:52Z","url":"https://wayve.ai/blog/dreaming-about-driving-imagination-rl"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"albertzeyer"},"title":{"matchLevel":"none","matchedWords":[],"value":"Learning to drive in a day"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/blog/learning-to-drive-in-a-day-with-reinforcement-learning"}},"_tags":["story","author_albertzeyer","story_17785506"],"author":"albertzeyer","created_at":"2018-08-17T20:48:08Z","created_at_i":1534538888,"num_comments":0,"objectID":"17785506","points":2,"story_id":17785506,"title":"Learning to drive in a day","updated_at":"2024-09-20T02:57:25Z","url":"https://wayve.ai/blog/learning-to-drive-in-a-day-with-reinforcement-learning"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"spolu"},"title":{"matchLevel":"none","matchedWords":[],"value":"The first example of reinforcement learning on-board an autonomous car"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/blog/learning-to-drive-in-a-day-with-reinforcement-learning"}},"_tags":["story","author_spolu","story_17448191"],"author":"spolu","created_at":"2018-07-03T07:25:03Z","created_at_i":1530602703,"num_comments":0,"objectID":"17448191","points":2,"story_id":17448191,"title":"The first example of reinforcement learning on-board an autonomous car","updated_at":"2024-09-20T02:42:41Z","url":"https://wayve.ai/blog/learning-to-drive-in-a-day-with-reinforcement-learning"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"paulbaumgart"},"title":{"matchLevel":"none","matchedWords":[],"value":"Scaling Autonomous Vehicles with Machine Learning"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://medium.com/@mhdempsey/scaling-autonomous-vehicles-with-machine-learning-our-investment-in-wayve-4d9bcf127932"}},"_tags":["story","author_paulbaumgart","story_17131532"],"author":"paulbaumgart","created_at":"2018-05-23T04:43:15Z","created_at_i":1527050595,"num_comments":0,"objectID":"17131532","points":2,"story_id":17131532,"title":"Scaling Autonomous Vehicles with Machine Learning","updated_at":"2024-09-20T02:29:49Z","url":"https://medium.com/@mhdempsey/scaling-autonomous-vehicles-with-machine-learning-our-investment-in-wayve-4d9bcf127932"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jcfrei"},"title":{"matchLevel":"none","matchedWords":[],"value":"Solving the long-tail of driving automation with end-to-end AI"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/thinking/e2e-embodied-ai-solves-the-long-tail/"}},"_tags":["story","author_jcfrei","story_41040553"],"author":"jcfrei","created_at":"2024-07-22T22:32:30Z","created_at_i":1721687550,"num_comments":0,"objectID":"41040553","points":1,"story_id":41040553,"title":"Solving the long-tail of driving automation with end-to-end AI","updated_at":"2024-09-20T17:27:35Z","url":"https://wayve.ai/thinking/e2e-embodied-ai-solves-the-long-tail/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"EhsanEtezad"},"title":{"matchLevel":"none","matchedWords":[],"value":"Microsoft-backed Self-driving Startup expands to Vancouver"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://www.bnnbloomberg.ca/microsoft-backed-self-driving-startup-wayve-expands-to-vancouver-1.2057888"}},"_tags":["story","author_EhsanEtezad","story_40042092"],"author":"EhsanEtezad","created_at":"2024-04-15T15:45:24Z","created_at_i":1713195924,"num_comments":0,"objectID":"40042092","points":1,"story_id":40042092,"title":"Microsoft-backed Self-driving Startup expands to Vancouver","updated_at":"2024-09-20T16:48:48Z","url":"https://www.bnnbloomberg.ca/microsoft-backed-self-driving-startup-wayve-expands-to-vancouver-1.2057888"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jedharris"},"title":{"matchLevel":"none","matchedWords":[],"value":"Solving the long tail with end to end AI"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/thinking/e2e-embodied-ai-solves-the-long-tail/"}},"_tags":["story","author_jedharris","story_39768961"],"author":"jedharris","created_at":"2024-03-20T16:37:48Z","created_at_i":1710952668,"num_comments":0,"objectID":"39768961","points":1,"story_id":39768961,"title":"Solving the long tail with end to end AI","updated_at":"2024-09-20T16:40:21Z","url":"https://wayve.ai/thinking/e2e-embodied-ai-solves-the-long-tail/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"danboarder"},"title":{"matchLevel":"none","matchedWords":[],"value":"Predicting the Future (2020)"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/blog/predicting-the-future/"}},"_tags":["story","author_danboarder","story_27786806"],"author":"danboarder","created_at":"2021-07-09T18:24:14Z","created_at_i":1625855054,"num_comments":0,"objectID":"27786806","points":1,"story_id":27786806,"title":"Predicting the Future (2020)","updated_at":"2024-09-20T08:58:30Z","url":"https://wayve.ai/blog/predicting-the-future/"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"dsr12"},"title":{"matchLevel":"none","matchedWords":[],"value":"Predicting the Future