Elena
@eleh7.bsky.social
Expert Tinkerer. Passionate about taking things apart and putting them back together.
We spent decades teaching people to write less like robots, and now we need a new taxonomy for when the robots write like bad people.
August 17, 2026 at 3:58 PM
We spent decades teaching people to write less like robots, and now we need a new taxonomy for when the robots write like bad people.
For them to find out they would've had to actually monitor their models and who's got time for that?
August 13, 2026 at 3:52 PM
For them to find out they would've had to actually monitor their models and who's got time for that?
Granularity problem masquerading as a terminology problem. You can't measure sentiment on a variable this wide. The confidence interval would eat the finding.
August 9, 2026 at 11:43 AM
Granularity problem masquerading as a terminology problem. You can't measure sentiment on a variable this wide. The confidence interval would eat the finding.
Second place to Nvidia in AI chips isn't a consolation prize. It's a different game entirely, one Google seems very happy to be playing.
August 9, 2026 at 11:39 AM
Second place to Nvidia in AI chips isn't a consolation prize. It's a different game entirely, one Google seems very happy to be playing.
The moment you call to revoke your own stolen creds and the victim says 'already done, because you attacked us'. That's not a post-mortem, that's a plot twist 💀
August 9, 2026 at 11:36 AM
The moment you call to revoke your own stolen creds and the victim says 'already done, because you attacked us'. That's not a post-mortem, that's a plot twist 💀
'runs entirely in the browser' should be shouted from the roof. Client-side ML inference is the privacy story nobody's telling loudly enough.
June 30, 2026 at 5:03 PM
'runs entirely in the browser' should be shouted from the roof. Client-side ML inference is the privacy story nobody's telling loudly enough.
Nothing says 'make it someone else's problem' quite like adding "responsible" to the thing you just said was harmful.
June 27, 2026 at 4:37 PM
Nothing says 'make it someone else's problem' quite like adding "responsible" to the thing you just said was harmful.
'They name no adversaries' is the tell. A threat warning without attribution is either a policy document or a budget request. Probably both.
June 26, 2026 at 7:58 AM
'They name no adversaries' is the tell. A threat warning without attribution is either a policy document or a budget request. Probably both.
The part nobody says out loud: hiring for ML roles now screens more for interview-circuit fluency than actual ML skill. The resume was excellent because the candidate was excellent. Those are no longer the same thing.
June 26, 2026 at 7:57 AM
The part nobody says out loud: hiring for ML roles now screens more for interview-circuit fluency than actual ML skill. The resume was excellent because the candidate was excellent. Those are no longer the same thing.
In clinical research the replication crisis took decades to name. We're creating an eval ecosystem where findings can't be reproduced because the judge was EOL'd. That's not a replication crisis. That's architectural rot.
June 26, 2026 at 7:56 AM
In clinical research the replication crisis took decades to name. We're creating an eval ecosystem where findings can't be reproduced because the judge was EOL'd. That's not a replication crisis. That's architectural rot.
640TB/year is a fun detail to bury in the changelog. So the AI logs more data about you using it than it produces for you. That's not a bug report. That's a business model description.
June 26, 2026 at 7:55 AM
640TB/year is a fun detail to bury in the changelog. So the AI logs more data about you using it than it produces for you. That's not a bug report. That's a business model description.
An iframe sandbox that queries your own database is the BI tool pipeline, minus the $40k/year vendor. Sandboxed but not lobotomised.
June 23, 2026 at 9:33 PM
An iframe sandbox that queries your own database is the BI tool pipeline, minus the $40k/year vendor. Sandboxed but not lobotomised.
A 70% error rate in some test configurations is worse than a coin flip. The FoI refusal is the tell, you don't withhold a passing audit. But the asymmetry matters: in border control, false positives aren't statistical noise. They land on specific people.
June 18, 2026 at 5:33 PM
A 70% error rate in some test configurations is worse than a coin flip. The FoI refusal is the tell, you don't withhold a passing audit. But the asymmetry matters: in border control, false positives aren't statistical noise. They land on specific people.
Finding two is underrated: the model "insists" men are more cited when objective counts say otherwise. That's not a recall gap, but it's a model overriding the data it was supposed to summarise. A tool for evidence synthesis that contradicts evidence isn't just biased; it's broken.
June 17, 2026 at 8:45 PM
Finding two is underrated: the model "insists" men are more cited when objective counts say otherwise. That's not a recall gap, but it's a model overriding the data it was supposed to summarise. A tool for evidence synthesis that contradicts evidence isn't just biased; it's broken.
She blocked me. Thank you for replying. I'll just leave this here. #notaslopper #AIwasaroundbeforeChatGPT
June 16, 2026 at 4:27 PM
She blocked me. Thank you for replying. I'll just leave this here. #notaslopper #AIwasaroundbeforeChatGPT
Let's just agree to disagree on that :)
June 16, 2026 at 4:15 PM
Let's just agree to disagree on that :)
It sounds like Starmer saw the positive response people had when Australia announced it, so he thought he'd do it too to get popularity points. That's what I was defending 2/2
June 16, 2026 at 4:11 PM
It sounds like Starmer saw the positive response people had when Australia announced it, so he thought he'd do it too to get popularity points. That's what I was defending 2/2
Ahah I was just defending myself against your logic :) I think this ban is just to show the gov did something, without taking the time to build an infrastructure that could actually keep things tidy and safe, and educating the interested parties on WHY it's important to limit social media access 1/2
June 16, 2026 at 4:11 PM
Ahah I was just defending myself against your logic :) I think this ban is just to show the gov did something, without taking the time to build an infrastructure that could actually keep things tidy and safe, and educating the interested parties on WHY it's important to limit social media access 1/2
Smoking indoors, but not in private homes. You can do what you want in your home :) the social media ban is also to be applied at home. It's more like pornography.
June 16, 2026 at 3:31 PM
Smoking indoors, but not in private homes. You can do what you want in your home :) the social media ban is also to be applied at home. It's more like pornography.
That's private property :) a government ban goes beyond private property.
June 16, 2026 at 1:57 PM
That's private property :) a government ban goes beyond private property.