neerajsi

I was arguing with a friend about this point today. I'd posit that our economy has been growing in a way where goods that require more human time and natural resources to produce are giving way to goods that are intangible.

Once we've met our basic material needs, we're tending to consume things that are replicable with low marginal costs, and which do not interfere with the production of other goods. So maybe we can actually support a continually growing digital and entertainment economy, at least for a few more generations.

Maybe these mathematical contributions will also impact the efficiency and capabilities of our material production systems as well, which is another way to keep the economy growing.

I'm optimistic that we'll do more with our resources rather than trying to optimize for doing the same more efficiently with less resources.

I work on systems code that is tricky to write, but with a good test harness. Being able to leave the agent unattended to try several paths when I'm generating the first draft of some code is very helpful.

Of course beating the code into shape for submission requires more manual work. But the draft stage is valuable to find unexpected friction points.

I wonder if it's the same behavior as copilot cli. I dug into it when I got annoyed and found that there's a fundamental tension between rendering the cell based UI and supporting the terminal's native copy paste. At least on copilot cli there were extra zero width characters being copied.

I disagree. For example, the Wright brothers did air foil engineering empirically without full knowledge of the underlying fluid mechanics. Early and innovative engineering will often be less based on existing scientific or mathematical knowledge, while still using some of the methods. It's still absolutely defined as engineering.

I wonder how they ensure durability. Is it possible that a cell going down would roll back a payment after it has occurred. Or do they depend on a non cell database?

Another way around this is that things that aren't coherent enough to be described by a finite set of inertial propositions simply aren't perceivable.

I just joined a new team and have been using copilot with opus models.

We have our core code in a weird dialect of C and rust. C I know well, but not rust. Our tests are in Python. The pipeline descriptions are in Yaml.

Outside of the core code there are so many arcana to learn. Writing syntactically and semantically correct yaml/Python test code would be a nightmare. The Agents have flaws, but they provide a huge leg up in improving the tests.

And they are great at providing a first pass review of the core code before bothering a human reviewer. Lastly we run some of our test failures through AI triage, which often enough finds the root cause or rules out simple failures.

This shows up in a higher checkin rate. I'm curious to see whether this will lead to quality end product since we have more support for the more manually written and reviewed core product code.

How does release management work? Are there stable branches for deployed properties with a cheery pick approach for bug fixes? Are all products in google3 on a coordinated set of release trains?

I think intuition is what is developed through calibration, so I personally like the word calibration.

Intuition and other forms of knowledge are stock quantities while calibration and instructions are types of flows which change the stock. I'd love to know if there a better word for learning through trial and evaluation than calibration.

Aren't pipelines even easier to target and destroy than boats?

Speculating: local ssds aren't as valuable in the cloud since they're effectively ephemeral. If the instance restarts, it would lose its storage. Trying to keep a workload affinitized to an SSD or to migrate data to a different SSD when an instance moves increases cost prohibitively.

If you want recognize all the common patterns, the code can get very verbose. But it's all still just one analysis or transformation, so it would be artificial to split into multiple files. I haven't worked much in llvm, but I'd guess that the external interface to these packages is pretty reasonable and hides a large amount of the complexity that took 16kloc to implement

https://www.anthropic.com/news/analysis-tool

Seems like they already built this capability.

From the previous article in the series, it looks like the biggest impediment to just using full llvm to compile the query is that they didn't find a good way to cache the results across invocations.

Sql server hekaton punted this problem in a seemingly effective way by requiring the client to use stored procedures to get full native compilation. Not sure though if they recompile if the table statistics indicate a different query plan is needed.

Looks like the ampersands appear in menu items to indicate the keyboard shortcut key to navigate to that item.

https://willus.com/mingw/colinp/win32/resources/menu.html

Yes, they did have the bug with the lock prefix. IOT people at Microsoft got NT booting on the Quark and we ran into that problem. I wound up writing a small tool to patch out all the lock prefixes.

