Latest posts 5
With AI here to do the heavy lifting, coders may want to consider the programming language, given its runtime performance and safety guarantees, suggested Carl Lerche, an Amazon Web Services principal engineer, speaking at ScyllaDB’s 2026 P99 virtual conference last week.
theregister.com/devops/2026/10…
theregister.com/devops/2026/10…
USearch is an amazing open source project that bundles the best vector similarity search library with open, free and responsive code & community. Well done!
USearch v3 is ready for a design-partner preview!View quoted post on X
3 years ago, I released USearch v2. It became my first widely deployed piece of software.
It grew to over 4’000 GitHub stars, a number unimaginable to me back then. Together with my complementary projects, that’s 17’000 stars today. The open-source infrastructure behemoths shipping these libraries by default, like ClickHouse & ScyllaDB, collectively account for over 550’000 stars. Almost 800’000 including opt-in integrations 🤯
Thank you to everyone who shipped, starred, or argued with this code! I know I can be slow to respond, and I appreciate your patience.
With help from AI agents, compute from neocloud partners and a lot of user feedback, USearch v3 finally has a release candidate. Only 1’000 days of work later 🫠
USearch v3 stays open-source, Apache 2.0, free for everyone, from open-source giants to neolabs, hyperscalers, government agencies, & even mobile publishers!
Still, I’ve seen many times how paying for an integrated software experience can work better and cost less than getting the software for free — and be in everyone’s best interest. So I want to take a page from ClickHouse and Databricks, two organizations that reshaped commercial analytics infrastructure over the past decade by building businesses around open-source foundations.
I want to do the same for Search: keep the infrastructure open and productize the experience around it.
Before putting a major breaking release on “main” branch and launching an integrated offering around it, I want to make the transition as smooth as possible for my users.
So I’m inviting a few design partners to join me in final benchmarking and help shape the UX and C ABI before the public release. Your workloads should have a say in what ships.
I have no doubt USearch v3 will be a “scale unlocker” for many organizations reliant on Search. With a little creativity and collaboration, it can become a “capability unlocker” as well.
If you’re working at a frontier intelligence lab, in autonomous driving, or in life sciences, and managing 100+ TB of data, DM me with a brief description of your workload. I’d love to collaborate!
Big congratulations to @glcst and @penberg and the rest of the team. Well deserved!
Turso is joining Supabase to give every agent its own database
turso.tech/blog/turso-is-…
Turso is joining Supabase to give every agent its own database
turso.tech/blog/turso-is-…
My laptop mic didn't work, so I asked claude command line to fix it. As expected, it just fixed. Now it loads the required skill automatically, like the matrix. So convenient, so scary
❯ my internal microphone doesn't work on the laptop, fix it
● Skill(superpowers:systematic-debugging)
Successfully loaded skill
Using systematic-debugging. Let me gather evidence across the audio stack layers before changing anything.
Ran 2 shell commands
Evidence so far: Lunar Lake-M audio controller is bound to snd_hda_intel (legacy HDA), not snd_sof_pci_intel_lnl (SOF). On Lunar Lake the internal mic array is usually a DMIC/SoundWire device that only SOF exposes. Let me verify that before concluding.
Ran 2 shell commands
Found a recent change. /etc/modprobe.d/sof-to-hda.conf forces dsp_driver=1 (legacy HDA) as a workaround. On Lunar Lake the internal DMIC array is only exposed by SOF — legacy HDA gives you only the analog codec (headset jack). Let me confirm the DMIC exists and understand the workaround it was papering over.
Ran 2 shell commands
DMIC endpoints are present in the NHLT table (LENOVO TP-R34) and SOF firmware + LNL topologies are installed. The workaround dates to 2026-07-21. Before removing it, I need to know whether the SOF error it worked around still exists.
Ran 2 shell commands
Journal retains the July boots. Let me read the actual failure that prompted the workaround.
