Pratique du Shadowing: I Tested OpenAI's Dots vs. Meta's Muse. What You Need to Know. - Apprendre l'anglais à l'oral avec la vidéo
Choisissez un mode d'entraînement
Chaque mode ouvre une salle d'entraînement en plein écran.
286 phrases
Si les phrases sont trop courtes ou trop longues, cliquez sur Edit pour les ajuster.
1
Today, I'm going to be comparing OpenAI's Dots and Meta's Muse on things like their features, their functionality, and their feel.
2
Also talk about pricing and where I think these apps are headed so you can use your time wisely.
3
And of course, I don't want to waste any more of your time.
4
So let's hop right in.
5
Okay, so we have Meta's Muse, which quickly became the number one app in the Apple App Store.
6
And then of course, we have ChatGBT's Dots.
7
Now my immediate reaction when I got in here
8
and started playing around with Dots is
9
that it felt like a bit of a rushed project to try to catch up
10
because they saw how muse was taking over
11
and they realized we have a better model
12
and you know we already have a lot of people inside
13
of our chat gpt ecosystem let's give them essentially what muse is
14
but in our ecosystem
15
and it just feels super rushed you can see even during dev day yesterday
16
when they announced this you know new product it literally failed
17
it you know she was trying to call it here
18
and it just wasn't responding and it was taking forever
19
and she ended up saying you know oh i think dotty
20
or whatever the name is it's having a slow day
21
and they ended up like just kind of pushing past it
22
but they like muted it on the live stream so everyone
23
that was there live in person saw the live demo they both have the ability to name your little muse
24
or your dots
25
and you can customize it a little bit as you can
26
see same exact thing here inside of the chat gbt
27
or the codex desktop app where i'm able to access my dot right here now the first thing
28
that surprised me was i was expecting dots to be more
29
of a grok bot feeling app where we would have a bunch of different bots
30
that we could create and we can have them talk to each other.
31
But this is way more like Muse, where it's kind of one thread or one agent
32
that you're talking to that kind of helps with everything on a personal level.
33
So we will look at the UI and we'll compare the differences and we'll talk about which one I like more.
34
But first, let's just talk about some of the foundational stuff.
35
So background work.
36
These are both persistent cloud agents that can have schedules and you can have conversations with, and they both have their own cloud computer.
37
The model powering dots right now is GBD6 Astra, and the model powering Muse is Muse Spark 1 .3.
38
I'll talk a little bit more about the differences between those models later in the video.
39
As far as the existing work setup, you already have a direct connection in Dots to ChatGPT work
40
and codecs if you have the desktop app and if you just allow access.
41
As far as Muse, it has its own computer and tools and they both have a really good connector ecosystem.
42
So kind of like the super simple single sign -on, just connect with a button rather than having to do a bunch of API keys and things like that.
43
And because they both have their own computers, they still can run scripts and you still could connect to APIs that might not have a native connector quite yet.
44
They can both create documents, presentations, interactive dashboards.
45
They can create all of that and keep them in the ecosystem.
46
As you can see here, if I go to my library inside of Muse, I can see these different artifacts, whether that be a document or whether it be more of a live dashboard web artifact.
47
And then inside of my dots, you can see that there's a bunch of ways that we can see the outputs here, whether those be markdown files or dashboards and other things, of course, especially because with dots,
48
your memory will start with your ChatGBT memory, and then it will start to develop its own context and learn over time.
49
So right away, my dot was connected to all my ChatGBT memory, and it was saying, oh, you know, you run this business, and you use N8N every day, and you do this.
50
And I was like, okay, that's really old memory.
51
That's ChatGBT memory.
52
Let me plug you into my actual local files
53
and my codex memory so that you have more real -time data of what I'm actually doing and what are my priorities.
54
But what I really like about Muse is in here, for the average person, let's say you're coming in here and you click on your Muse, mine is named Charlie in this case,
55
there's a little identity tab over here where we can see the soul file and the memory file.
56
So the memory file, we can actually like edit things.
57
This is like facts and preferences and commitments like that.
58
And then if we go back over here and we go to the soul, this is like Muse's persona.
