pradn

Pretty cool you got to see this stuff in the real world!

Questions:

1) Was the problem able to be solved with one machine?

2) Was the problem reliably able to be solved within a time-bound, or was it "spiky"?

3) Was the solution all or nothing? Aka would you get a partial/suboptimal solution after a time-bound?

> Distributed systems are different because they fail often

The key here is not just the rate of failure, but the rate of failure in a system of multiple nodes.

And - "distributed systems problems" don't only arise with several servers connected by a network. Any set of nodes with relations between them - files on disk linked logically, buffers on different IO devices - these are also going to face similar problems.

> Google here uses SD 1.4, as the core of the diffusion model, which is a nice reminder that open models are useful to even giant cloud monopolies.

A mistake people make all the time is that massive companies will put all their resources toward every project. This paper was written by four co-authors. They probably got a good amount of resources, but they still had to share in the pool allocated to their research department.

Even Google only has one Gemini (in a few versions).

You're right that the roles are somewhat flexible: somewhere between raw technical output and technical/design leadership and "program/project management".

These ceremonies are meaningful if you invest meaning into them.

For my college graduation, the dean did something small to make the ceremony more meaningful. He asked us all to stand up, look back to our family, and applaud them for supporting us all these years. For me, looking back at my parents and thanking them for all they've done for me was a beautiful moment. And it was a full-circle moment for our family, the culmination of a long journey of immigrating to the US, moving around in search of stability. We had moved to be able to buy a house and to get us good, affordable educations. Both of those dreams were fulfilled at that time.

I don't recall much else from that ceremony. Not the speakers, but a few of the interactions afterward with my fellow students and their families.

Google changed their terminal level to L4, which is a level where you can be do designs with guidance, and implement big chunks of projects on your own. The previous terminal level was L5 ("senior"), where you're expected to own multi-quarter long projects and be a tech-lead of ~5-10 people. Oddly it does end up being structured as a pyramid - lots of L3/L4 folks than L5 folks.

I'm totally with this idea - use the language you know best for mundane tasks.

I used to switch to shell or Python to do one-off scripts. But there's not a super great reason why I can't do the same in C++, which is what I know best. All the build stuff is easy to do in my company's repo, so that's not a big blocker.

It's really nice that Go and Rust have one standard package management system, and a built-in standard formatter. Both of these things seem obvious now, but they were innovative when they were introduced. They add a new set of "batteries" to the "batteries-included" mindset. And they put community and ease-of-use first, which are crucial for adoption.

The AI boom is, in theory, a godsend for Intel and AMD. You can focus on creating good tensor computation hardware, without having to worry about getting gamers on board. No need for "Game ready drivers" or compatibility with complex legacy graphics APIs.

Of course, there's the elephant the room for general purpose tensor machines, which is CUDA - famously closely guarded Nvidia. But with the new wave of "above-CUDA" APIs like TensorFlow, PyTorch, and Keras, there's an opportunity to skip the CUDA layer altogether.

Yes, there's humans in the loop for the knowledge graph. I'm unsure if snippets are reviewed in the same way.

You are correct - if a workflow/agent company starts bundling data and API access - they'll multiplicatively increase the capability of their agents.

LLMs themselves are becoming a commodity, plus or minus prompt-following/format-following growing pains. In a year or two, we'll have pretty decent general LLMs that can make use of databases and tools/APIs.

It's a race to see who can integrate all these things in a good way. It really is an execution problem, not an idea problem - it's so obvious.

One sad part of the GenAI wave happening right now is that we're past the golden age of open APIs.

It's hard to read data with widespread anti-abuse checks (CAPTCHAs), lack of open-format data (RSS support being spotty), and restricted APIs (ex: Twitter API). Companies have all the incentives to prevent bot use, and select for human eyeballs.

If we had a Yahoo Pipes sort of golden age, GenAI agents would have a vaster playground to play in, and would be more useful for us.

Consider building an agent for choosing what to do on weekends for a group of friends. The agent would need to keep state for past activities (X, Y, and Z went upstate to Storm King last week) and users' preferences (ex: liking dosas or Calder, dietary restrictions). This part is easy enough - you could just keep a notebook that's passed as context. Older context gets simply deleted or condensed into high level points.

But would it be easy for the agent to:

1) Look up nearby restaurants and events? (Perhaps Resy/OpenTable allow listing restaurants, but it's likely they have tons of anti-abuse tech. Is there even a place where you could see a list of public events - Google pays a third-party for this feed.)

2) Actuate on behalf of the user? (Do Resy and OpenTable allow authority delegation so the agent could book restaurants for users? There's no standard way to do this across venue types - concerts, museums, cooking classes. Is it realistic for agents to click through these sites on their own?)

Friction is signal.

I recall I used to individually send snaps to like 30 friends each time, back before Snapchat invented Stories. I had to think a tiny bit, and put a tiny bit of effort, for each person. We lost that when people started posting to stories only.