Ran 1 shell command
Root cause is now clear. Reading the July logs closely:
- DMICs detected in NHLT tables: 2 — the internal mic is a 2-ch DMIC array, reachable only via SOF ("Digital mics found on Skylake+ platform, using SOF driver")
- The error 104 that prompted the workaround was failed to create module drc.2.1 on pcm0 (HDA Analog), dir 0 — direction 0 is playback. So the original bug broke speakers, not the mic.
The workaround traded a broken speaker pipeline for a broken internal mic. Let me capture the current baseline and check what's changed since July.
Ran 2 shell commands
Baseline captured. Note pactl is absent — let me identify the sound server, then check whether the dsp_driver param can be flipped at runtime so I can test SOF reversibly instead of guessing.
Ran 2 shell commands
Now I have everything I need for a hypothesis.
Root cause: The internal mic is a 2-channel DMIC array wired to the SoC audio DSP. Only the SOF driver can reach it. /etc/modprobe.d/sof-to-hda.conf forces dsp_driver=1 (legacy snd_hda_intel), which has no DMIC path — so the array simply never appears. WirePlumber confirms only one source exists (the analog headset-jack mic).
The dsp_driver param is read-only at runtime, so I'll test by reloading the module stack — fully reversible, no reboot needed.
Ran 2 shell commands
Testing SOF now — audio will drop for a few seconds and come back.
Ran 1 shell command
Clean switch. Now the critical checks — did the DMIC appear, and did error 104 return?
Ran 1 shell command
Confirmed. device 6: DMIC Raw now exists, and there's no error 104. Note the key difference from July: the loadable firmware library now loads at 2.14.1.1, matching the booted firmware. In July it loaded 2.13.0.1 against firmware 2.14.1.1 — that mismatch is exactly why the drc module failed to instantiate. The August firmware refresh fixed it.
Error 104 struck at stream start, not probe, so let me actually exercise both pipelines — including the playback path that originally failed.
Ran 1 shell command
Both mics now present in PipeWire — 59. Digital Microphone is the internal array, and it's the default. Playback on the formerly-failing pipeline returned rc=0. Let me confirm the DMIC actually captures audio rather than just enumerating.
Ran 1 shell command
audioop was removed in recent Python. Computing levels directly:
Ran 1 shell command
The internal mic captures real audio — peak −35 dBFS, RMS 95.7 (room ambience, not a dead channel). Hypothesis confirmed. Now making it permanent.
Ran 1 shell command
No options line remains (those are just symbol aliases). Confirming and rebuilding initramfs, since Ubuntu copies /etc/modprobe.d into it:
Ran 1 shell command
Initramfs rebuilt for all kernels, including 7.0.0-31 (installed, will be used at next boot). Final verification pass:
Ran 3 shell commands
Fixed. The internal mic works now — no reboot needed.
Root cause
Your internal mic isn't an analog mic. On this ThinkPad E14 Gen 7 (Lunar Lake) it's a 2-channel DMIC array wired to the SoC audio DSP — the kernel sees it as DMICs detected in NHLT tables: 2. Only the SOF driver can reach it; legacy snd_hda_intel has no path to it at all.
/etc/modprobe.d/sof-to-hda.conf, created 2026-07-21, forced dsp_driver=1 to pin the legacy driver. So the array was never enumerated — WirePlumber only ever saw the analog headset-jack mic.
That workaround was written for a real bug, but a different one. The July logs show it failed at:
FW reported error: 104 - Other failure of module instance initialization request
failed to create module drc.2.1
sof_pcm_setup_connected_widgets: pcm0 (HDA Analog), dir 0
dir 0 is playback — the original bug broke your speakers, and the fix sacrificed the mic to get them back.
And that bug is already gone. The July boot loaded firmware 2.14.1.1 but a loadable library at 2.13.0.1. The drc module lives in that library, so the version mismatch made it fail to instantiate. Your 2026-08-25 linux-firmware/firmware-sof-signed refresh aligned both at 2.14.1.1 — the workaround has been obsolete for about a month.
