Remco van Mook

4.1K posts

Remco van Mook

Remco van Mook

@rvmNL

Father, tech startups, interconnection guru, Internet numbers bigwig, collector of geeky T-shirts. Also @[email protected] and @rvmnl elsewhere.

Deventer, Netherlands Katılım Eylül 2011
547 Takip Edilen995 Takipçiler
Sabitlenmiş Tweet
Remco van Mook
Remco van Mook@rvmNL·
2023 will be for AI training data what 1950 was for carbon dating.
English
2
2
13
1.2K
Remco van Mook
Remco van Mook@rvmNL·
Imagine playing chess against a distracted and not particularly gifted 5 year old. On a monopoly board. And he's just found a rock in his pocket that is now on the board as a "super king turbo".
English
0
0
1
66
Firefox for Web Developers
Firefox for Web Developers@FirefoxWebDevs·
HTML might be getting a new type of tag, which… hasn't happened this millennium! It looks like this: <?marker name="…"> See the video in the next post for a quick intro to how it works.
English
23
39
690
100.3K
Remco van Mook
Remco van Mook@rvmNL·
@esrtweet Chaperoned development is how I'm trying to approach it. Gets pretty intense and the models aren't built for that kind of interaction really.
English
0
0
1
53
Eric S. Raymond
Eric S. Raymond@esrtweet·
We only have one term for the evolving mode of writing software cooperatively with an AI - "vibecoding". I think we need at least one more term, because there's a spectrum of techniques being explored and "vibecoding" can describe at most one end of it. At one end of the spectrum is what I think of as YOLO coding, and I'm tempted to compound-word it as "yolocoding". That's where you describe an application that you want and just tell the computer to grind on your informal description until it thinks it's done. I never do this. I hear stories about people doing this ("I wrote an entire videogame in 3 days!") and getting good results. I'm not sure I believe them. I inhabit the other end of the spectrum, where you ride the AI like a fine warhorse. You start by tossing an informal spec at it, then you and the computer iterate on both the spec and the code. It does the moving and supplies the grunt, you do the steering. I'm not sure what to call this. "Symbocoding", maybe. You and the AI are symbiotic, more than the sum of the parts. I think when people talk about "vibecoding" they usually mean something more towards the yolocoding end of the spectrum. I don't think that's where you'll get high-quality code from, and where the big productivity gains are. I could be wrong, but whether I'm right or wrong I think we at least need to be able to articulate the difference between fire-and-forget and a highly interactive, carefully directed approach. Yolocoding. Symbocoding. Are these the right words? I don't know, but I'm sure the distinction is important.
English
122
25
354
18.2K
Remco van Mook
Remco van Mook@rvmNL·
@zuhaitz_dev Fetching individual bits sucks so mostly doesn't happen. But if you take an 8 bit character and just look at the last bit of that character, 0,n and N map to False and 1,y and Y map to True 🙂
English
0
0
0
46
Zuhaitz
Zuhaitz@zuhaitz_dev·
What is better, 8 booleans in 8 bytes or 8 booleans in 8 bits. It is not as easy as just choosing one. The first one is simpler in terms of CPU instructions, as having only 1 byte for 8 booleans will require shifts and masks, but the ALU is not the bottleneck, memory is. In the first scenario it would need to fetch x8 bits of data from the memory. Which, for 8 booleans is not much, but for a large array (like a boolean flag for each pixel in an image) it will fill the L1, L2 and L3 caches a lot faster. A cache miss is a lot more painful than an extra AND/SHIFT. It also depends on how we access the data. For looping, bit-packing will be the best as there's more data packed, reducing cache misses again. If it's for random access, a byte array surely wins (if the data set is small enough to fit). I must also say that with bit-packing we can find race conditions as we cannot access bits atomically. It is important to consider everything. But at the end of the day, if you don't worry about performance, use a byte array.
Programmer Humor@PR0GRAMMERHUM0R