59
So, you know, if you have built a Hermes agent in the past
60
or OpenClaw and you've kind of worked with these memory and soul files, you will be familiar.
61
But even if you're not, it just makes it super easy and visual because you can just sort of see it from this interface.
62
Whereas with Dots, it's a little bit more abstract as far as how it's understanding memory
63
and you don't have as much to do with sort of like the profile.
64
I don't know why that looks super pixelated right now, but overall, I'd say like the Muse interface is just a better experience, especially if you're thinking of someone who's non -technical.
65
By the way, guys, I've got this completely free SOP for you about getting your first AI automation client.
66
It's gonna go over the exact steps
67
that has been proven for hundreds of our AIS Plus members to get their first paid gigs.
68
It goes over the one sentence service pitch that can get you started today, why your first client should cost you money, the five minute video that answers can this person actually deliver before you've actually received any money,
69
what to do when you have zero case studies.
70
There's so many good things in here that are gonna help you out.
71
Even if you already do have clients, I would recommend grabbing this because like I said, it's yours completely free.
72
So if you wanna grab this, there's a link for it down in the description.
73
Let's get back to the video.
74
Now here's another difference.
75
With communication, you can obviously do this stuff from your phone.
76
You can use the Muse app or the ChatGPT app to talk to your dot or to your Muse.
77
With the dots, you can instantly integrate them into Slack, which is pretty cool.
78
So this was literally like a three -click setup where i added this app into my slack
79
so now i could chat with my actual dot right here
80
and it would go ahead and respond if i open up the desktop app real quick you can see in here
81
that you can see that it's thinking.
82
So you see the top, it says thinking, there's a little light bulb, and then in my Slack, that's how it's gonna respond.
83
But then also, inside of a channel, I could have it talk to me in here by inviting it to a channel, so your team could actually message your dot, which I think is pretty cool.
84
And once again, you can respond in here, and it will show in the app that it has received this message, that it's thinking, and that it's trying to reply inside of that thread.
85
And as you can see now, it is thinking, and then it's going to shoot a message back into the thread right there.
86
It said, hey, I'm here.
87
So that is pretty cool.
88
But with Muse, what you can do is you have a bunch of different chats you can manage.
89
So you've got your main chat, and then you can set up different side chats for different topics.
90
And you also have a native WhatsApp connector, which is pretty cool.
91
So on the app in Muse, this is view only.
92
So what I would do is inside of my WhatsApp, whether I'm on my desktop or I'm on my phone, I can just say, hey, how are you today?
93
And this obviously is Charlie, my Muse responding in real time.
94
And then in the Muse app, I would be able to see this interaction, but not respond in here.
95
So it's a separate thread as far as my main chat, my WhatsApp channel, and then these side chats.
96
And then what else is very interesting is the ecosystem.
97
So because Dots is already integrated into Codex and ChatGPT, if you are already living heavily in Codex and ChatGPT,
98
then Dots probably makes more sense at least right away when you wanna figure out what this feels like.
99
But with Muse, it has some of these, you know, like the meta ecosystem.
100
So Instagram, Facebook, Facebook Marketplace, and it just feels a little bit more, you know, detached from like this coding stuff.
101
Anyways, that's just a very quick rundown we're going to dive into some more stuff here
102
but overall after playing with these each for about a day this is how i feel
103
if i could just sum it up dots fits your current work environment
104
if you're using codex htbt it can use local skills it
105
doesn't feel like you're kind of starting over explaining a lot of things like priorities
106
and things like that and it offers a direct route into codex
107
and slack whereas muse is much easier to evaluate for free
108
because you can get started on this for free it's inexpensive paid entry editable memory whatsapp access strong meta integration overall.
109
Like if I wanted to have my parents or people that I know that aren't technical at all, I would probably recommend Muse.
110
It just has a better experience.
111
And when you talk about this actual pricing on the Muse Free, you can get up to 100 million tokens a week and you can get started.
112
Like I haven't yet paid for Muse.
113
And then you can get on subscriptions and you can increase your tokens.
