Science on alper.nl

Sat, 13 Jun 2026 21:52:25 +0200

The value of real skills is going up day by day for those of us who have them.

Learning things is hard but that’s the whole reason people go to universities. Nobody needs a university education to ask Claude to do their work for them.

https://www.dailycal.org/news/campus/academics/failing-grades-soar-as-professors-see-greater-ai-usage-dwindling-math-skills-in-uc-berkeley/article_16fad0bf-02cb-4b8c-8d88-888ffd9f8608.html

Wed, 21 Jan 2026 22:49:40 +0100

Psychedelics being responsible for humans developing quicker and smarter than other species is a delightful notion. The proof isn’t very solid, but that’s fine for a “notion”.

https://davidluke543044.substack.com/p/reawakening-the-stoned-ape

Thu, 05 Jun 2025 22:22:58 +0000

For the many people who may not be intimately familiar with the workings of Turing machines, this visual and interactive guide should get them up to speed really nicely.

It’s mostly just the mechanics of the machine and doesn’t go into the more fun and formal parts such as reductions and proofs. But it’s a start.

https://samwho.dev/turing-machines

Tue, 27 May 2025 06:11:11 +0000

A twofer on software estimation that also tallies with my intuition. People are really great at estimating the median which is impossible to communicate to people without deep intuition for software engineering and statistics.

It makes total sense as a guideline to multiply your initial guess x1.6 for the mean, x5 for the 95th percentile and x10 for the 99th percentile. This is also why deadlines are irrelevant and counterproductive in software engineering. There are much better ways to achieve the same goals.

https://thesearesystems.substack.com/p/task-estimation-conquering-hofstadters

https://erikbern.com/2019/04/15/why-software-projects-take-longer-than-you-think-a-statistical-model.html

Tue, 27 May 2025 05:44:41 +0000

The Return To Office debate has been very contentious and hardly productive. Here in HBR is an ethics based approach to create solutions here that work for everybody. A lot of it is common sense but this is coherent and a useful lens.

https://hbr.org/2025/05/the-workplace-psychological-contract-is-broken-heres-how-to-fix-it

Tue, 27 May 2025 05:40:59 +0000

We use reservoir sampling a lot day to day and here’s a great visual explanation of how this algorithm does this seemingly impossible task.

https://samwho.dev/reservoir-sampling

Wed, 23 Apr 2025 23:58:57 +0000

A dive into Stafford Beer’s systems thinking that uncovers the ANT-style depth that’s often glossed over. That notwithstanding, I still think the glib and reductive POSIWID has its uses.

https://joelitobarski.com/posiwid/

Tue, 25 Feb 2025 22:30:25 +0000

Lots of people are going to have a problem with the truth going forward, so get those Wikipedia backups handy and start thinking about how to host it in a more distributed and resilient way.

https://www.404media.co/wikipedia-prepares-for-increase-in-threats-to-us-editors-from-musk-and-his-allies

Tue, 18 Feb 2025 17:11:08 +0000

Quanta’s write-ups of math discoveries are always enjoyable to read. In this case an undergraduate (!) figured out something new about hash tables and how fast they can be. Not sure yet when this will be shipped in code, but still it’s a pretty amazing discovery.

https://www.quantamagazine.org/undergraduate-upends-a-40-year-old-data-science-conjecture-20250210/

Sat, 08 Feb 2025 20:49:25 +0000

LLMs are fundamentally matching the patterns they’ve seen, and their abilities are constrained by mathematical boundaries. Embedding tricks and chain-of-thought prompting simply extends their ability to do more sophisticated pattern matching. The mathematical results imply that you can always find compositional tasks whose complexity lies beyond a given system’s abilities.

