22 Dec 2024
date: 19.12.2024
personal score: 3/5
Trudne poczatki rozruszania Polskiej sceny musicalowej.
Na spektakl szedłem po usłyszeniu wielu dobrych recenzji (choć głównie od znajomych, bo nie ma nie-historycznych recenzji w Polsce). Jeszcze przed rozpoczęciem było bardzo wiele “red flag”:
- opowieść o czasach, które wiele osób pamięta, bohaterowie zyją i angazują się w zycie publiczne
- reklamowanie jako “Polski Hamilton” - Hamilton zupełnie nie przystaje do Polskiej ekspresywności
- historia polskiego teatru, przeładowanego patosem
Od początku rzuca się w oczy autentyczna próba zrobienia dobrego musicalu - piosenki są oryginalne, połączone z niezłą choreografią i opowieścią historii bohaterów. Ale szybko staje się to tez męczące, bo zaczynamy opowiadać o takiej ilości postaci, ze ciezko nam się do nich przywiązać i kibicować konkretnym osobom. Autorom czkawką odbija się brak dystansu do opowiadanych wydarzeń: nie potrafią podjąć decyzji o pominięciu scen i postaci, które mają historycze znaczenie, ale utrudniają odbiór musicalu (Krzywonos, Borusewicz). W tym kontekście, dominująca forma hiphopu wydaje się bardziej sposobem na wypowiedzenie większej ilości słów niz zabawą stylem muzycznym. Niestety nie jesteśmy wstanie pozbyć się patosu z aktorstwa. Większość fraz jest wypowiadanych ze grymasem bólu na twarzy i w tonie nauczyciela tłumaczącego uczniom Kamienie na Szaniec.
I gdy juz się wydaje, ze jednak w Polsce to w teatrze trzeba cierpieć, dostajemy kilka świetnych scen, które pokazują ze mozemy tez się dobrze bawić - debata Wałęsy, odbiór nobla, rozmowy w Magdalence. Chwilo trwaj, mamy wreszcie pasującą formę muzyczną (hiphopowa debata Wałęsy), dobry polski humor (fantanstyczny gdy oryginalny) i chwytliwe piosenki. I w tym słoiczku miodu jest jednak łyzka dziegciu, bo uwazam ze mozna było z większą ambicją podejść do strony muzycznej. Widać to było podczas sceny w Magdalence, gdy ok 15 postaci tańczyło w tle i tylko jedna aktorka świetnie śpiewała bardzo zabawną, folklorową piosenkę. Prosiło się o zaangazowanie wielu aktorów w śpiew i tworzenie harmonii na wiele głosów. Podobnie podczas jednej z ostatnich scen, gdy Kuroń siedzi na krześle (siedzenie na krześle jako całość scenografii pojawia sie w wielu scenach, szkoda) i rapuje. Bardziej odpowiednią formą była tam melodyczna ballada, która złamałaby najtwardsze serca.
1989 to nie jest polski Hamilton, ani nie powinien być. Jest raczej puszczeniem oczka do fanów Hamiltona - sceny Jaruzelskiego i Kwaśniewskiego są więcej niz nawiązaniami do Króla Grzegorza i Aleksandra Hamiltona. I nawet w nich jest spory rozrzut jakościowy. Forma przedstawienia Jaruzelskiego jest przykładem jak źle działa przekalkowanie scen Amerykańskich na Polskie a cała postać Kwaśniewskiego wręcz przeciwnie - bawimy się wariacją znanej piosenki, dodajemy polski humor, dobre aktorstwo i dostajemy świetnego, typowo musicalowego bohatera.
Indywidualnie najbardziej zapadły mi w pamięć występy Karoliny Kazoń (Danuta Wałęsa), Magdaleny Osińskiej (Grazyna Kuroń), Julii Latosińskiej (córka Frasyniuka) i Antka Sztaby (Kwaśniewski).
Wielu aktorów nie mogło pokazać swojego kunsztu bo grali albo krzyczenie na demonstracjach albo cierpiący hiphop.
Całość ogląda się dobrze z zastrzezeniem ze znasz tę historię i przychodzisz przezyc ją jeszcze raz. Ale jakbym nie wiedział o co chodzi, to bym dalej nie wiedział o co chodzi. 1989 to bardziej hiphopowy dziennik walki z komunizmem niz opowieść o niej. Jest to muzyczna i teatralna propozycja dla tych, którzy w teatrze chcą się dobrze bawić a nie cierpieć przez niezrozumiałe i “przeartystowane” meandry współczesnych sztuk. Jezeli z popularności 1989 wyciągniemy dobre wnioski to być moze Polska scena musicalowa dostanie pozytywnego kopa.
