Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
manv1
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
18 ms
·
121.
▲
by
manv1
4y ago
You could argue that human brains are actually big state machines, at least for 90% of humanity.
122.
▲
by
manv1
4y ago
Many dogs will look at what you're pointing at and not your finger. It depends on the dog. And there are dogs that are literally bred to point...but they're usually pointing at game. There are dogs that will literally drag you to
123.
▲
by
manv1
4y ago
The 030 was the first real MMU 68k.
124.
▲
by
manv1
4y ago
If you're looking at layoffs as a way to reduce costs short-term they do work, even accounting for severance.
125.
▲
by
manv1
4y ago
I suppose you haven't been reading the normal press.
126.
▲
by
manv1
4y ago
Amazon has historically said that they don't care about their stock price. The losses in question are a mark-to-market loss on Rivian. If you look at their operational numbers they're mostly fine. That's the problem with the
127.
▲
by
manv1
4y ago
Just imagine how many people that this happens to who don't know enough to post online on a forum that lots of people read. For the CloudFlare people here, this is an upsell opportunity that's being missed. The whole point of the
128.
▲
by
manv1
4y ago
Analog hole.
129.
▲
by
manv1
4y ago
I mean, it'd be trivial to write stream replay for MQTT. It's literally just stashing messages and sending them back on connect. Not sure what the issue is there.
130.
▲
by
manv1
4y ago
Funny, the IoT space has bought into MQTT but the general internet space has not. MQTT scales and works. And it's easy, fast, and small. I've been trying to get our guys to do MQTT-based pub/sub, and they're rather do th
131.
▲
by
manv1
4y ago
Everyone says a memory-safe C would be slower, but has anyone actually tested that recently? It seems that a memory safe C would be faster, in that you wouldn't have to learn yet another language and runtime to deploy your stuff.
132.
▲
by
manv1
4y ago
You haven't met people who are, basically, dumb as posts. They're good people for the most part, but they are literally dumb as posts. As a less offensive example, there are people who literally cannot comprehend a=b, b=c, a=c. Th
133.
▲
by
manv1
4y ago
Many of the larger IT orgs I know do releases on Tuesday - because nobody wants to deal with crap on the weekends. Everyone is around, and the system is actually being used. This kind of deployment also brings more discipline to the process
134.
▲
by
manv1
4y ago
If Intel cuts its dividend the stock will crater. Since executive compensation is stock-based, tied to the stock performance, or both, there's an obvious incentive to keep the dividend going. The dividend has become more political than
135.
▲
by
manv1
4y ago
The cost should be billed to the department with the users that were affected. The laptops are assigned to those employees, so normally any kind of compliance/spend should be associated with them as well. It doesn't matter if thos
136.
▲
by
manv1
4y ago
The limits are because at this point there's no way to tell the tool what you're trying to do. The various photo modes are a step in that direction, but they've been stuck. Once they find a way to interact with the processing
137.
▲
by
manv1
4y ago
MS is defeated in that what they do doesn't matter in the industry anymore. They make plenty of money selling people more dog food.
138.
▲
by
manv1
4y ago
AWS + Node.js + Lambda + Aurora + API Gateway + ElasticCache + CloudFront + SQS + MediaLive + MediaConvert. We scaled up to hundreds of thousands of viewers + telemetry worldwide with no issues. Should be breaking into millions real soon no
139.
▲
by
manv1
4y ago
Oh, they brought in Waldemar Horwat and he worked on the Rosemary runtime. For those who don't know, he wrote the TMON low-level Mac debugger when he was a teenager and has gone on to do all kinds of stuff.
140.
▲
by
manv1
4y ago
I have Presto!Links, Presto!Mail, the Magic Bag Game Pack, and MagicXchange in boxes. I was the tech lead on MagicXchange/Mac for a while and worked at Active Paper on the Presto stuff, and was the lead on Presto!Links (which was subse
141.
▲
by
manv1
4y ago
Can any of the AIs come up with the Fosbury flop? It sounds vaguely like AlphaGo can. If so, we should celebrate, panic, or do both. https://en.wikipedia.org/wiki/Fosbury_flop Most of human "progress" has bee
142.
▲
by
manv1
4y ago
I still have my two Magic Links, a bunch of SRAM cards, and a bunch of shrink wrapped software. Lost my telebug cable years ago. GM is where Andy Rubin learned how to leverage carrier partnerships. He took that knowledge to Android and the
143.
▲
by
manv1
4y ago
The reason you do this is because you have an instruction that means nothing. You save space by recycling it. Why bother making a new instruction? That's against the whole idea of RISC. It's not their problem if the compiler write
144.
▲
by
manv1
4y ago
The story was that Atkinson saw the Parc and it's GUI and said to himself "holy shit they're doing their graphics this way" and write QuickDraw based on that. The Parc didn't do its graphics that way at all, and Atk
145.
▲
by
manv1
4y ago
What you need to do is codify this imbalance in ownership. If you're bringing more to the table you get more. It begs the question: what's the other person for? As that other person does more they might get more...or they might no
146.
▲
by
manv1
4y ago
"GCC, in turn, allowed the development of new languages like Perl and python" wut?
147.
▲
by
manv1
4y ago
You're inside the return window.
148.
▲
by
manv1
4y ago
Jerry Pournelle's column wouldn't have existed if he had owned an Apple Computer.
149.
▲
by
manv1
4y ago
It's better for you to sell it on the secondary market, the 16" MBP M1 should be worth about $1400-1600 USD in about a month after the M2 ships.
150.
▲
by
manv1
4y ago
The great news? Price drops on the M1 MBP!
More ›