LLMs are still very useful in a bunch of domains but here’s an article explaining (from a paper with a novel bound on computational complexity) why improvements in reasoning seem to have run out of steam.

https://www.quantamagazine.org/chatbot-software-begins-to-face-fundamental-limitations-20250131/

Wed, 05 Feb 2025 22:45:13 +0000

The scientists are all on Bluesky which is understandable given the war on science and truth taking place on Twitter.

https://www.nature.com/articles/d41586-025-00177-1?linkId=12635521

Tue, 14 Jan 2025 08:44:58 +0000

The Agents chapter from Chip Huyen’s book “AI Engineering” is clear and enjoyable to read. She’s right that “the concept of an agent is fairly simple” but building something functional still looks like a massive lift.

https://huyenchip.com/2025/01/07/agents.html

Wed, 08 Jan 2025 08:55:50 +0000

Let’s see how this develops but good to see a positive healthcare story for AIs.

https://www.theguardian.com/society/2025/jan/07/more-breast-cancer-cases-found-when-ai-used-in-screenings-study-finds

Wed, 08 Jan 2025 08:53:07 +0000

A very useful thought experiment whenever anybody tries to pretend LLMs are ‘human’ because they sound human.

https://bsky.app/profile/colin-fraser.net/post/3ldoyuozxwk2x

Mon, 06 Jan 2025 06:14:08 +0000

RARE: Retrieval-Augmented Reasoning Enhancement for Large Language Models

Medical QA depends heavily on domain-specific knowledge that is not always available within pre-trained models, necessitating knowledge-based retrieval from external sources.

In addition medical knowledge evolves rapidly, and new treatments or updated guidelines may not be included in the model’s pertained corpus.

The question example for the reasoning process in Figure 1 is on a multiple-choice question. That seems overly simple.

In parallel, Commonsense Question Answering shares similar complexities with Medical QA, particularly in its reliance on structured multi-step reasoning and iterative evidence retrieval.

rStar (Mutual Reasoning Makes Smaller LLMs Stronger Problem-Solvers)

The rStar approach seems worth diving into. That will be the paper I read next.

Monte Carlo Tree Search

enabling the open source LLMs (LLAMA3.1) to achieve competitive performance with top closed-source LLMs like GPT-4 and GPT-4o.

We’ll come to this later in the paper. Their conclusion is that they can trick out LLAMA to get similar performance to GPT-4 in these domains.

Upper Confidence Bound applied on trees (UCT)

In contrast, rStar incorporates five distinct actions that enable more adaptive exploration:

A1: Propose a One-Step Thought. This action generates the next reasoning step based on previous steps, allowing the LLM to build the solution incrementally.

A2: Propose Remaining Thought Steps. This action enables the LLM to produce all remaining reasoning steps in one inference, similar to CoT, for simpler questions.

A3: Generate Next Sub-question and Answer. This action decomposes the main problem into a sequence of sub-questions, each solved in turn. A4: Re-answer Sub-question. This action allows the LLM to re-answer a previously generated sub- question, increasing accuracy by using few-shot prompting.

A5: Rephrase Question/Sub-question. This action rephrases the question to clarify conditions and reduce misunderstandings, enhancing the LLM’s interpretation of the problem.

I need to trace the rStar algorithm after reading the original paper. The explanation here is too short.

These queries target information that can either support or refute the content of each statement, ensuring comprehensive factual verification.

How does this approach deal with (non-)negation that LLMs often have a lot of trouble with? From a language perspective it could just as easily say I can or can’t eat grapefruit (iykyk) based on the temperature that day but especially in a medical context these kind of errors can be catastrophic.

RARE achieves substantial gains, outperforming rStar by 5.17% on MedQA, 2.19% on MedMCQA and 2.39% on MMLU-Medical.

Even if these numbers are statistically significant (which they don’t say), these increases are really modest. I would not call this in any way “substantial”.

Looking at Table 1, RARE is as much of an increase over rStar as rStar is over the next best approach so from that perspective maybe you could call it significant. The difference between worst and best framework here is around 10% across CoT, RAG, SC, rStar, RARE.

evaluated on StrategyQA (SQA), CommonsenseQA (CQA), Social IQA (SIQA) and Physical IQA (PIQA)

The main question I have is from what percentage accuracy such a system would be reasonably possible to use in a real world context. Even at 90-95% that would seem like it would be too low to rely on when the stakes are high.