If brands can be put on the balance sheet as "goodwill", how much money was burnt by obliterating this brand name - one almost as generic as Kleenex, for its product category (small streaming-first HDMI dongles)?

Yes, totally. Automated, fallible systems.

Ultimately it's a cost issue. If they wanted to, they could fact check the top million question queries - but they don't want to spend on that.

It's certainly wrong, but its not out of place w/ the sheer number of recipes and blogs saying the same thing. It's parroting "common wisdom" which is wrong.

It's not as egregious as some of the generative AI results spewed out by the lowest-cost-inference LLMs they hooked up to Google search a few months back.

The words "friend" (everyone knows how shallow a "Facebook friend" is), "share" ("ride-sharing" instead of calling a taxi), and "community" (is the entire customer base of Facebook really a community?) have been shorn of their sociable, human meanings. It's as if a corporation were mining the good will humans have accreted to those words over millenia.

Sometimes there are communities in these spaces - NUMTOTs or small Discord servers. Other times its just marketing foo foo.

That makes sense. The only reason to separate the metadata out is to avoid having to parse the main content files for front-matter. Plus the metadata files themselves could be in several formats, if necessary. I don't see a great need for that, however.

I was thinking of how to setup a simple static blog without the baggage of a full-blow framework. Of course, I'd be re-inventing the wheel, but at least it'll make sense to me, personally.

What I came up with:

* Each page gets its own folder.

* An optional METADATA file in each folder has stuff like the type of the page (blog post, photo gallery, code listing).

* The index file can be in Markdown or rST or something. I'd just branch to a different parser based on the file name suffix.

This is simple and lets me evolve the blog over time.

For a personal blog, an ideal mix might be just branching to a different HTML converter based on the

Quite helpful, thanks!

Eventual consistency gives you more availability, at the cost of consistency. It's fine for financial institutions to prefer this trade-off, because the any over/under error is a likely a fraction of the cost of unavailability.

Also, in the real world, we can solve problems at a different level - like legal or product.

The alternative is that everyone in the Microsoft security ecosystem gets off kernel-mode drivers. Once ebpf-for-windows lands, it should be possible for Microsoft, CrowdStrike, and everyone else to run their filters in user-land. That puts everyone on a level playing field, and makes the ecosystem more secure overall.

https://github.com/microsoft/ebpf-for-windows

It is a bit odd how they do it in Docs. In Cloud, announcements happen when the feature has already fully rolled out. Note, this is separate from region expansion, which might be delayed for some heavy-weight features (like new categories of VMs requiring special hardware.)

In Google Cloud at least, we're quite aware we're not the market leader, so there's a pervasive humbleness. We're proud of certain technical achievements (ie: Spanner), but the world can catch up quickly (ie: CockroachDB, Foundation, CosmosDB, etc). This might be a departure from the feeling in older divisions of the company, haha.

The NYC chapter of Papers We Love looks to be starting back up, after a gap of about a year and then two years before that. I think they're getting Datadog to host the events. They have a convenient office next to Penn Station.

https://www.meetup.com/papers-we-love/

There are a few examples of oral cultures ensuring perfect transmission across a large expanse of time.

One example is the several different recitation styles used to memorize Sanskrit verses. These methods give memorizers multiple ways to remember a line, and also prevent errors like the inadvertent mixing of adjacent words (" euphonic combination"). The "checksumming schemes" are far more elaborate than you'd imagine. [1] The result is perfect transmission of text and its pronunciation, including pitch accent.

Another example is a "multi-party verification" scheme in some Aboriginal Australian cultures. "... storytelling among contemporary Aboriginal people can involve the deliberate tracking of teaching responsibilities. For example, a man teaches the stories of his country to his children. His son has his knowledge of those stories judged by his sister’s children—for certain kin are explicitly tasked with ensuring that those stories are learned and recounted properly—and people take those responsibilities seriously. ... the 'owner-manager' relationship, requiring a story to be discussed explicitly across three generations of a patriline, constitutes a cross-generational mechanism which may be particularly successful at maximising precision in replication of a story across successive generations". [2]

[1]: https://en.wikipedia.org/wiki/Vedic_chant

[2]: "Aboriginal Memories of Inundation of the Australian Coast Dating from More than 7000 Years Ago", Patrick D. Nunn, https://doi.org/10.1080/00049182.2015.1077539

No, we’re quite aware the world outside has been catching up. There’s even a famous doc by a senior director about it…

The very very important stuff is hidden, and the only two examples anyone ever gives are core search ranking algorithms and the self-driving car.

Even the battle-tested hyper-optimized, debugged-over-15-years implementation of Paxos is accessible. Though I’m sure folks could point out other valuable files/directories.

OWNERS files rarely get in the way - you can always send a code change to an OWNER. They are also good for finding points of contact quickly, for files where the history is in the far past and changes haven't been made recently.

Readability really does help new engineers get up to speed on the style guide, and learn of common libraries they might not have known before. It can be annoying - hell, I'll have to get on the Go queue soon - but that's ok.

Yes, oops - fixed!