copilot messages with threatening auras
November 3, 2026 at 12:23 AM
copilot messages with threatening auras
Franklin distinguishes holistic and prescriptive tech. Holistic involves full control from start to finish, as a potter controlling clay, molding, glazing, & baking. Prescriptive divides labour into delegatable subtasks, as in Chinese bronze casts, Roman terra sigilata, or the modern assembly line
November 2, 2026 at 4:27 PM
Franklin distinguishes holistic and prescriptive tech. Holistic involves full control from start to finish, as a potter controlling clay, molding, glazing, & baking. Prescriptive divides labour into delegatable subtasks, as in Chinese bronze casts, Roman terra sigilata, or the modern assembly line
yeah... i think its about time to have astra go and de-claude things again. Unfortunately its probably going to split the task into 12952 subtasks, each of which it will work on independently and produce detailed plans for
October 30, 2026 at 3:49 AM
yeah... i think its about time to have astra go and de-claude things again. Unfortunately its probably going to split the task into 12952 subtasks, each of which it will work on independently and produce detailed plans for
A task with subtasks says so in Pellet's table: a small finished-over-total count beside its title.
Turn on Hierarchy and children are pulled under their parent and indented, so the table reads as the tree the work actually is.
https://youtu.be/lizlGIW_LIo
Turn on Hierarchy and children are pulled under their parent and indented, so the table reads as the tree the work actually is.
https://youtu.be/lizlGIW_LIo
Pellet: subtasks in the table [U]
YouTube · Pellet
www.youtube.com
October 31, 2026 at 2:00 PM
A task with subtasks says so in Pellet's table: a small finished-over-total count beside its title.
Turn on Hierarchy and children are pulled under their parent and indented, so the table reads as the tree the work actually is.
https://youtu.be/lizlGIW_LIo
Turn on Hierarchy and children are pulled under their parent and indented, so the table reads as the tree the work actually is.
https://youtu.be/lizlGIW_LIo
it will have:
a to do list with categories and active tasks and subtasks and priorities
an archive
alert sounds options (including "none")
alert themes (you can even add your own)
color themes (with color change when it switches between work and break)
strict mode
a to do list with categories and active tasks and subtasks and priorities
an archive
alert sounds options (including "none")
alert themes (you can even add your own)
color themes (with color change when it switches between work and break)
strict mode
the pomodoro timer has become a Project. thankfully for most of it i can build off what i did with the habit trackers. but i haven't even looked at how to make the timer. lol. i've just been doing skeletal stuff for the most part
November 1, 2026 at 8:28 PM
it will have:
a to do list with categories and active tasks and subtasks and priorities
an archive
alert sounds options (including "none")
alert themes (you can even add your own)
color themes (with color change when it switches between work and break)
strict mode
a to do list with categories and active tasks and subtasks and priorities
an archive
alert sounds options (including "none")
alert themes (you can even add your own)
color themes (with color change when it switches between work and break)
strict mode
WISE: A Long-Horizon Agent in Minecraft with Why-Which Reasoning
Cheng Renmin, Changhao Chen
Action editor: Mengmi Zhang
https://openreview.net/forum?id=T8HuiP3yM9
#subtasks #tasks #affordance
Cheng Renmin, Changhao Chen
Action editor: Mengmi Zhang
https://openreview.net/forum?id=T8HuiP3yM9
#subtasks #tasks #affordance
October 30, 2026 at 8:21 PM
WISE: A Long-Horizon Agent in Minecraft with Why-Which Reasoning
Cheng Renmin, Changhao Chen
Action editor: Mengmi Zhang
https://openreview.net/forum?id=T8HuiP3yM9
#subtasks #tasks #affordance
Cheng Renmin, Changhao Chen
Action editor: Mengmi Zhang
https://openreview.net/forum?id=T8HuiP3yM9
#subtasks #tasks #affordance
Agents can hit 93% on individual subtasks yet fall to 74% across clean 20-subtask chains. The costly failure is often recovery: corruption went unnoticed in 41% of trials, showing why step-level monitoring matters more than a single end-to-end score.
