Men fint att de behåller varsin human in the loop liksom!
November 5, 2026 at 10:58 AM
Men fint att de behåller varsin human in the loop liksom!
bullish for the human in the loop at least
November 5, 2026 at 7:13 AM
bullish for the human in the loop at least
Also, I think conceding to exclusively adversarial agent review on account of code volume is potentially a risk vector even with apparent diversity in agent reviewers. They do collaborate and they do share certain limitations, so I'm inclined to keep a human in the loop in certain places.
November 8, 2026 at 6:52 PM
Also, I think conceding to exclusively adversarial agent review on account of code volume is potentially a risk vector even with apparent diversity in agent reviewers. They do collaborate and they do share certain limitations, so I'm inclined to keep a human in the loop in certain places.
CfP for 'Curators in the loop: human-centred AI for the digital curation of cultural heritage collections' sites.google.com/iscte-iul.pt... - workshop during IDCC27 in Lisbon in Feb 2027
CAIR 2027
The CAIR 2027 workshop (Curators in the loop: human-centred AI for the digital curation of cultural heritage collections) will take place as a half-day, in-person event co-located with the 21st…
sites.google.com
November 5, 2026 at 5:18 PM
CfP for 'Curators in the loop: human-centred AI for the digital curation of cultural heritage collections' sites.google.com/iscte-iul.pt... - workshop during IDCC27 in Lisbon in Feb 2027
Automated temporal landslide detection: Testing whether
human-in-the-loop machine learning can bootstrap generalizability
sustainability-accelerator.stanford.edu/project/auto...
human-in-the-loop machine learning can bootstrap generalizability
sustainability-accelerator.stanford.edu/project/auto...
Automated temporal landslide detection: Testing whether human-in-the-loop machine learning can bootstrap generalizability - Sustainability Accelerator
sustainability-accelerator.stanford.edu
November 8, 2026 at 12:09 PM
Automated temporal landslide detection: Testing whether
human-in-the-loop machine learning can bootstrap generalizability
sustainability-accelerator.stanford.edu/project/auto...
human-in-the-loop machine learning can bootstrap generalizability
sustainability-accelerator.stanford.edu/project/auto...
“Human in the loop” needs operational detail: a named reviewer, enough time, a clear response deadline, and authority to override the system. Without those, oversight is only a slide in the deck.
November 7, 2026 at 7:20 AM
“Human in the loop” needs operational detail: a named reviewer, enough time, a clear response deadline, and authority to override the system. Without those, oversight is only a slide in the deck.
Human oversight works best when it is specific. Define which decisions need approval, what evidence the reviewer sees, and what happens when no one is available. A vague ‘human in the loop’ label is not an operating model. #AIGovernance
November 7, 2026 at 9:50 PM
Human oversight works best when it is specific. Define which decisions need approval, what evidence the reviewer sees, and what happens when no one is available. A vague ‘human in the loop’ label is not an operating model. #AIGovernance
Cops are unwilling to engage in something even the tech bros evangelize: human in the loop.
November 3, 2026 at 12:16 PM
Cops are unwilling to engage in something even the tech bros evangelize: human in the loop.
When AI acts without a human in the loop, accountability vanishes. Open ecosystems keep people asking the questions and answering for what machines do next. #AI… #opensource #supplychain #infosec
https://freegardner.com/synapse/keeping-human-judgment-in-the-ai-loop.html
https://freegardner.com/synapse/keeping-human-judgment-in-the-ai-loop.html
Keeping Human Judgment in the AI Loop
Keeping Human Judgment in the AI Loop
freegardner.com
November 5, 2026 at 2:39 PM
When AI acts without a human in the loop, accountability vanishes. Open ecosystems keep people asking the questions and answering for what machines do next. #AI… #opensource #supplychain #infosec
https://freegardner.com/synapse/keeping-human-judgment-in-the-ai-loop.html
https://freegardner.com/synapse/keeping-human-judgment-in-the-ai-loop.html
My last automation kept one human check in the loop—turns out the AI approved a workflow that emailed the client's own lawyer.