from Monocular Cameras in Bird\u2019s-Eye View"},"url":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"https://wayve.ai/blog/fiery-future-instance-prediction-birds-eye-view/"}},"_tags":["story","author_dsr12","story_26922129"],"author":"dsr12","created_at":"2021-04-24T03:55:27Z","created_at_i":1619236527,"num_comments":0,"objectID":"26922129","points":1,"story_id":26922129,"title":"Predicting the Future from Monocular Cameras in Bird\u2019s-Eye View","updated_at":"2024-09-20T08:23:51Z","url":"https://wayve.ai/blog/fiery-future-instance-prediction-birds-eye-view/"},{"_highlightResult":{"author":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"wayvey"},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: A Heart Rate Monitor App Written in Vanilla JavaScript"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/richrd/heart-rate-monitor"}},"_tags":["story","author_wayvey","story_25768112","show_hn"],"author":"wayvey","children":[25769464,25769839,25771007,25785907,25797079],"created_at":"2021-01-13T21:19:37Z","created_at_i":1610572777,"num_comments":6,"objectID":"25768112","points":45,"story_id":25768112,"title":"Show HN: A Heart Rate Monitor App Written in Vanilla JavaScript","updated_at":"2024-11-01T22:47:12Z","url":"https://github.com/richrd/heart-rate-monitor"},{"_highlightResult":{"author":{"fullyHighlighted":true,"matchLevel":"full","matchedWords":["wayve"],"value":"Wayve"},"story_text":{"matchLevel":"none","matchedWords":[],"value":"Despite the Liquid Glass controversy, our team is quite excited about the new on-device ASR model and Foundation Model Framework. It makes building local-first AI apps much easier, it was very easy to build a basic scribe app compared to our first iteration of our meeting note taker. Everything runs locally with zero cloud dependencies. With the option to combine with cloud models for more complex tasks, we hope to see more local-first AI apps.
Microsoft just announced Foundry local as well, and it'll be interesting to see where they go with it.
There are still some limitations with the foundation model, such as a 4,096 context window, but it will probably get better over time: \nhttps://gist.github.com/samhenrigold/3aad01b762ccc87e34e6115..."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Swift Scribe: On-device AI scribe using Apple's new Foundation Models"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/slipboxai/swift-scribe"}},"_tags":["story","author_Wayve","story_44269925","show_hn"],"author":"Wayve","children":[44269965,44331723],"created_at":"2025-06-13T16:25:25Z","created_at_i":1749831925,"num_comments":1,"objectID":"44269925","points":13,"story_id":44269925,"story_text":"Despite the Liquid Glass controversy, our team is quite excited about the new on-device ASR model and Foundation Model Framework. It makes building local-first AI apps much easier, it was very easy to build a basic scribe app compared to our first iteration of our meeting note taker. Everything runs locally with zero cloud dependencies. With the option to combine with cloud models for more complex tasks, we hope to see more local-first AI apps.
Microsoft just announced Foundry local as well, and it'll be interesting to see where they go with it.
There are still some limitations with the foundation model, such as a 4,096 context window, but it will probably get better over time: \nhttps://gist.github.com/samhenrigold/3aad01b762ccc87e34e6115...","title":"Show HN: Swift Scribe: On-device AI scribe using Apple's new Foundation Models","updated_at":"2025-06-20T20:31:11Z","url":"https://github.com/slipboxai/swift-scribe"},{"_highlightResult":{"author":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"wayvey"},"title":{"matchLevel":"none","matchedWords":[],"value":"Video inside centrifuge shows we don't understand fluid physics"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://www.newscientist.com/article/2407923-video-inside-centrifuge-shows-we-dont-fully-understand-fluid-physics/"}},"_tags":["story","author_wayvey","story_38671817"],"author":"wayvey","children":[38671928,38671932],"created_at":"2023-12-17T10:19:17Z","created_at_i":1702808357,"num_comments":5,"objectID":"38671817","points":11,"story_id":38671817,"title":"Video inside centrifuge shows we don't understand fluid physics","updated_at":"2024-09-20T15:55:12Z","url":"https://www.newscientist.com/article/2407923-video-inside-centrifuge-shows-we-dont-fully-understand-fluid-physics/"},{"_highlightResult":{"author":{"fullyHighlighted":true,"matchLevel":"full","matchedWords":["wayve"],"value":"Wayve"},"story_text":{"matchLevel":"none","matchedWords":[],"value":"We needed a speaker diarization solution that could run every few seconds alongside transcription on iOS and macOS. But native Swift support was either limited or locked behind paid licenses. Since diarization is a common need in speech-to-text workflows, we decided to open source our work and give back to the community.
