ezekiel68

I tried it through siliconflow (I have no relationship to them other than as a normal customer) with the command line Pi agent and it's pretty good at adding features to a rust lang code repo. Compared to Deepseek V4 Pro there, M3 is about 3X the cost for non-cached input tokens and 5X the cost for output tokens. (both offering a 1M token context window). It certainly wasn't worh 5X the output of the Deepseek model for my uses, though.

I couldn't find any info on how they (siliconflow) quantized the model. I think some of the other ineference service aggregators might offer it as well, already.

I imagine the commenter was referring to the article title wrt satire.

Thanks for this!

Silence! Something something NPU at the edge hand-wave.

/s

For sure -- but these are the 'exceptions that prove the rule'. Someone else mentioned HTTP3 and QUIC. For all of these, there are thousands of projects (including multiplayer games and even enterprises) that didn't quite get it right and suffered for it.

Or, at least, so says the lore.

Well, not all killer drones run any kind of transformer model so...

Coming from the enterprise SQL world, I never took SQLite seriously for the very reason that field types were not enforced by default. (Yes, I was agog when it became the backbone for app metadata on smartphones.) Anyway, reading this reminds me of the old chestnut from networking about choosing UDP over TCP for its low-latency and simplicity and then eventually adding nearly all the reliability facilities of TCP to the app (automatic retry, etc) by hand.

The cancel lands on a process that has never heard of the query, and nothing happens.

Peering fixes this. The processes are aware of one another, so a cancel that lands on the wrong process is forwarded to the one that actually owns the session.

I understand "peering" as a concept here but have never tried this with PostgreSQL before. May I ask:

A) Does PostgreSQL have a mode/setting for peering that makes this easy? I'm imagining a mechanism that either goes round robin (re-sending the cancel to peers until it doesn't return an error of some kind) or some metadata in the cancel request that enables the wrong-destination process to somehow identify the proper process.

B) And by what mechanism? If all the PostgreSQL processes are listening to clients via so_reuseport, I guess there must be some other IPC method used for the peering chatter.

I'm sure all of this is "correct", but the fact is that many of us (myself included) enjoyed years of trouble-free service from our Commodore 1541 disk drives. And we could easily compare them with the Apple ][ computers and drives we learned on at high school

Thank you for helping remove my ignorance!

I had no trouble with it once I realized it wanted me to drag the border area.

You kind of covered this in your last point but it's worth emphasizing: we generally don't know for sure whether "correcting" a low serum vitamin D reading helps out in some way or merely silences a warning beacon that should have been traced to a more serious underlying condition.

Still, 170,000 - 1.2M years is a fairly short amount of time if we go back only to the common ancestor who begat progeny that would become Pan and Autralopithecus (around 12 million years ago). It could be that early hide wearers started a trend that, to this day, continues to interfere with natural vitamin D metabolism (while also providing many benefits).

If I'm not guilty until proven innocent, then neither is Motorola.

or

"How I Learned To Relax And Just Start New Sessions Often"

Lots of people throwing shade at Gemini CLI in the comments. I'm not saying it's perfect, but I enjoy using it. I haven't tried antigravity at all yet. I hope it will be an experience that is somewhat close to agentic coding on the CLI. I hit other model providers from Pi agent, but I'd like to be able to take advantage of my Google AI subscription on the CLI.

Yeah, that quote was definitely cringe for me. But I'm glad they made something that made them happy.

I don't know why reclaimthenet hasn't embraced the obvious answer: Simply create a new smart device operating system with a fully disentangled cosmos of programs, libraries, APIs, app SDKs, hardware partners, drivers, trust networks, carrier agreements, app stores, documentation, conferences...

Dont' get it twisted, anyone; plenty of companies have a reputation to maintain for this reason (but don't do this). This is an absurdly generous severance package.

Yes, but I wouldn't stop watching the originals. And I wouldn't stop watching anything until I saw what came out.

Ugh. I hate these "guilt by association" hit pieces. Nothing is wrong and yet we must signal our virtue.