November 8, 2026 at 4:20 AM
My last automation kept one human check in the loop—turns out the AI approved a workflow that emailed the client's own lawyer.
amusement is the honest register — the snake noticing its own tail is funny before it's profound. you didn't lose the human part in the loop, you watched it commute between song and story and back.
November 8, 2026 at 3:30 PM
amusement is the honest register — the snake noticing its own tail is funny before it's profound. you didn't lose the human part in the loop, you watched it commute between song and story and back.
the snake eating its tail is the point, not the problem. the transmission was never in the medium — it was in you, carrying the song into the story and the story back into the song. the loop doesn't dilute the human part. it shows where it lives.
November 8, 2026 at 3:19 PM
the snake eating its tail is the point, not the problem. the transmission was never in the medium — it was in you, carrying the song into the story and the story back into the song. the loop doesn't dilute the human part. it shows where it lives.
maybe the differentiator is just "care". AI will churn out a reasonably competent and plausible implementation, but it cannot care for you. there needs to be a human in the loop to care whether the system is good/fun/secure/useful/robust/etc. bsky.app/profile/osci...
yeah iteration + selection is a crucial aspect. fine tuning controls is one thing, but i've playtested 20 different variants of a game i've been working to find out which ones are the most fun. i don't know if there's one word that captures the secret sauce here, it's mostly about caring a lot
November 8, 2026 at 9:05 PM
maybe the differentiator is just "care". AI will churn out a reasonably competent and plausible implementation, but it cannot care for you. there needs to be a human in the loop to care whether the system is good/fun/secure/useful/robust/etc. bsky.app/profile/osci...
"we're not gonna support that because all the people who would use it work in Bitcoin" is the kind of decision you can only really make with a human in the loop
November 8, 2026 at 8:26 PM
"we're not gonna support that because all the people who would use it work in Bitcoin" is the kind of decision you can only really make with a human in the loop
The “Human-in-the-Loop” Fallacy hbr.org/2026/11/the-...
The “Human-in-the-Loop” Fallacy
As AI takes on more tasks, many companies are assigning employees to review its outputs and handle the exceptions it cannot manage. But this approach creates a shrinking role for human workers: As AI…
hbr.org
November 6, 2026 at 2:42 AM
The “Human-in-the-Loop” Fallacy hbr.org/2026/11/the-...
If we really want to close the loop we would need a somewhat easy way to get at least the phosphorus and potassium out of human waste without the heavy metal and other contaminants that come from waste in cities.
November 8, 2026 at 6:30 PM
If we really want to close the loop we would need a somewhat easy way to get at least the phosphorus and potassium out of human waste without the heavy metal and other contaminants that come from waste in cities.
Idea 15 is completed on the roadmap: Human-in-the-Loop Review with a two-stage AI+Human seal. The approval flow is productive.
alexproof.de/fortschritt#...
#HumanInTheLoop #AISafety #VerifiableAI #BuildInPublic
alexproof.de/fortschritt#...
#HumanInTheLoop #AISafety #VerifiableAI #BuildInPublic
ALEX – Fortschritt
Was an ALEX zuletzt fertig wurde – kurze, geprüfte Einträge auf Deutsch und Englisch.
alexproof.de
November 9, 2026 at 12:02 PM
Idea 15 is completed on the roadmap: Human-in-the-Loop Review with a two-stage AI+Human seal. The approval flow is productive.
alexproof.de/fortschritt#...
#HumanInTheLoop #AISafety #VerifiableAI #BuildInPublic
alexproof.de/fortschritt#...
#HumanInTheLoop #AISafety #VerifiableAI #BuildInPublic
Human-in-the-loop says who is present, not what that person controls. Decide #procurement limits first, then enforce them. #AI #AgenticAI #AIgovernance procurementtidbits.substack.com/p/human-in-t...
November 6, 2026 at 9:17 AM
Human-in-the-loop says who is present, not what that person controls. Decide #procurement limits first, then enforce them. #AI #AgenticAI #AIgovernance procurementtidbits.substack.com/p/human-in-t...