cleverNotSmart redd.it/1r2m4ui

English
32
12
513
61.4K
Remco van Mook
Remco van Mook@rvmNL·
@brankopetric00 The correct answer is, as usual, it depends. What's the amount of time that passed between the requests, what's the DNS reply TTL, how long is the keep alive timeout, did your host have any other traffic for the default gateway, ...
English
0
0
0
74
Branko
Branko@brankopetric00·
Senior DevOps Interview: What happens when you type google[.]com into your browser and hit Enter? Focus specifically on the TCP/IP stack. 1. ARP request (to find the gateway MAC). 2. DNS (UDP). 3. TCP 3-Way Handshake (SYN, SYN-ACK, ACK). 4. TLS Handshake. 5. HTTP GET. Which of these steps is skipped if you visit the site a second time (Keep-Alive)?
English
31
18
334
87.1K
Robert Graham
Robert Graham@robertgraham·
The correct value of pi 𝜋 is 3.141592653589793. It's the maximum precision for a 64-bit (double) floating point number according to the IEEE 754 standard, used by NASA and rocket engineers for all their calculations for rockets in the solar system. You'll never practically need greater precision, and you'll rarely do floating point at less than 64-bits, so there's little point to use fewer or more digits. Now, gamers might choose 3.1415927 for 32-bit floating point, but they understand the compromise. For everyone else, pi 𝜋 is exactly 3.141592653589793.
World of Engineering@engineers_feed

How many π digits do we need? 3.1415 ➡️ design the finest engines 3.1415926535 ➡️ obtain the circumference of the Earth within a fraction of an inch 3.1415926535897932384626433832795028842 ➡️ measure the radius of the universe to an accuracy equal to the size of a hydrogen atom

English
184
163
3.9K
596.8K
Niels Hoekman
Niels Hoekman@NJHoekman·
Als Epstein-bestanden voldoende zijn om zonder bevel servers te confisqueren, zijn we iets fundamenteels kwijt. De Nederlandse politie heeft zonder gerechtelijk bevel een server van VPN-aanbieder Windscribe in beslag genomen. Dat bevestigt het bedrijf zelf. Volgens de autoriteiten wordt de server teruggegeven nadat deze “volledig is geanalyseerd”. Windscribe noemt de actie uitzonderlijk en zorgwekkend, en terecht. Waarom dit zo problematisch is: - Inbeslagname van digitale infrastructuur hoort aan strikte wettelijke voorwaarden gebonden te zijn - VPN-providers bestaan bij de gratie van privacy en vertrouwen - Er was geen rechterlijke toets vooraf - Op de server bevonden zich naar verluidt ongelakte Epstein-bestanden, wat onvermijdelijk vragen oproept over motieven en proportionaliteit Dit gaat over rechtsstatelijkheid, digitale privacy en de vraag hoe ver autoriteiten mogen gaan “voor onderzoek”. Als we accepteren dat servers zonder bevel kunnen worden meegenomen, analyse eerst, juridische onderbouwing later. Dan zetten we een gevaarlijk precedent. Zeker in een tijd waarin encryptie, privacy en digitale autonomie al onder druk staan. Transparantie en rechtsbescherming zijn geen luxe. Ze zijn de fundering van vertrouwen in de overheid.
Niels Hoekman tweet media
Nederlands
116
392
811
37.8K
Remco van Mook
Remco van Mook@rvmNL·
@cryptoron None of that post makes sense. Unlike some other countries, we still have a mostly solid form of due process in place, and none of this is in line with standard operating procedures of the Dutch high tech crime unit. They're way smarter than this.
English
0
0
1
60
𝗥𝗼𝗻𝗮𝗹𝗱 𝗣𝗿𝗶𝗻𝘀 🇿🇦
The Dutch police are not known to seize equipment like this without proper legal authorization. Also, in digital forensics it is standard practice to keep a seized server powered on or otherwise technically accessible until investigators can perform a live memory (RAM) capture in a forensic lab
Windscribe@windscribecom

THIS IS NOT A DRILL: The Dutch authorities, without a warrant, just seized one of our VPN servers saying they'll give it back after they "fully analyze it". Windscribe uses RAM disk servers so the only thing the authorities will find is a stock Ubuntu install. The bigger worry is the unredacted Epstein files we had on there...