I've used this 'suggestion' workflow in azure devops. https://devblogs.microsoft.com/devops/introducing-the-new-pu...

mean, "not enough data" is an explanation why neural networks can't do a bunch of things... One of the things humans are really good at is learning things from a few examples

I dispute the search space problem for something like folding clothes. Like a lot of human actions in space, folding clothes and other motor tasks are hierarchical sequences of smaller tasks that are strung together, similar to a sentence or paragraph of text.

We can probably learn things from each other from few examples because we are leaning on a large library of subtasks that all have learned or which are innate, and the actual novel learning of sequencing and ordering is relatively small to get to the new reward.

I expect soon we'll get AIs that have part of their training be unsupervised rl in a physics simulation, if it's not being done already.

Reading this code for the first time, this seems to be a consequence of the separation between allocating and fd and "installing" a pointer to a file there. Allocating the fd already needs to acquire a lock. So if the install happens together with allocation, there wouldn't be a need to use synchronize_rcu to kick out other threads. The lock would do that.

In other words the LLM does not contain the knowledge of what the words represent.

This is probably true for some words and concepts but not others. I think we find that llms make inhuman mistakes only because they don't have the embodied senses and inductive biases that are at the root of human language formation.

If this hypothesis is correct, it suggests that we might be able to train a more complete machine intelligence by having them participate in a physics simulation as one part of the training. I.e have a multimodal ai play some kind of blockworld game. I bet if the ai is endowed with just sight and sound, it might be enough to capture many relevant relationships.

At least in the Windows group, we use ri and fi oppositely from how you describe. RI = sharing code with a broader group of people toward trunk. FI = absorbing code created by the larger group of people on the dev team. Eventually we do a set of release forks that are isolated after a final set of FIs, so really outside customers get code via FI and then cherry pick style development.

She can't make a prediction if the parents don't know.

The point is not that she predicted the gender of the child. She predicts what gender the parents expect the child to be without being explicitly told.

She didn't tell me anything about the success rate. Instead she told what factors she used to make her guess. She can tell if she was right or wrong right away because the parents can say what they are expecting.

I once met a gate agent in the Maui airport. She looked at me and my pregnant wife and congratulated us on our upcoming baby boy. I asked her how she knew. She said it was the way we carried ourselves and looked at each other. Not sure what specifically she saw, but she told us if another couple she observed who were having a girl. She said the man looked like the type who didn't treat women that well, but who was clearly trying to be on his best behavior.

Most of the times situations are a complete cypher, but in more unisual cases when observing groups of people who know each other, you can tell how they feel from analyzing the behaviors.

This author is clearly in a privileged position in the wedding group. She's in the background with the specific job of looking at everyone and capturing their feelings. They don't try to hide their behavior or respond to her, and the usual western taboo against staring at strangers doesn't apply.

This was definitely a good read!

Very interesting project!

I wonder if there's a way to make this set of techniques less brittle and more applicable to any language. I guess you're looking at a new backend or some enhancements to one of the parser generator tools.

I experienced the uk system briefly and it seemed decent for the simple thing I needed: wait for the NHS or pay a modest fee for private service. The price of private service is bounded by the fact that you can wait for the NHS.

Thanks for the explanation. Sometimes it's worth it to keep weird shit working, but it's hard to predict which subset of weird shit will happen in practice.

Wow, thanks for the links. I was previously unaware of this issue. I checked the referenced bug and I see that it is fixed and will be released in an upcoming version of Windows.

I've worked in and around srwlock for a long time. I don't think it's all that reasonable to think you should get "shared starve exclusive" by default, which is what you're asking for. We used to demand that kind of behavior in some windows filesystem code and it's always been a design mistake.

Seems like rfl people should do the work downstream like the rt folks did for a while. Maybe be an upstream for Android and the various corporate kernels.

Linus can decide whether to pull from their versions or not but basically maybe they should run a bit freer than having to deal with these maintainers.

A soft fork might clarify things like the old gcc fork did.

Yes, I think the pager attack is also an interesting case study. It's one thing to execute a supply chain compromise for information gathering, where the target may never know what happened. On the other hand, flaunting your abilities in that area will just lead you to being cut out of supply chains.