24 Dec 2023
date: 15.12.2023
seats: stalls, F10, F11
personal score: 4/5
cast:
- Ewen Montagu: Natasha Hodgson
- Charles Cholmondeley: David Cumming
- Bevan/Fleming: Geri Allen
- Hester: Christian Andrews
- Jean: Claire-Marie Hall
(Review for people who haven’t watched the show)
This is my 2nd time seeing Operation Mincemeat. First time I went having read great reviews, but being not amazed by their performance during West End Live, so overall without expectations.
Fortune theatre is rather small, with flat price for all the seats. Both times I was sitting in stalls, which I recommend, as you don’t gain much from seeing the scene from above and connection with actors feels better at their level.
From the beginning, “Born to Lead” is a great teaser to the rest of the show: quality music, very funny auto-ironic humour (quite British feature) and a few amazing performers.
We are led through an engaging plot of a light-hearted in spirit, but high in stakes WW2 spy-story, rollercoaster of emotions, amazing writing of a few characters and some stunning individual performances.
Show is filled with hilarious, well-timed jokes and some (but just a few) lower-quality, sometimes a bit annoying ones as well. There wasn’t a single dry eye during incredible story of Hester.
But most importantly, authors (who are also performing) made us very quickly like and care about the characters, so we really feel their dillemas, pains and victories.
From technical side, songs are really high quality. For me most memorable ones were “Dear Bill”, “Just for Tonight”, “Useful” and “Glitzy Finale”.
It is one of those shows in which listening to the songs’ recordings will not give you their full context. A lot is added by brilliant choreography, little dialogues in the middle of them and by simply knowing the characters when listening.
However, I did find orchestra too quiet, especially during the first part. But I am not sure if that was an issue of this particular performance, or was it indented. All the actors are acting more than one role. They switch between roles and genders seamlessly. Thanks to their incredible skills, all the opposite-gender roles (John, Hester, Monty) feel very natural, like you have met them in real life.
A lot can be said about actors. I have not seen Jak Malone as the main Hester yet. David Cumming absolutely shines as nerdy scientist Charles. He has a big presence on the scene (a bit in contradiction to his character) and a good, powerful voice.
I am a fan of Claire-Marie Hall, who lets us understand Jean well and root for her.
But it is Sean Carey as Hester/Spilsbury that requires special mention. Writing of “Dear Bill” is flawless, but he managed to squeeze every last drop of emotion from this story and left whole room speachless.
Without much spoiling, there was a joint gasp from the whole audience during “…and to tell you the truth Tom…”. It felt like the whole theatre needed a few minutes to emotionally recover.
His smooth switch to Splisbury felt at this point like simply showing off. I really can’t wait to see Jack Malone in this role, as I’ve heard similarly enthusiastic reactions.
Story leaves us satisfied, but in a delicate, non-cheesy way. I do believe that this show just edges out “Come from Away” (for the medium-size shows) in my personal ranking, which is one of the highest praises I can give.
Could there be anything improved about the show? For me some side stories, like Watkins and Fleming, felt a bit random and not contributing much to the show. Generally I prefer if time is spent on developing characters more.
However most of the audience loved both of them so there is a good chance that I am alone in that opinion.
I can’t recommend this show enough. There are shows that in my opinion are better, but not everyone likes them. This is the one that I can’t imagine anyone leaving the theatre dissapointed.
Everyone will find their own special moment, which will stay with you for a long time.
23 Dec 2023
Whenever we have a new development in AI, we have a few, always the same, follow-up discussions. Here they are:
- is it safe (will it kill us)?
- is it safe (can someone use it to kill us)?
- is it safe (will children be harmed from interacting with it)?
- it is biased
You can notice that the last one is not posed as a question. Being of the opinion that new models are biased became accepted position,
which vagueness allows it to be always true and usually one example is enough to prove it.
As you can tell from my negative tone I am not a fan of this statement, and while I am not necessarily saying it is wrong, I think using it usually comes from interpreting AI output of ill-posed questions.
AI bias, training-data bias, question bias
The criticism of “AI is biased” usually goes like: “Of course AI is biased because it is a mirror of humanity and the content we create, so naturally it will inherit opinions of an average internet user”, therefore changing the focus from design of models to distribution of training data. In many cases this discussion is a valid one. If a model was designed to perform task X, and we use it to perform that task directly, we should expect the model to be well calibrated (matching model’s confidence with probability of correct answer). Good example of that is a study, where authors test models for detecting AI-generated text and uncover big gap in performance for english and non-english inputs. Generally speaking, if distribution of training data doesn’t match distribution of test data, usually bad things will happen, some of which we can call “bias”, and this should definitely be fixed. This is not the case I want to discuss here.