By enhancing LLMs with retrieval-augmented reasoning, RARE bridges the gap between open source models and state-of-the-art proprietary systems.

The framework has only been tested on open source models like LLaMA 3.1 and not on larger proprietary models such as GPT-4. This is due to the high number of API calls required by RARE’s iterative retrieval and reasoning process, making evaluation on closed source models prohibitively costly.

So here they repeat the statement that they’ve bridged the gap but they say they haven’t used this approach with a model like GPT-4 because the number of API calls would make it too expensive.

That leaves on the table that these kind of many-call approaches are open to OpenAI because they can do these numbers of calls much more affordably from inside the house. No real gap has been closed here and it shows again how big of an advantage OpenAI has.

That raises the question: What makes GPT-4 so good? Why does it perform so much better than open source models?

RARE is designed to identify a single reasoning trajectory that leads to a correct answer but does not necessarily optimise for the best or shortest path that maximises robustness.

Any integration into medical workflows must be supervised by qualified practitioners to ensure patient safety and ethical use.

Thu, 02 Jan 2025 07:17:08 +0000

Adapting While Learning: Grounding LLMs for Scientific Problems with Intelligent Tool Usage Adaptation

A paper where they fine tune an LLM to be able to answer some questions itself and figure out for which questions it needs to use a specialized tool. Intelligent tool usage seems like it would expand the use cases for LLM driven systems much more than any kind of scaling (real or imagined).

However, scholars note that their abilities are capped at approximately high-school levels

That seems like a noteworthy statement especially if you are looking to LLMs to provide “novel thinking”. It would seem much more that high school problems are abundantly available and relatively trivial so they see a specific focus.

For numerical answers in the MATH and SciBench datasets, we consider answers correct if they fall within a small tolerance range of the true value, specifically within ±5%.

Don’t really see why you could not get exact answers in a mathematical domain.

This performance gap on public benchmarks is likely due to the larger parameter count and specific optimization of state-of-the-art models on the open-source datasets.

Same as with the high school questions. These datasets are easily available and draw attention so the models overfit on them.

The model Ours-Pn demonstrates performance comparable to Base-Pf , both showing a significant improvement over the base model. This similarity indicates successful internalization of distilled knowledge from tools. The transition from Ours-Pn to Ours-Pi showcases further improvement in answer accuracy, resulting from the model’s enhanced ability to intelligently switch to tools for harder questions.

This is the core proposition of the paper. Looking at Table 1 with the accuracy percentages there is something of an improvement but it does not really look dramatic or so convincing that you could use these systems in any critical context.

We’re looking at increases of 10-20% and an accuracy that’s still well under 90% (which I’m also not convinced would be usable).

We introduced a novel two-component fine-tuning approach to enhance Large Language Models (LLMs) in solving scientific problems of varying complexity.

One of the key issues with the paper I have is how much work the term “scientific problems” is doing. If this is published, people are going to think that the LLM is solving actual novel issues where in this case it’s just filling in relatively basic question/answer pairs that are well understood. Calling them problems is problematic.

The most interesting part of the paper is the appendix where you can see the actual questions and answers in the various datasets and the prompts they used (with example responses). The answers mostly are multiple choice which already influences how many of them you should expect to be correct.

Tue, 31 Dec 2024 14:23:39 +0000

Monolith: Real Time Recommendation System With Collisionless Embedding Table

I didn’t get that much from this paper, probably because it’s pretty high level and I don’t have a strong background in recommendation systems.

The approach is their Cuckoo Hashmap for embedding from which they can update parameters on the fly using existing data engineering pipeline technology.

Instead of reading mini-batch examples from the storage, a training worker consumes realtime data on-the-fly and updates the training PS. The training PS periodically syn- chronizes its parameters to the serving PS, which will take effect on the user side immediately. This enables our model to interactively adapt itself according to a user’s feedback in realtime.