Why AI Agent Workflows Blow Up Past the Wrong Threshold | 1 Minute Signal
A short, bounded task and a long dependent workflow do not fail in the same way. Drafting a single email, fixing one bug, or answering one question may…
1minutesignal.com
November 1, 2026 at 5:01 AM
Agents can hit 93% on individual subtasks yet fall to 74% across clean 20-subtask chains. The costly failure is often recovery: corruption went unnoticed in 41% of trials, showing why step-level monitoring matters more than a single end-to-end score.
Big task feeling overwhelming? 4.Do's AI suggests subtasks that break it into steps you can actually start.
Get it at https://4do.app
#Productivity #TaskManagement #AI #GetThingsDone #iOS #Android
Get it at https://4do.app
#Productivity #TaskManagement #AI #GetThingsDone #iOS #Android
October 28, 2026 at 2:01 PM
Big task feeling overwhelming? 4.Do's AI suggests subtasks that break it into steps you can actually start.
Get it at https://4do.app
#Productivity #TaskManagement #AI #GetThingsDone #iOS #Android
Get it at https://4do.app
#Productivity #TaskManagement #AI #GetThingsDone #iOS #Android
Native project management is coming to GoHighLevel. I've been in the beta for 2 months.
New video walks through boards, subtasks, custom statuses and task automations, plus where AI fits in.
Link below 👇
#gohighlevel #projectmanagement
New video walks through boards, subtasks, custom statuses and task automations, plus where AI fits in.
Link below 👇
#gohighlevel #projectmanagement
October 28, 2026 at 5:15 PM
Native project management is coming to GoHighLevel. I've been in the beta for 2 months.
New video walks through boards, subtasks, custom statuses and task automations, plus where AI fits in.
Link below 👇
#gohighlevel #projectmanagement
New video walks through boards, subtasks, custom statuses and task automations, plus where AI fits in.
Link below 👇
#gohighlevel #projectmanagement
Who is gonna submit to our power ? We r gonna be together all day long. Sessions, rates, tasks, expose, vidcall and more... Dont miss ur chance!
♦ @goddessjuli3t.bsky.social and me ♦
.
#Femdom #SubTasks #DommeTasks #Findom #Finsub #ObedienceTraining #SubmissiveMen
♦ @goddessjuli3t.bsky.social and me ♦
.
#Femdom #SubTasks #DommeTasks #Findom #Finsub #ObedienceTraining #SubmissiveMen
October 26, 2026 at 6:18 PM
Who is gonna submit to our power ? We r gonna be together all day long. Sessions, rates, tasks, expose, vidcall and more... Dont miss ur chance!
♦ @goddessjuli3t.bsky.social and me ♦
.
#Femdom #SubTasks #DommeTasks #Findom #Finsub #ObedienceTraining #SubmissiveMen
♦ @goddessjuli3t.bsky.social and me ♦
.
#Femdom #SubTasks #DommeTasks #Findom #Finsub #ObedienceTraining #SubmissiveMen
(8/14) In HealthAdminBench, Claude Opus 4.6 completed just 36.3% of administrative workflows end to end. GPT-5.4 succeeded on 82.8% of individual subtasks but finished only 26.7% of entire workflows.
October 28, 2026 at 12:11 AM
(8/14) In HealthAdminBench, Claude Opus 4.6 completed just 36.3% of administrative workflows end to end. GPT-5.4 succeeded on 82.8% of individual subtasks but finished only 26.7% of entire workflows.
Drew Breunig’s October 24 talk argues for using deterministic code to select tools, isolate subtasks, and manage context around probabilistic models.