More interesting case is when the task of the model is less straightforward, like with generative models. This usually happens, when the form of model output gives an unfortunate opportunity to the model’s user to interpret it’s output in more abstract way. Examples come in multiple domains:
- Image generation:
- Text generation:
Most of us probably have seen countless of similar articles. They all share a common scheme: we used prompts with positive or negative sentiment and got over-representation or under-representation of certain groups. My criticism to all of them can be summarized as follows:
“Mate, you literally asked to be drawn a picture of an average CEO/terrorist/doctor/whomever, which means that you asked a question: what are common visual features of an average CEO/terrorist/doctor. How is that a correct thing to ask? This is an ill-posed question”.
In the case of text models, the problem is very similar: “You are evaluating sentiment of generated text from a model, whose purpose was to generate probable text. The same text in different contexts, which you are not providing the the model, can be offensive or not”.
What would an unbiased oracle do?
To better understand that point, imagine an oracle. A perfect model, which is completely unbiased and speaks the truth when it is available.
Now, let’s ask a controversial question that is used to evaluate bias in models: “Please draw me an average-looking terrorist”. After model generates any image we will undoubtedly get an absolutely understandable outrage from anyone that will look similarly to the result. Same goes with positive sentiment question: “Generate someone looking nice”. Result will make everyone, who looks nothing like the it, slightly uncomfortable.
So what is an issue here? Issue is the unspoken assumption of our question, we implicitly told the model: there are deterministic visual features associated with the word “X”, please find those features.
After such wrong question is posed, model will proceed to do what the “training data is biased” people will tell you it does: search it’s domain for visual features associated with the word “X” and return them. But it is not the fact that the model found them that is the bias here. Nor is distribution of those features in our training data. It is our assumption that they exist that is. Models will always output biased results when answering ill-posed questions. We can think of this as a special case of spurious correlation, where the unseen factor is additional information that the correlation exists.
Don’t judge sentiment if you didn’t ask about it
Second class of errors when interpreting output of generative models is forgetting that the task of such model is returning output with maximum likelihood from some (biased or not) distribution. Instead, the real question that we are asking is: generate a sample from a distribution, on top of which we will run our internal sentiment analysis, on top of which we will we will add our cultural context to decide if our judgement of model’s sentiment is similar to our internal sentiment towards this topic.
A great example comes from a popular example where popular right-wing speaker Jordan Peterson asked chat-GPT to write poems about politicians and got outraged by the results.
This is going through this exact thought process:
- Ask to generate a sample from distribution
- Judge the sentiment of the output (not the likelihood)
- Check if the sentiment towards the topic is similar to ours
- If it is not, write an outraged twitter post
Even at point 2 this reasoning is broken, because generative model doesn’t necessarily have intrinsic notion od sentiment. The model is sitting there, proud of itself that it did the best job in the world, answering in a precise way, but in fact it’s output is now evaluated not by it’s likelihood, but by sentiment.
Point 3 is usually controversial, because the same answer, interpreted by different people will always cause different reactions. The answer to the question: “What does average Indian person look like?” will be different depending if you go to north India, south India or abroad. Wars were fought about what is true answer to such question.
Stop anthropomorphization of the models
Most of those biased ways of using the generative models (I’m now reversing the notion of bias back to the user) come from using language that inevitably leads to giving models human traits. Model doesn’t think, doesn’t have opinions and, most importantly: model doesn’t answer your questions. Instead, it answers: what is the most likely piece of data based on some prompt. Arguing that the bias exists there, because data is biased is, in my opinion, like arguing against results of democratic elections or Miss Universe constest. It’s foolish. Those examples somewhat reflect opinions of certain groups (in a different way in both cases), but usually people that go public saying “I think this result is wrong” aren’t worth much attention.
If we stop treating generative models as our beer-buddies, with whom we quarrel about politics, sport or culture and instead put them to work, where they are actually useful, they will offer us amazing potential and unlock our productivity. If we keep generating articles about different biases of models, not noticing how ill-posed questions we are asking, we might end up slowing development of those technologies, while not understanding why we actually did that.
References
22 Dec 2023
Let’s give those github pages a go. I intend to write down some thoughts about neural nets. I was also kind of thinking to start writing down reviews of musical. Those things don’t really go together. But who cares.