Tue, 31 Dec 2024 10:17:34 +0000

Eight Things to Know about Large Language Models

A bunch of stuff that maybe was somewhat surprising a year ago but by now should be common knowledge for anybody even half following the developments in this field.

Some interesting bits in there but for the rest it’s a bit rah-rah because the author works at Anthropic.

In particular, models can misinterpret ambiguous prompts or incentives in unreason- able ways, including in situations that appear unambiguous to humans, leading them to behave unexpectedly.

Our techniques for controlling systems are weak and are likely to break down further when applied to highly capable models. Given all this, it is reasonable to expect a substantial increase and a substantial qualitative change in the range of misuse risks and model misbehaviors that emerge from the development and deployment of LLMs.

The recent trend toward limiting access to LLMs and treating the details of LLM training as proprietary information is also an obstacle to scientific study.

Witches Kitchen

Fri, 27 Dec 2024 20:22:46 +0000

From the Grothendieck biography, funny to see that the legend would express himself in German.

Riemann-Roch’scher Satz: der letzte Schrei: der Diagramm […] ist kommutatif!
Um dieser Aussage über f:X->Y einen approximativen Sinn zu geben, musste ich nahezu zwei Stunden lang die Geduld der Zuhörer missbrauchen. Schwartz auf weiss (in Springer Lecture Notes) nimmt’s wohl an die 400,500 Seiten.
Ein packendes Beispiel dafür, wie unser Wissens und Entdeckungsdrang sich immer mehr in einem lebensentrückten logischen Delirium auslebt, während das Lebens selbst auf Tausendfache Art zum Teufel geht – und mit endgültiger Vernichtung bedroht ist. Höchste Zeit unsern Kurs zu ändern!

—Alexander Grothendieck

Thu, 26 Dec 2024 23:31:13 +0000

Attention Is All You Need

I thought I’d dive back into history and read the original paper that started it all. It’s somewhat technical about encode/decoder layouts and matrix multiplications. None of the components are super exciting for somebody who’s been looking at neural networks for the past decade.

What’s exciting is that such a simplification generates results that are that much better and how they came up with it. Unfortunately, they don’t write how they found this out.

The paper itself is a bit too abstract so I’m going to look for some of those YouTube videos that explain what is actually going on here and why it’s such a big deal. I’ll update this later.

Thu, 26 Dec 2024 23:12:15 +0000

Tree of Thoughts: Deliberate Problem Solving with Large Language Models

I came across this paper after the recent o3 high score on the ARC-AGI-PUB test. It’s a quick read and details how to scale LLMs at inference stage by generating new states at every node and so create a tree on which to perform DFS/BFS search algorithms.

A specific instantiation of ToT involves answering four questions: 1. How to decompose the intermediate process into thought steps; 2. How to generate potential thoughts from each state; 3. How to heuristically evaluate states; 4. What search algorithm to use.

For each of these steps they can deploy the LLM to generate the desired results which then scaled over the search space balloons the number of calls that need to be done (costing almost 200x the compute).

This isn’t your normal LLM stochastic parrot anymore. We’ve gone one up the abstraction chain and here we have a computer science algorithm running with LLM calls as its basic atoms.

Sun, 03 Nov 2024 22:14:36 +0000

I don’t really know where to find a Honeycrisp apple locally, but since moving to Germany I have been eating tons of amazing juicy, sweet, crisp apples. It turns out that this is deliberate and these are not the apples of my childhood.

https://www.scientificamerican.com/article/apples-have-never-tasted-so-delicious-heres-why

Literature review of measures against the common cold

Mon, 08 Jan 2024 15:57:36 +0000

I refer to this piece of writing regularly:

https://srconstantin.wordpress.com/2017/12/11/preventing-respiratory-tract-infections/

It’s a review of the scientific literature to see what actually works against respiratory tract infections to either prevent them or to get better more quickly.

