Bastiaan Veelo

It is easy to miss, also as a subscriber. I started to bookmark your analyses because they need sitting down and paying attention. X should let subscribers see a list of subscriber-only posts. Perhaps subscribers should maintain a list of links to your nuggets as a service to each other outside of X (X takes care of access control).

@FPCHD @intel I have to repeat: stop wasting your time on stupid accounts. 99% of what you see on X is noise and will not help you as an investor.

We continue to run circles around Wall Street research. Our Intel @intel estimates have been materially more accurate and significantly ahead of the Street. In January, we projected Intel would deliver EPS of $0.15 for Q1 2026. At the time, Wall Street consensus stood at just $0.04, and even the most bullish analyst was only at $0.07. Intel ultimately reported $0.29. While no one captured the full upside, our estimate was dramatically closer to reality than the Street. But the stronger proof is Q2. We projected Q2 2026 EPS of $0.18 in January, several months before results were reported. Yesterday, Intel management guided to approximately $0.20. Meanwhile, Wall Street consensus was only $0.10, and even the most bullish analyst was below our forecast at $0.17. The numbers speak for themselves: Q1 2026 vs Actual Reported Results Wall Street consensus error rate: 86% Our error rate: 48% Q2 2026 vs Management Guidance Wall Street consensus error rate: 50% Our error rate: 10% This is not an isolated case. It reflects what happens when rigorous quantitative research replaces slow moving, backward looking analyst models. We provide institutional grade equity research designed to be accessible to all investors. Our process is disciplined, data driven, and increasingly demonstrates a consistent ability to outperform Wall Street expectations. We also keep the receipts. Every forecast is timestamped, documented, and verifiable through our published research and company filings. If you want research that leads the market instead of following it, subscribe today. Excerpt 1: from our January research report. Link: x.com/alojoh/status/… Excerpt 2: Intel's SEC filiings.

«it will be approved automatically in all 27 EU countries.» I don’t think that is correct. If they vote yes, member states can approve without further evaluation. It remains an individual approval. If they vote no, member states can still approve with a full individual evaluation. I think if voted yes most member states will approve, but it is not automatic.

As expected the Netherlands will present their results on Tesla's FSD in the next meeting of the Technical Committee - Motor vehicles (TCMV) on 5 May. 🇪🇺 The vote will take place at the end of June. If at least 50% of EU countries vote in favour of FSD, it will be approved automatically in all 27 EU countries.

@ankkala @nonzenz In D, `debug` is a special keyword for conditional compilation. Code that is behind this condition is only present in debug builds, and is ignored in purity checks. ```d void fun() pure { debug writeln("log"); } ```

@nonzenz I would say that a print statement in general is an effect. For example you may be making a CLI program where the user sees every print, so they're only to be done from impure functions. But during development you may want to "debug log" from a pure function, I would allow that

more programming languages should come with the notion of a "pure function". you could mark functions as pure and they could only call other pure functions

@folaoftech Did you check COBOL?

The most lucrative programming language in 2026☕

@DonaldClark @Kasparov63 He doesn’t talk when he writes on the board. Presenters that use pointers typically talk while pointing. That is the point (but not his key message). What graphics would you use that are less primitive and more relevant to this talk?

@Kasparov63 His key message - don't turn your back on audience, which does consistently across entire lecture. Use graphics - he doesn't and when he does they are primitive and almost irrelevent.

This is deservedly viral and a great lecture, but I'm afraid it doesn't help much if you are dealing with religious or political bigotry that defies any logic!

Good! I was looking at $HYNSE:xlux but it is extremely low volume. When I bought my first $TSLA in 2019 I made the mistake to buy them at xswx, also much lower volume than xnas. Then the Swiss got into a quarrel with the EU by which I was refused to trade at xswx, so my shares were essentially held hostage until I was able to transfer them to xnas. Lesson learned.

@Brigadirk There will be a US listing soon

Deep dive into SK Hynix dropping soon...

@ErikR391987 @ReubenR80027912 So it was not an intended pun. It really is as bad as it looks.

Truly incredible stuff here. This is such an egregious error for the A section of the NYT (not a typo, genuinely not knowing what NATO stands for) that it should lead every piece on the decline in journalism

@MRMR__101 @deNiksnut @pasimatalamaki @frontlinekit Current. Speed over ground != speed through water.

@deNiksnut @pasimatalamaki @frontlinekit mhm.. it's also not accurate on yacht speed

📢 I don't know what to say... Our Interceptor drone is officially the fastest in Ukraine. All other 10 inch interceptors are running at around 160-200kmph MAXIMUM. Today in our first tests we hit 217. This is just the beginning 🔥🔥🔥

@unclebobmartin We are talking about @WalterBright ‘s Empire game, right? Which version of the source code did you use, in which language? en.wikipedia.org/wiki/Empire_(1…

Yesterday I went through a process of decreasing the processing time of the adversary algorithm in the Empire game. As the game proceeds the computational load increases dramatically. I told Codex to instrument the main loop and start gathering data once the computation time exceeded 500ms. Then, step by step it dug into find the bottlenecks. It proposed solutions and implemented them. This took several hours, but the result was a dramatic improvement. Codex found many places where computations were being needlessly repeated, and used a strategy of caching to streamline behavior. As one bottle neck was solved, it would find the next, and then the next, and so on. All I had to do was monitor the process and say "yes" from time to time. Eventually I told it that it has solved enough and had it back out all the instrumentation leaving the improvements in place and all the tests still passing as before.