English
45
22
697
155.3K
Remco van Mook retweetledi
Corey Quinn
Corey Quinn@QuinnyPig·
@dvmikun Remember, everything's a database if you hold it wrong.
English
3
3
19
1.9K
solst/ICE of Astarte
solst/ICE of Astarte@IceSolst·
Help my arrow keys are gone and it has one-shotted me, idk how to use terminal or edit text anymore Do not get a mini keyboard
solst/ICE of Astarte tweet media
English
148
6
390
41K
Remco van Mook
Remco van Mook@rvmNL·
@cryptoron Herinnert me er aan dat ik ergens in de jaren 90 Venezuela offline heb gehaald door op IRC "+++ATH0' te sturen.
Nederlands
0
0
1
98
Remco van Mook
Remco van Mook@rvmNL·
@SwiftOnSecurity "Be liberal in what you accept, and conservative in what you send" - Postel's law on protocol design. Society could certainly use a bit more TCP/IP.
English
0
0
1
88
SwiftOnSecurity
SwiftOnSecurity@SwiftOnSecurity·
TCP/IP is a social construct
English
46
56
531
34K
Remco van Mook
Remco van Mook@rvmNL·
@withzombies Meanwhile some of the largest companies on the planet officially require people to take vacation days to attend work related conferences, and pay attendance and travel out of their own pockets. So yes, many of us are managers but the mindset issue is elsewhere.
English
0
0
0
29
𝗥𝗬𝗔𝗡 𝗦𝗧𝗢𝗥𝗧𝗭
I wish people would stop putting their conferences on weekends. We're a grown up industry that makes money. We need to start acknowledging that attending conferences is valuable work time.
English
3
1
29
2.5K
Ben Roberts 🇬🇧🇰🇪
Ben Roberts 🇬🇧🇰🇪@benrobertskenya·
Is this possible proof that Blue Tech Wave Media @MediaBtw and @NRSTeam_ are the same organisation? They are sending out near identical emails to African IP Address resource holders in Africa at the same time ?
Ben Roberts 🇬🇧🇰🇪 tweet mediaBen Roberts 🇬🇧🇰🇪 tweet media
English
1
1
7
673
Remco van Mook retweetledi
Zephyr
Zephyr@zephyr_z9·
>be OpenAI >“we’re AGI-maxxing bro trust the plan” >revenue? lol lmao >burn cash faster than a Saudi prince in a Monaco casino >walk into PIF like “salam, got a PowerPoint and a prayer” >Saudis: “how many GW you want?” >Sam: “yes” >sign MOU on a napkin written in camel milk ink >immediately leak $500B figure to WSJ >journalists: “is any of this real?” >OpenAI: “it’s real in the sim” >Oracle hands Larry Ellison another yacht, signs $300B cloud deal >delivery date: 2027, right after Half-Life 3 >NVIDIA shows up: “we’ll invest $100B if you actually build something” >Jensen whispers: “pls buy more H100s I have yacht 2 to fund” >sign LOI with Samsung for FLOATING data centers >yes, boats full of GPUs, what could go wrong >900K wafers of DRAM per month, half the planet’s supply >Samsung execs already scuba-diving for sunken sticks >AMD joins the circle-jerk: “we’ll give you 6 GW of MI450s” >fine print: OpenAI gets 10 % of AMD if they don’t melt first >entire cap table now denominated in watts and prayers >Stargate capacity at 7 GW, still can’t run ChatGPT without gaslighting you >every LOI signed with crayon, every MOU blessed by a shaman >rating agencies: “this is fine, everything is fine” >SEC: “is this securities fraud?” >OpenAI: “no, it’s sovereign cloud infrastructure, habibi” >final slide: “step 4: profit” >step 1-3 still TBD
Zephyr tweet media
English
70
183
1.8K
180.9K
Dr. Peering
Dr. Peering@DrPeering·
A fun update of the RIPE Atlas Terms & Conditions from @ripencc & @RIPE_Atlas soon. Turns out folks were using “inappropriate” names (and they don’t mean Proby McProbeface) for probes! What would your inappropriately #RIPEAtlas probe name be? Inappropriate answers only please.
Dr. Peering tweet media
English
1
0
0
415
Dilan Yesilgöz - Zegerius
Dilan Yesilgöz - Zegerius@DilanYesilgoz·
De VVD kiest voor rust in jouw portemonnee. In plaats van o.a. tanken en je hypotheek nóg duurder maken. ✅ Zuinige overheid ✅ Lagere accijns ✅ Hypotheekrenteaftrek behouden ✅ Lagere lasten op o.a. energie en inkomen ❌ Geen hogere lasten op spaargeld en erfenissen
Nederlands
479
43
203
51.5K
SwiftOnSecurity
SwiftOnSecurity@SwiftOnSecurity·
It occurs to me GenZ have never seen that lesbian go-kart movie we all watched as kids, 2 Girls 1 Cup
English
47
36
435
32K