Sun, 05 Mar 2023 13:51:46 +0000

The hype train around AI now should be very familiar if you paid any attention during the previous crypto cycle. It’s the same people pushing this technology in the world in the same overblown and damaging way.

We’ve learned to make “machines that can mindlessly generate text,”Bender told me when we met this winter. “But we haven’t learned how to stop imagining the mind behind it.”

https://nymag.com/intelligencer/article/ai-artificial-intelligence-chatbots-emily-m-bender.html

Fri, 06 Jan 2023 19:45:18 +0000

Everything we continue to learn about the very earliest people is completely fascinating.

Next time I go to Turkey, a visit to the Taş Tepeler is definitely in order.

https://www.spectator.co.uk/article/does-an-unknown-extraordinarily-ancient-civilisation-lie-buried-under-eastern-turkey-/

Fri, 06 Jan 2023 18:55:08 +0000

Machine learning looks like a very challenging field to keep up with let alone to study it. This guide by Vicki Boykis gives a bunch of pointers around reading, mapping and learning that can be applied to any large field you want to get into.

https://vickiboykis.com/2022/11/10/how-i-learn-machine-learning/

Tue, 27 Dec 2022 14:02:17 +0000

Even a cursory dip into the subreddits and other social media posts, whitepapers, and other documentation and community discussions surrounding generative AI, makes it hard to miss the spite and negativity directed at human creativity. They are almost necessary preconditions to be part of the development community. Ironically, of course, the developers celebrate their own creativity in building these software tools, despite the fact that they implicitly rebuke the work and lives of everyone who creates or thinks meaning is confined to living things.

That perspective which I’ve also noticed in the people pushing this stuff is incredibly damning. They hover up the sum creative works of humanity and privatize it into shitty models.

https://davidgolumbia.medium.com/chatgpt-should-not-exist-aab0867abace

Sun, 21 Aug 2022 23:03:55 +0000

Any kind of workshop that you do virtually is going to be more difficult but in my experience a skilled facilitator and a motivated group can overcome these challenges.

https://www.gsb.stanford.edu/insights/thinking-inside-box-why-virtual-meetings-generate-fewer-ideas

Sun, 17 Jul 2022 09:02:34 +0000

https://www.currentaffairs.org/2022/07/the-dangerous-populist-science-of-yuval-noah-harari

For how popular and questionable Yuval Noah Harari’s writing has been, it has barely seen any scrutiny. This critique by Darshana Narayanan is on point and most of the things in it were—I hate to say it—glaringly obvious from the get-go.

Sat, 09 Apr 2022 23:21:57 +0000

Lovely bit about the broodje kroket and about Dutch colonial history.

https://theswordandthesandwich.substack.com/p/notable-sandwiches-22-broodje-kroket?s=w

Sat, 09 Apr 2022 22:33:22 +0000

On old but interesting article about a study that indicates that people will use their cars a lot more once we get fully autonomous vehicles.

https://jalopnik.com/zombie-miles-and-napa-weekends-how-a-week-with-chauffe-1839648416

Property dealings in Berlin

Mon, 07 Mar 2022 12:01:57 +0000

https://parisinstitute.org/the-real-academy-in-exile/#ref1

Back when I read this exposé in November it struck me as gripping but also as something that could become messy. Turns out that’s true since the piece now has the mark “Censored” and some choice parts of text have been blanked out. I did a quick search but can’t find any updates what might have happened. If I find time later, I’ll look in the Wayback Machine to see what pieces of text disappeared.

The kind of property deals that are described in the article are par for the course in any up and coming city. Recently there was a piece about resistance to the city giving away one of its largest buildings to a private art gallery: https://www.freitag.de/autoren/der-freitag/portraet-wut-motiviert-mich

The academic backroom dealings should also be considered to be normal especially in small subfields such as Turkey studies. Academics are playing funding games more than they research and the web of foundations that float around are ways to creatively bookkeep and move money around. None of this is really surprising.