What I did
Rather than guess, I switched the module stack to SOF at runtime (reversible, no reboot) and tested the formerly-failing path before committing to anything:
┌─────────────────────────────────────┬───────────────────────────────────────┐
│ Check │ Result │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ Firmware library version │ 2.14.1.1 — matches booted FW │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ Error-104-class messages │ 0 │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ Speaker playback on pcm0 HDA Analog │ rc=0, clean │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ DMIC Raw capture device │ present (hw:0,6) │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ Live capture │ −24.8 dBFS peak, RMS 112 — real audio │
└─────────────────────────────────────┴───────────────────────────────────────┘
Then I commented out the options line (keeping the full history and a revert recipe in the file), and rebuilt initramfs for all three kernels — Ubuntu bakes /etc/modprobe.d into it, so editing the file alone would not have survived a reboot. I verified the next-boot image (7.0.0-31) contains no active override.
Both mics are now available, with the internal one as default:
58. ... Stereo Microphone (headset jack)
59. ... Digital Microphone * (internal array)
One thing to know
You have 7.0.0-31 installed but are running 7.0.0-30 — a reboot was already pending. I validated SOF on -30, so next boot lands on a kernel I couldn't test. Risk is low (newer kernel, and the cause was firmware-side), but if audio misbehaves after rebooting, uncomment the last line of /etc/modprobe.d/sof-to-hda.conf and run sudo update-initramfs -u to get speakers back immediately.
❯ my internal microphone doesn't work on the laptop, fix it
● Skill(superpowers:systematic-debugging)
Successfully loaded skill
Using systematic-debugging. Let me gather evidence across the audio stack layers before changing anything.
Ran 2 shell commands
Evidence so far: Lunar Lake-M audio controller is bound to snd_hda_intel (legacy HDA), not snd_sof_pci_intel_lnl (SOF). On Lunar Lake the internal mic array is usually a DMIC/SoundWire device that only SOF exposes. Let me verify that before concluding.
Ran 2 shell commands
Found a recent change. /etc/modprobe.d/sof-to-hda.conf forces dsp_driver=1 (legacy HDA) as a workaround. On Lunar Lake the internal DMIC array is only exposed by SOF — legacy HDA gives you only the analog codec (headset jack). Let me confirm the DMIC exists and understand the workaround it was papering over.
Ran 2 shell commands
DMIC endpoints are present in the NHLT table (LENOVO TP-R34) and SOF firmware + LNL topologies are installed. The workaround dates to 2026-07-21. Before removing it, I need to know whether the SOF error it worked around still exists.
Ran 2 shell commands
Journal retains the July boots. Let me read the actual failure that prompted the workaround.
Ran 1 shell command
Root cause is now clear. Reading the July logs closely:
- DMICs detected in NHLT tables: 2 — the internal mic is a 2-ch DMIC array, reachable only via SOF ("Digital mics found on Skylake+ platform, using SOF driver")
- The error 104 that prompted the workaround was failed to create module drc.2.1 on pcm0 (HDA Analog), dir 0 — direction 0 is playback. So the original bug broke speakers, not the mic.
The workaround traded a broken speaker pipeline for a broken internal mic. Let me capture the current baseline and check what's changed since July.
Ran 2 shell commands
Baseline captured. Note pactl is absent — let me identify the sound server, then check whether the dsp_driver param can be flipped at runtime so I can test SOF reversibly instead of guessing.
Ran 2 shell commands
Now I have everything I need for a hypothesis.
Root cause: The internal mic is a 2-channel DMIC array wired to the SoC audio DSP. Only the SOF driver can reach it. /etc/modprobe.d/sof-to-hda.conf forces dsp_driver=1 (legacy snd_hda_intel), which has no DMIC path — so the array simply never appears. WirePlumber confirms only one source exists (the analog headset-jack mic).
The dsp_driver param is read-only at runtime, so I'll test by reloading the module stack — fully reversible, no reboot needed.
Ran 2 shell commands
Testing SOF now — audio will drop for a few seconds and come back.
Ran 1 shell command
Clean switch. Now the critical checks — did the DMIC appear, and did error 104 return?