@renderhjs @ClintRutkas Cert signing is a pain in many ways…

@ClintRutkas Winui3 was my first Windows GUI framework. It was very painful to onboard and learn the hard way to avoid UWP , AI SDK and more undercooked - often not compiling out of box. Also cert signing cost is a big turn off. Quality needs go up, less junk offerings

Dev using Windows? Love to chat with you

@ericjing_ai @Scobleizer «Zero human code.» And it shows. This is an utter waste of energy.

We let Genspark Claw build a complete multiplayer game from scratch. Zero human code. 100+ autonomous coding sprints. 17,500 lines. Isometric engine, economy system, 8 languages, email auth, 200+ AI-generated pixel art objects. It's live. It's open source. Play it yourself. 🦞 genclawverse.ai 📦 github.com/genspark-ai/cl…

@BarrakAli @farzyness And Zeiss.

@farzyness ASML quietly holding the keys to the entire AI revolution is the most underrated geopolitical fact of the decade.

Seems like AI's entire supply chain bottleneck is in Europe.

@lister6520 @DJSnM @igor_os777 True.

@BVeelo @DJSnM @igor_os777 Well it is in the sense that its a debugging nightmare. I've been bitten such as releasing memory by the fake pointer instead of the real one. In the end I would find it not worth the trouble and just put a -1 in the index or just adapt the task to use 0 index.

FORTRAN, ancient giant among scientific languages, has arrays starting at 1 by default—unless explicitly told otherwise. Want to start indexing from -42? Totally allowed. Mixing arrays with arbitrary indexing schemes led to some impressively baffling code, explaining why FORTRAN still frightens new devs.

@lister6520 @DJSnM @igor_os777 It is not much more of a cheat than arrays in C are a cheat.

@BVeelo @DJSnM @igor_os777 Yes I know, I've used such form but it's a cheat of sorts. I was thinking more about handling it natively like Fortran or even the OPTION BASE statement in Basic.

@DJSnM @igor_os777 Oh, meant to reply to @lister6520.

@DJSnM @igor_os777 char dum[4]; char* arr = dum - 1; for (int i = 1; i <= 4; i++) arr[i] = ‘a’ + i; // arr: [‘b’, ‘c’, ‘d’, ‘e’]

@christesla @SawyerMerritt @wholemars @Tesla They clarified that a decision is yet to be made. Apparently they felt the need to soften the certainty that could be picked up from Tesla’s own messaging.

@SawyerMerritt @wholemars @Tesla How did they manage to write so many words but say nothing.

NEWS: Dutch regulators (RDW), the key authority that would clear the path for Tesla FSD approval across much of Europe, has issued a response to @Tesla’s 𝕏 post today. "We know that Tesla's application has the interest of many people. The RDW gets a lot of (media) questions about this. Given this great interest and the many speculations, we would like to give a short response to Tesla's message and request. Normally, we never make statements about requests from manufacturers because of the market-sensitive information. Assessment process: In the message, Tesla states that they are in the final phase of the assessment process. That's right, Tesla and the RDW are currently going through the final stages of the assessment process. About 18 months ago, the joint intensive testing program began. During the term and in this final phase, the RDW thoroughly looks at the test results and analyzes the data. During this final phase, our inspectors will review all data and test results and, after completion of this process, a decision will be made on the approval of the driver's assistance system FSD Supervised. For the RDW, (traffic) safety is paramount."

@TeslaBoomerMama Is there a strategy? The lack of strategy seems to be the problem.

Your view on Nato allies not participating in the Strait of Hormus strategy?

@mechanical_4u “connected by a series of reduction gears” if that were true, all those 9s would not turn into 0s all at the same rate.

Mechanical odometers in older vehicles rely on a simple but precise system of numbered drums connected by a series of reduction gears. As the vehicle moves, a cable driven by the transmission rotates the internal mechanism, advancing each digit in a controlled sequence. Most classic designs were limited to six digits, meaning the highest readable mileage was 999,999. Once the final drum completes its full rotation, the gearing forces every number wheel to return to zero and begin counting again from 000,000.

@zakarum4 @corsix Correction: for strings we do have annotations to select one of two different formats.

@zakarum4 @corsix If you write those functions by hand, it is not a generic (de)serializer. Our (de)serializer does file I/O in legacy binary format without any need to annotate or extend the existing data structures. It is all generated at compile time.

Speaking as someone who studied joint mathematics and computer science, this right here is how you tell apart the mathematicians from the computer scientists.