Fri, 31 Dec 2021 12:07:22 +0000

I had no idea that most of Tesla’s FSD videos were fake. Huh.

https://blog.piekniewski.info/2021/12/18/farcical-self-delusion/

German adjective inflections demystified

Thu, 25 Mar 2021 10:08:16 +0000

I see lots of people still struggle with the German adjectives and especially the fact that there are three tables, one for each of definite articles, indefinite articles and when there’s no article, for 48 separate inflections.

Image

It looks pretty intimidating and back in school my brain refused to do learn this. I crammed it for a test, immediately forgot it after and German classes turned into agony.

There are various mnemonics going around to make the table easier but I’m not in favor of rote memorization if there’s also a logic behind something. As with all language, the entire thing is based on the principle of energy conversation while maintaining essential information and harmony. That’s my non-linguist intuition. Linguists feel free to chime in.

The entire table can be boiled down to five rules.

Definite nom./acc. singular: all -e (because a definite article has all the information and is leading, so no further decoration or information required on the adjective) except for masc. acc. which is -en (because acc always has -en and it sounds weird without it)
Indefinite nom./acc. singular: all congruent with the article: einer gut er Mann, für ein en gut en Mann, ein e gut e Frau, für ein e gut e Frau etc. (because the article still has all the information but it is weaker/indefinite so the adjective helps out)
Definite and indefinite gen./dat. singular: all -en (all the information is in the article so no need to repeat it but good to have some differentiation from the nom./acc. cases)
Definite and indefinite plural: all -en (because it’s plural)
No article: the adjectives take the endings that would normally have been carried by the article (because there is no article and otherwise the case information would not be there) except for masc./neut. gen. where the -s is on the noun

That’s it. I hope this adds some logic to something that otherwise feels totally random for German learners.

With practice you can look these inflections up in your head and with even more practice, anytime somebody uses the wrong inflection, it starts to sound wrong (that lack of harmony), just like it would to a native speaker.

Mon, 25 Jan 2021 21:22:48 +0000

When I still did photography, I read everything I could find about how cameras work and this is the mental model I built up back then.

The lovely thing about this webpage is that it makes each of the mechanics interactive so you can play with them and see what happens.

https://ciechanow.ski/cameras-and-lenses/

Sat, 05 Sep 2020 13:57:00 +0000

An uncannily thorough treatment of why DARPA works and how you might go about replicating it. Maybe people at the European policy level should read this instead of trying all these initiatives where everything good dies in a committee.

https://benjaminreinhardt.com/wddw

Sun, 21 Jun 2020 23:14:52 +0000

I’ve always felt that programming is more similar to languages than to mathematics. The number of hard maths problems that we solve as part of this work are really limited, the number of language problems on the other hand is immense.

https://www.nature.com/articles/s41598-020-60661-8

Wed, 27 Nov 2019 09:54:17 +0000

Turns out large parts of science and popular science are all contaminated by Epstein. Evgeny Morozov reveals his part and extricates himself.

https://newrepublic.com/article/154826/jeffrey-epsteins-intellectual-enabler

Sun, 20 Oct 2019 14:47:58 +0000

“the resilience of a system corresponds to its adaptive capacity tuned to the future.”

https://lorinhochstein.wordpress.com/2019/08/20/tuning-to-the-future/

“While counterfactual reasoning helps restore our feeling that the world makes sense, the problem with it is that it doesn’t help us get better at avoiding or dealing with future incidents. The reason it doesn’t help is that counterfactual reasoning gives us an excuse to avoid the messy problem of understanding how we missed those obvious-in-retrospect actions and signals in the first place.”

https://lorinhochstein.wordpress.com/2019/08/22/the-problem-with-counterfactuals/

The writings of Lorin Hochstein about resilience engineering are immensely valuable and illuminating. It’s nice to be able to read the thinking of people who are current with the cutting edge of a profession.