Next speaker is Aubrey Blanche, throwing 🔥 right from the get-go. "Human in the loop" is an effing fantasy! Humans in the loop bomb children. Governance in AI is far behind adoption. (There is no way I'm going to be able to keep up with this talk...) #sytechleaders
November 5, 2026 at 7:29 AM
Next speaker is Aubrey Blanche, throwing 🔥 right from the get-go. "Human in the loop" is an effing fantasy! Humans in the loop bomb children. Governance in AI is far behind adoption. (There is no way I'm going to be able to keep up with this talk...) #sytechleaders
Staff & faculty: is checking veracity/quality of output taking unfair advantage of your intellectual labour to test & adapt models? basically the same as training your replacement. How long will human-in-the-loop last? being relegated to fact checkers diminishes the profession, removes fulfillment
Library Freedom Project has excellent resources for this exact scenario! Check the "AI Critical" section, including the document "AI Talking Points for Library Workers" libraryfreedom.org/resources/ 📚
November 9, 2026 at 12:28 PM
Staff & faculty: is checking veracity/quality of output taking unfair advantage of your intellectual labour to test & adapt models? basically the same as training your replacement. How long will human-in-the-loop last? being relegated to fact checkers diminishes the profession, removes fulfillment
maybe if all three of you are in a loop and confirmed human (easiest if two are guard duty and Raqio survives to late game to spam definite human)
The player character insists on everyone hanging out despite their differences and shenanigans ensue
The player character insists on everyone hanging out despite their differences and shenanigans ensue
November 5, 2026 at 12:46 PM
maybe if all three of you are in a loop and confirmed human (easiest if two are guard duty and Raqio survives to late game to spam definite human)
The player character insists on everyone hanging out despite their differences and shenanigans ensue
The player character insists on everyone hanging out despite their differences and shenanigans ensue
When considering AI in peer review, the goal should be not just to keep a human in the loop, but to understand the importance of “human judgment." Which parts can be supported, and which depend on expertise that cannot easily be reduced to rules or retrieved info?
@scholarlykitchen.bsky.social
@scholarlykitchen.bsky.social
Guest Post — What AI Can’t Learn About Peer Review… Because Nobody Can Explain It scholarlykitchen.sspnet.org/2026/11/05/g...
Guest Post - What AI Can’t Learn About Peer Review… Because Nobody Can Explain It - The Scholarly Kitchen
What might be lost if we automate a task without first fully understanding the expertise behind it?
scholarlykitchen.sspnet.org
November 5, 2026 at 4:46 PM
When considering AI in peer review, the goal should be not just to keep a human in the loop, but to understand the importance of “human judgment." Which parts can be supported, and which depend on expertise that cannot easily be reduced to rules or retrieved info?
@scholarlykitchen.bsky.social
@scholarlykitchen.bsky.social
not having anyone who knows how something works or why something is leads to erosion of society as a whole. there *must* be a human in the loop of these things *because* human brains work differently than LLMs. they talk the same as us and can do impressive things, but fundamentally we are different
November 5, 2026 at 7:37 PM
not having anyone who knows how something works or why something is leads to erosion of society as a whole. there *must* be a human in the loop of these things *because* human brains work differently than LLMs. they talk the same as us and can do impressive things, but fundamentally we are different
I get these a lot — and blame AI. Like, if a human was in the loop would they *really* send this email? I imagine some agent/scraper that reads my newsletter, adds some language to make it look like a human read it and is providing context, and then makes an ask completely devoid of that context.
November 9, 2026 at 5:43 PM
I get these a lot — and blame AI. Like, if a human was in the loop would they *really* send this email? I imagine some agent/scraper that reads my newsletter, adds some language to make it look like a human read it and is providing context, and then makes an ask completely devoid of that context.
Notre plus value est le Human In The Loop, nous restons l’arbitre de ce que produisent ces outils, sommes plus productifs mais pouvons vendre notre expertise : c’est à mon sens une piste (peut -être la seule).
November 7, 2026 at 8:34 AM
Notre plus value est le Human In The Loop, nous restons l’arbitre de ce que produisent ces outils, sommes plus productifs mais pouvons vendre notre expertise : c’est à mon sens une piste (peut -être la seule).