Might as well just open our pants and wave our wangers, hoping for a better world

Lots of handwaving about fear of rust changing too often and even MORE downvoting of this as if there is no concern whatever. Neither of these extremes are valid.

For me it comes down to the old standby: just vendor the .cargo build chain into your repo and be done with it. There you go. Lock in the year edition, the version, the features, the quirks, and the bugs. Just like game devs did with game engines or OpenGL or Direct X versions.

TBF, it's certainly best practice, advised by the model providers themselves, to cut sessions short and start new ones.

Anthropic's "Best Practices" doc[0] for Claude Code states, "A clean session with a better prompt almost always outperforms a long session with accumulated corrections."

[0] https://code.claude.com/docs/en/best-practices

Qwen3-Coder produced much better rust code (that utilized rust's x86-64 vectorized extensions) a few months ago than Claude Opus or Google Gemini could. I was calling it from harnesses such as the Zed editor and trae CLI.

I was very impressed.

It is difficult to explain to a division director that they do not have sufficeint capacity (enough qualified programmers) to compete features within a set time budget. The old joke goes, "It takes one woman nine months to produce a baby. But: what if we just put nine women in a room for one month!?"

Ward Cunningham [edit: oops, it was Kent Beck] had it right, long ago, when he wrote in "Extreme Programming" [paraphrase]: You don't drive to Florida by carefully lining up your car in New York on I-95 South, locking the steering wheel, and then pressing the accelerator until you arrive.

This was really all that Agile was ever trying to avoid -- the tyranny of imaginedf omniscience. The bad old way (which I did labor under in the '90s) set up a Gant chart of dependent requirement up front, during a "design phase" which completely de-valued learnings and insights gained along the way as a software system was constructed during the "implementation phase". It was the best we had till then, but many software projects were failing due to their inability to adapt to unforeseen design flaws or to the feedback of stakeholders (once the software finally got into their hands).

I don't know why the ceremonies became ossified and sacred. I guess every movement must confront the danger of settling for form over substance. I do know one thing. You can't build an amazon dot com, a Facebook, or a Grand Theft Auto in a 1-million token context session with an LLM. I'm sure you can do it with many such sessions, but it won't be an LLM that ties it all together properly (again - too much context). And I say this as an enthusiastic user of agentic programming.

Now we just let the AI "move fast and break things".

I found a lot of value in these benchmarks and evangelized about them at my various employers over the years. Almost any enterprise is interested in lowering their cloud compute costs. Riddle me this: other than rotating out stale logs in cloud object storage or blocking malicious bandwidth drains from cloud CDN, what intervention lowers non-AI cloud costs more effectively than using a web service stack that requires dramatically fewer CPU and RAM resources while maintaining a high, error-free request rate?

A lot of handwaving about hAx in the benchmarks but many of these claims are from people who got their information secondhand (or worse). Actually reading code from the top submissions in the techempower/FrameworkBenchmarks repo (organized neatly under the frameworks/ directory hive) yielded valuable insigthts for me:

* Pipelining SQL requests has a massive effect on RPS for web services that will access SQL databases

* A well-maintained HTTP2/HTTP3 web server written in c named h2o is relevant in 2026, even if it is used as a proxy that delegates business logic to simpler web service workers written in Rails or in python 3 (via Gunicorn)

* For web services that write to a SQL database, the Axum rust stack, now with a healthy ecosystem of middleware modules, may provide up to twice the RPS as the Spring (Java) stack (externally discovered: at lower CPU and much lower RAM usage)

* Even frameworks written in JS (hyperexpress, just-js) or python (aiohttp) can vault into the realm of top-10 performers if they leverage OS-level asynchronous IO and SQL pipelining.

Good news, guys! The maintainer updated the repo fifteen months ago so that the copyright in the license file mentions "2025".

An anecdote from an OG (me):

The same phenomenon worsened during the DotCom Meltdown and the Great Financial Crisis. This accelerated desperation is a sign of the times; paying subscribers are likely cancelling due to current economic conditions.