Sun, 22 Sep 2019 20:58:39 +0000

Not surprising that those things at the MIT Media Lab that looked too good to be true, were in fact fake.

https://gizmodo.com/mit-built-a-theranos-for-plants-1837968240

Mon, 02 Sep 2019 23:12:03 +0000

I’ve heard “if you’re so smart, then why aren’t you rich”more than a couple of times in my life. The answer it turns out is: unluckiness.

https://www.technologyreview.com/s/610395/if-youre-so-smart-why-arent-you-rich-turns-out-its-just-chance/

The Theory of Society

Wed, 05 Jun 2019 21:40:37 +0000

I was arguing against Systems Theory today (a necessary evil when you live in Germany) which prompted the thought of searching for the conjunction of Bruno Latour and Niklas Luhmann. That led me to this gem.

It turns out that they met each other for a debate in 1996 in Bielefeld and Latour DESTROYED his opponent (full article).

Luhmann, as expected, failed to engage with the theme of the conference, Science and Technology Studies and didn’t come out of his bubble. The same bubble that he has managed to trap most of the German humanities in.

[Luhmann] only managed to address the theme of the conference—science and its sociological study—with half a sentence where he curtly asserted that science is an autopoietic subsystem of modern society.

An autopoietic subsystem, my ass. Latour quickly riposted into a frontal assault at the entirety of Luhmann-ism.

No, according to Latour, this theory didn’t have anything to offer to him and neither he concluded any of those gathered there. A quick perusal of the conference program should have sufficed to ascertain that the empirically obsessed STSers could not recover their objects in this theory. This may be bemoaned from the high vantage point of the Theory of Society as poor, theoretically “flat” sociology, nonetheless, Latour replied, this empirical Zoology of STS gives an account of this society as it is and not how it may appear from the distance of the chilling heights of systems theory. Latour clarified that fundamentally, systems theory represents everything that he and his colleagues in science studies have been battling for 20 years—yes, really battle and not just criticize. The purification of science, the simplification of the social by a demarcation with its environment, Luhmann’s work as the embodiment of the “cognitive turn” in epistemology—for Latour these were the old buzzwords that had to miss what’s special about science: its materiality. And in doing so of course also what’s specific for modern society, the large technological networks.

Latour battling ‘the purification of science [and] the simplification of the social’, a true hero of science and society.

Mon, 27 May 2019 22:30:29 +0000

Fijn zo’n land onder de zeespiegel dat wordt beheerst door klimaatontkenners. En gezien het Nederlandse asielbeleid is het niet heel waarschijnlijk dat klimaatvluchtelingen elders op een warm welkom zullen kunnen rekenen.

https://twitter.com/Jamrat_/status/1132546604366475264

Sun, 12 May 2019 20:16:22 +0000

This article about forecasting dovetails nicely with the book about measuring risk that I just read. It makes the case for foxlike thinking, something of which I don’t need to be persuaded.

Among those, they identified a small group of the foxiest forecasters—bright people with extremely wide-ranging interests and unusually expansive reading habits, but no particular relevant background—and weighted team forecasts toward their predictions. They destroyed the competition.

Tetlock and Mellers found that not only were the best forecasters foxy as individuals, but they tended to have qualities that made them particularly effective collaborators. They were “curious about, well, really everything,”as one of the top forecasters told me. They crossed disciplines, and viewed their teammates as sources for learning, rather than peers to be convinced. When those foxes were later grouped into much smaller teams—12 members each—they became even more accurate. They outperformed—by a lot—a group of experienced intelligence analysts with access to classified data.

https://www.theatlantic.com/magazine/archive/2019/06/how-to-predict-the-future/588040/

Sun, 28 Apr 2019 13:59:00 +0000

As you’ll see, everything is slightly approximate, slightly exaggerated — so that each step in the argument takes you further from what you’re prepared to agree to.

It turns out Harari’s books only go downhill after Sapiens.

http://www.thewhitereview.org/reviews/yuval-noah-harari/

Tue, 26 Mar 2019 23:02:55 +0000

Theranos employees are suffering some adverse consequences for participating in that scam (and rightly so).

https://edition.cnn.com/2019/03/14/tech/theranos-employees/index.html