We initially tried sherpa-onnx, which works, but running both diarization and transcription models slowed down older devices. CPU-only inference just isn\u2019t ideal for near real-time workloads, so we wanted the option to offload segmentation and speaker embedding to the GPU or ANE. Supporting M1 Macs in particular meant pushing more of the workload to the ANE.
Instead of shoehorning the ONNX model into CoreML with C++, we converted the original PyTorch models directly to CoreML. This approach required some monkey-patching in the PyTorch and pyannote code, but the initial benchmarks look promising.
We\u2019d love feedback! We're currently working on adding VAD and integrating Parakeet for transcription, but still wrestling with CoreML model conversion."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: FluidAudio \u2013 Swift Speaker Diarization on CoreML"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/FluidInference/FluidAudio"}},"_tags":["story","author_Wayve","story_44546863","show_hn"],"author":"Wayve","created_at":"2025-07-13T01:58:39Z","created_at_i":1752371919,"num_comments":0,"objectID":"44546863","points":6,"story_id":44546863,"story_text":"We needed a speaker diarization solution that could run every few seconds alongside transcription on iOS and macOS. But native Swift support was either limited or locked behind paid licenses. Since diarization is a common need in speech-to-text workflows, we decided to open source our work and give back to the community.
We initially tried sherpa-onnx, which works, but running both diarization and transcription models slowed down older devices. CPU-only inference just isn\u2019t ideal for near real-time workloads, so we wanted the option to offload segmentation and speaker embedding to the GPU or ANE. Supporting M1 Macs in particular meant pushing more of the workload to the ANE.
Instead of shoehorning the ONNX model into CoreML with C++, we converted the original PyTorch models directly to CoreML. This approach required some monkey-patching in the PyTorch and pyannote code, but the initial benchmarks look promising.
We\u2019d love feedback! We're currently working on adding VAD and integrating Parakeet for transcription, but still wrestling with CoreML model conversion.","title":"Show HN: FluidAudio \u2013 Swift Speaker Diarization on CoreML","updated_at":"2025-07-13T19:22:05Z","url":"https://github.com/FluidInference/FluidAudio"},{"_highlightResult":{"author":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["wayve"],"value":"wayvey"},"title":{"matchLevel":"none","matchedWords":[],"value":"Disco Elysium studio laying off nearly 25% of its staff"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://www.pcgamer.com/disco-elysium-studio-laying-off-nearly-25-of-its-staffincluding-the-final-remaining-writer-from-the-original-games-credits/"}},"_tags":["story","author_wayvey","story_39413624"],"author":"wayvey","created_at":"2024-02-17T21:15:17Z","created_at_i":1708204517,"num_comments":0,"objectID":"39413624","points":5,"story_id":39413624,"title":"Disco Elysium studio laying off nearly 25% of its staff","updated_at":"2024-09-20T16:22:58Z","url":"https://www.pcgamer.com/disco-elysium-studio-laying-off-nearly-25-of-its-staffincluding-the-final-remaining-writer-from-the-original-games-credits/"},{"_highlightResult":{"author":{"fullyHighlighted":true,"matchLevel":"full","matchedWords":["wayve"],"value":"Wayve"},"title":{"matchLevel":"none","matchedWords":[],"value":"Making group conversations more accessible with sound localization"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://research.google/blog/making-group-conversations-more-accessible-with-sound-localization/"}},"_tags":["story","author_Wayve","story_44477626"],"author":"Wayve","created_at":"2025-07-06T03:37:39Z","created_at_i":1751773059,"num_comments":0,"objectID":"44477626","points":3,"story_id":44477626,"title":"Making group conversations more accessible with sound localization","updated_at":"2025-07-06T14:17:53Z","url":"https://research.google/blog/making-group-conversations-more-accessible-with-sound-localization/"}],"hitsPerPage":50,"nbHits":11540,"nbPages":20,"page":0,"params":"query=Wayve&tags=story&hitsPerPage=50&advancedSyntax=true&analyticsTags=backend","processingTimeMS":13,"processingTimingsMS":{"_request":{"queue":16,"roundTrip":19},"afterFetch":{"merge":{"mergeLoop":{"prepareNextHit":3,"total":3},"total":3},"total":3},"fetch":{"query":4,"scanning":4,"total":9},"total":13},"query":"Wayve","serverTimeMS":30}