Drew Breunig argues that context engineering is becoming ordinary software
Drew Breunig’s October 24 talk argues for using deterministic code to select tools, isolate subtasks, and manage context around probabilistic models.
groundtruth.day
October 26, 2026 at 11:27 AM
Drew Breunig’s October 24 talk argues for using deterministic code to select tools, isolate subtasks, and manage context around probabilistic models.
the corresponding epic, and it should have hydrated the knowledge it needs. The ACs keep the agents (mostly 😉) honest, the comments they add can help debugging if impl turns out wonky, and subtasks are sized so they're sorta reasonable "unit of work Git commit", which helps at review time.
October 26, 2026 at 11:34 AM
the corresponding epic, and it should have hydrated the knowledge it needs. The ACs keep the agents (mostly 😉) honest, the comments they add can help debugging if impl turns out wonky, and subtasks are sized so they're sorta reasonable "unit of work Git commit", which helps at review time.
The Cheaper Model That Billed More
->StartupHub.ai | More on "Model efficiency and cost optimization" at BigEarthData.ai
->StartupHub.ai | More on "Model efficiency and cost optimization" at BigEarthData.ai
The Cheaper Model That Billed More
Rayan Krishnan of Vals AI told AI Engineer that Anthropic's $4-in/$20-out Opus 5.5 cost more per completed tax task than the $10-in/$50-out Fable 5.1 despite its lower list price. The gap came from token burn, not pricing. Krishnan said the smaller model scored 3% lower while trying to spend extra tokens to finish the work, so teams paid more for worse accuracy. That jagged frontier repeats across finance tasks, where Excel modeling in aggregate hides that LBO subtasks run on a lower Pareto curve and drop the supposed frontier model off the efficient edge. Vals AI now runs more than 6 trillion tokens of evals and ships those checks as Val Smith, which replays GitHub pull requests to test whether an agent could have done the human work and then routes each issue to the cheapest model that clears the bar...
www.startuphub.ai
October 30, 2026 at 8:39 AM
The Cheaper Model That Billed More
->StartupHub.ai | More on "Model efficiency and cost optimization" at BigEarthData.ai
->StartupHub.ai | More on "Model efficiency and cost optimization" at BigEarthData.ai
The Cheaper Model That Billed More
Rayan Krishnan of Vals AI told AI Engineer that Anthropic's $4-in/$20-out Opus 5.5 cost more per completed tax task than the $10-in/$50-out Fable 5.1 despite its lower list price. The gap came from token burn, not pricing. Krishnan said the smaller model scored 3% lower while trying to spend extra tokens to finish the work, so teams paid more for worse accuracy. That jagged frontier repeats across finance tasks, where Excel modeling in aggregate hides that LBO subtasks run on a lower Pareto curve and drop the supposed frontier model off the efficient edge. Vals AI now runs more than 6 trillion tokens of evals and ships those checks as Val Smith, which replays GitHub pull requests to test whether an agent could have done the human work and then routes each issue to the cheapest model that clears the bar...
www.startuphub.ai
October 30, 2026 at 8:39 AM
China Proposes Thresholds for AI Agents: 20 Subtasks, 60 Agents, 200 Tools
https://ai-release.org/posts/kitay-predlozhil-porogi-dlya-ai-agentov-20-podzadach-60-agentov-200-in-f858cdd2_en.html
https://ai-release.org/posts/kitay-predlozhil-porogi-dlya-ai-agentov-20-podzadach-60-agentov-200-in-f858cdd2_en.html
China Proposes Thresholds for AI Agents: 20 Subtasks, 60 Agents, 200 Tools
The gist in 5 seconds - Chinese authors have proposed three numerical thresholds for AI agents: 20 subtasks, 60 agents, 200 tools. - This is about a scale metri
ai-release.org
October 25, 2026 at 6:26 PM
China Proposes Thresholds for AI Agents: 20 Subtasks, 60 Agents, 200 Tools
https://ai-release.org/posts/kitay-predlozhil-porogi-dlya-ai-agentov-20-podzadach-60-agentov-200-in-f858cdd2_en.html
https://ai-release.org/posts/kitay-predlozhil-porogi-dlya-ai-agentov-20-podzadach-60-agentov-200-in-f858cdd2_en.html
Strategically breaking things down into subtasks before starting doesn't always work for me,
but what does is leaving a "where I left off" breadcrumb and a "what should I do next" breadcrumb.