Ran 1 shell command
Confirmed. device 6: DMIC Raw now exists, and there's no error 104. Note the key difference from July: the loadable firmware library now loads at 2.14.1.1, matching the booted firmware. In July it loaded 2.13.0.1 against firmware 2.14.1.1 — that mismatch is exactly why the drc module failed to instantiate. The August firmware refresh fixed it.
Error 104 struck at stream start, not probe, so let me actually exercise both pipelines — including the playback path that originally failed.
Ran 1 shell command
Both mics now present in PipeWire — 59. Digital Microphone is the internal array, and it's the default. Playback on the formerly-failing pipeline returned rc=0. Let me confirm the DMIC actually captures audio rather than just enumerating.
Ran 1 shell command
audioop was removed in recent Python. Computing levels directly:
Ran 1 shell command
The internal mic captures real audio — peak −35 dBFS, RMS 95.7 (room ambience, not a dead channel). Hypothesis confirmed. Now making it permanent.
Ran 1 shell command
No options line remains (those are just symbol aliases). Confirming and rebuilding initramfs, since Ubuntu copies /etc/modprobe.d into it:
Ran 1 shell command
Initramfs rebuilt for all kernels, including 7.0.0-31 (installed, will be used at next boot). Final verification pass:
Ran 3 shell commands
Fixed. The internal mic works now — no reboot needed.
Root cause
Your internal mic isn't an analog mic. On this ThinkPad E14 Gen 7 (Lunar Lake) it's a 2-channel DMIC array wired to the SoC audio DSP — the kernel sees it as DMICs detected in NHLT tables: 2. Only the SOF driver can reach it; legacy snd_hda_intel has no path to it at all.
/etc/modprobe.d/sof-to-hda.conf, created 2026-07-21, forced dsp_driver=1 to pin the legacy driver. So the array was never enumerated — WirePlumber only ever saw the analog headset-jack mic.
That workaround was written for a real bug, but a different one. The July logs show it failed at:
FW reported error: 104 - Other failure of module instance initialization request
failed to create module drc.2.1
sof_pcm_setup_connected_widgets: pcm0 (HDA Analog), dir 0
dir 0 is playback — the original bug broke your speakers, and the fix sacrificed the mic to get them back.
And that bug is already gone. The July boot loaded firmware 2.14.1.1 but a loadable library at 2.13.0.1. The drc module lives in that library, so the version mismatch made it fail to instantiate. Your 2026-08-25 linux-firmware/firmware-sof-signed refresh aligned both at 2.14.1.1 — the workaround has been obsolete for about a month.
What I did
Rather than guess, I switched the module stack to SOF at runtime (reversible, no reboot) and tested the formerly-failing path before committing to anything:
┌─────────────────────────────────────┬───────────────────────────────────────┐
│ Check │ Result │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ Firmware library version │ 2.14.1.1 — matches booted FW │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ Error-104-class messages │ 0 │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ Speaker playback on pcm0 HDA Analog │ rc=0, clean │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ DMIC Raw capture device │ present (hw:0,6) │
├─────────────────────────────────────┼───────────────────────────────────────┤
│ Live capture │ −24.8 dBFS peak, RMS 112 — real audio │
└─────────────────────────────────────┴───────────────────────────────────────┘
Then I commented out the options line (keeping the full history and a revert recipe in the file), and rebuilt initramfs for all three kernels — Ubuntu bakes /etc/modprobe.d into it, so editing the file alone would not have survived a reboot. I verified the next-boot image (7.0.0-31) contains no active override.
Both mics are now available, with the internal one as default:
58. ... Stereo Microphone (headset jack)
59. ... Digital Microphone * (internal array)
One thing to know
You have 7.0.0-31 installed but are running 7.0.0-30 — a reboot was already pending. I validated SOF on -30, so next boot lands on a kernel I couldn't test. Risk is low (newer kernel, and the cause was firmware-side), but if audio misbehaves after rebooting, uncomment the last line of /etc/modprobe.d/sof-to-hda.conf and run sudo update-initramfs -u to get speakers back immediately.