Where can I find the source for these? I'm currently over engineering my own book and you have done interesting ideas is love to reuse, without having to reimplement.

I've got a pixel 9, I don't see any meaningful changes here worth upgrading. The hardware looks essentially the same, except for a better camera and a ruby but if more battery. What am I missing?

I have a framework 13, but I couldn't imagine running a local llm on it, how do you do it? Do you have a eGPU?

If you have AMD, you can benefit from the unified memory.

Your Boris?

Could you publish your gitlab skill to give an example?

Missing from this is Nostr, which is very similar to Atoroto

So you have a write up on how to do this?

Not full text, but:

- config you cli to allow run agents in parallel

- ask to use tmp dir for all files, and git worktree copy

- ask to export all to files, progress, errors, reports, as it goes.

Now setup for each agent run (process with own context):

- planning-with-files, caveman, ask-questions-if-underspecified(if not in CI)

- these skills must be loaded into each subagent.

After, run assembling context: - audit-context-building, trailmark-structural

At this point all only operator runs. No other agents.

Now ask for forks(parallel), each loads context built before:

- dba-review

- supply-chain-risk-auditor (ask to trigger only on lock files changes

- spec-to-code-compliance

- openai-gh-fix-ci (if audit of PR)

- graph-evolution, differential-review

- dimensional-analysis

- mutation-testing, property-based-testing (ask to mutate and proptest around changes).

So not of these skills are load into shared context. They have own to avoid pollution and loss of focus.

So all forks run done. Ask for join.

- fp-check, second-opinion

Ask for final report.

All skills above from github and public (not my, easy to find).

I have private skills like:

- github-context-aggregation (links, comments, descriptions)

- structural-transformation-review (thing topology, category theory, duals)

- layered-context-chunking (2 layers at a time - like db+backend, backend+fe to more focus)

- product-feature-interactions (product line eng combinations)

Good mental model for what you are doing is

- https://github.com/microsoft/conductor/blob/main/examples/pa...

- https://github.com/features/actions

Engineers are conditioned to find problems, it's why we have the saying "don't let the perfect be the enemy of the good". It's paralysing. Often we raise problems outside of our lane, and that drives people nuts. This article is suggesting that we never raise issues, I've only ever seen the opposite, where we cannot keep our haranguing to ourselves and we fail to know when making a point makes sense or how to make our point and move on

Name the company, hiding it is irresponsible

Author doesn’t spell out why they are not naming them, but my guess is they are trying to not promote the product to malicious actors who would be interested in the sleep data of others.

I guess that’s not a huge problem, though, since all users are presumably at least anonymous.

less sleep data, i imagine, and more the whole “send remote electrical impulses” thing

It’s probably safe to assume they are all like that.

Why do you find it implausible that he got money to develop this challenge? I work in the non profit space and we get these kind of gifts all the time, with minimal strings attached or sometimes the deliverables don't see the light of day. 20k is actually small in my experience and they often come from random people who support our mission. Sure we do investigate and sometimes refuse based on findings, but we don't know or find everything. Taking money doesn't mean you know and support everything the person who have the money ever did, thought, said...

Just because you can't find it doesn't mean it didn't happen. Also just because you give it hard to imagine, does not mean that it didn't happen, or that people with better imaginations as you could accomplish that.

I was a generative linguist in my previous career and no linguist that I know has ever heard of this Chomsky challenge thing. It’s hard to prove that something doesn’t exist, but I’m fairly sure it doesn’t.

He could have created this, and then Epstein decided not to launch it? It can exist without you and others knowing about it.

It's conceivable, but I doubt it, for the reasons I mentioned. If Chomsky created this thing then presumably his wife can dig out some kind of record of it (as she apparently handled the communication with Epstein's office). So far we haven't seen it.