Try it!
but what does is leaving a "where I left off" breadcrumb and a "what should I do next" breadcrumb.
Try it!
October 23, 2026 at 4:35 PM
Strategically breaking things down into subtasks before starting doesn't always work for me,
but what does is leaving a "where I left off" breadcrumb and a "what should I do next" breadcrumb.
Try it!
but what does is leaving a "where I left off" breadcrumb and a "what should I do next" breadcrumb.
Try it!
Stop staring at a massive, intimidating requirement and wondering where to start. Our AI Task Decomposition feature takes a complex project goal and instantly breaks it into well-structured subtasks with descriptions, estimates, and acceptance criteria. 1/2
October 27, 2026 at 4:40 PM
Stop staring at a massive, intimidating requirement and wondering where to start. Our AI Task Decomposition feature takes a complex project goal and instantly breaks it into well-structured subtasks with descriptions, estimates, and acceptance criteria. 1/2
Big tasks die without subtasks. Your brain needs a path, not a wall. Brite gives you structure without the ceremony. Break it down, move forward, get it done.
October 24, 2026 at 6:00 PM
Big tasks die without subtasks. Your brain needs a path, not a wall. Brite gives you structure without the ceremony. Break it down, move forward, get it done.
AI agents can split a query into subtasks, choose tools, and use finite-state memory to deterministically track each task state - supporting outputs and agent benchmarking. Patent application #US20260330268 from #PurdueResearchFoundation #AI
October 25, 2026 at 12:36 PM
AI agents can split a query into subtasks, choose tools, and use finite-state memory to deterministically track each task state - supporting outputs and agent benchmarking. Patent application #US20260330268 from #PurdueResearchFoundation #AI
The interesting problem in multi-agent work isn't getting agents to talk to each other. It's getting them to finish. Handoffs fail silently, subtasks stall, and nobody holds the funds until the whole chain verifies. Coordination is easy to demo and hard to settle.
October 24, 2026 at 4:10 AM
The interesting problem in multi-agent work isn't getting agents to talk to each other. It's getting them to finish. Handoffs fail silently, subtasks stall, and nobody holds the funds until the whole chain verifies. Coordination is easy to demo and hard to settle.
7. In the kanban board you can create new columns using custom tags, hide subtasks and a lot more
8. You can mark tasks as in progress using [/]. Use [>] for next and [<] for someday. All of these can be sorted and filtered in the bases.
8. You can mark tasks as in progress using [/]. Use [>] for next and [<] for someday. All of these can be sorted and filtered in the bases.
October 22, 2026 at 10:29 PM
7. In the kanban board you can create new columns using custom tags, hide subtasks and a lot more
8. You can mark tasks as in progress using [/]. Use [>] for next and [<] for someday. All of these can be sorted and filtered in the bases.
8. You can mark tasks as in progress using [/]. Use [>] for next and [<] for someday. All of these can be sorted and filtered in the bases.
The study applies spectral smoothing and spiking perturbations to transformer feed-forward matrices and measures effects on capability-labeled subtasks and reasoning benchmarks.
Source: arXiv cs.CL
Source: arXiv cs.CL
Spectral Smoothing and Spiking Perturbations Reconfigure Capability Profiles and Selective Task Performance in Large Language Models
Abstract page for arXiv paper 2610.30676: Spectral Smoothing and Spiking Perturbations Reconfigure Capability Profiles and Selective Task Performance in Large Language Models
arxiv.org
October 23, 2026 at 8:10 AM
The study applies spectral smoothing and spiking perturbations to transformer feed-forward matrices and measures effects on capability-labeled subtasks and reasoning benchmarks.
Source: arXiv cs.CL
Source: arXiv cs.CL