114
Whereas in ChatGPT with dots, you have to be on a paid subscription in order to actually get access to building your first dot.
115
Now there's something out about right now they're not actually attributing your dot usage to your weekly subscription
116
but i think that's just to sort of build some i guess adoption inside
117
and then they're definitely going to have
118
that ding against your subscription now also right now we talked about how muse is powered by spark
119
and dots is powered by astra
120
so let's talk a little bit about these models here's some rough benchmarks astra is a little bit more intelligent
121
but when we think about like professional work
122
and knowledge work spark doesn't do bad at all
123
so creating some of these documents and helping you with some of these routine tasks and figuring out how to unblock yourself, things like that.
124
Now, when it comes to building apps and software and automations, obviously, Ash is going to be a bunch better.
125
With coding, it's going to be a lot better.
126
But Muse is also going to be a little bit faster.
127
So I guess what I'm trying to say here is
128
if I was building a bunch of little automations and scripts
129
that I wanted to run and things like that and dashboards with one of these tools, I'd probably choose Dots.
130
But like I said, for general knowledge work, and if I'm just using this kind of as an assistant to help me keep on track, manage emails, remind me of things, manage the team, I would probably be pretty comfortable using Spark.
131
So let me just show you, I've said quite a few times that I think the interface of Muse and the experience of Muse is just better.
132
Let me show you why.
133
Inside of Muse, you obviously come in here and you create one thread.
134
So with GrokBot, it can be a little intimidating to people because there's the ability to keep creating more bots, but here you just create one bot.
135
Now, what you see right here is you see today.
136
You see things that have been worked on by date.
137
You also click on this next tab, which is approvals, if there's anything for you to approve.
138
So super nice, easy interface to approve things if you've got a bunch of automations running.
139
You have a schedule tab.
140
So you can see here are your reminders.
141
You can see here are things that are running.
142
You can even check in on the heartbeat that runs every 30 minutes.
143
And the heartbeat will proactively look at your Google Docs and your email, and it will give you alerts about things.
144
And it will even give you reminders like, hey, use me, set this up.
145
You know, like it just feels like it's more user -friendly there.
146
And then of course you have the identity of looking at things like your soul and memory.
147
Now also you'll notice on the left -hand side, we can search.
148
So we can search through a bunch of different chats to figure out, oh, I was talking about this with Muse, where'd that go?
149
You can also look at your different chats, like I talked about.
150
You can manage them here, which I really liked, because it's not as overwhelming as having to build a bunch of different bots.
151
It's more like, oh, you can have a couple side chats
152
if you have one very specific project that you want my help with.
153
Otherwise, use the main chat, and here's the channel that you've connected me to.
154
And this is plural.
155
So what I'm assuming is they're going to start to add more, probably Slack, probably Teams, Discord, things like that.
156
A lot of times I hear from people, okay, if I was to use AI, what would I even use it for?
157
Well, Muse does a good job of like giving you ideas.
158
You can set goals, which I think is really cool.
159
Optimize my sleep, health goals, relationship goals, career goals, interest goals.
160
You can come in here and look for ideas.
161
It'll tell you what it should be thinking about
162
and how you should be kicking off like these different automations and things like that.
163
building artifacts.
164
It's really, really cool.
165
And then if you go to the feed, you can see that you can actually customize what you want to see.
166
You can see things about sleep, because I've been talking about sleep.
167
You can look at OpenAI.
168
You can look at all these different things.
169
And if you like this, you can discuss it.
170
And you can just shoot off a task to your muse to say, hey, can we dig into this a little bit more?
171
Build me an artifact for this.
172
Let's chat about this, whatever.
173
You can also like things, which is going to change the algorithm to give you more things that are actually relevant to you in your feed.
174
I think this is a really cool experience.
175
And then if I go to my artifacts, you can see that I have, you know, all of these here documents or web We can also create images.
176
We can create videos and then you can create podcasts, which I think is really cool
177
So maybe you want to turn your feed daily tech news into a morning podcast a three -minute morning podcast
178
Or you want a weekly podcast on what your team accomplished this week
179
or how your workouts were this week This is very very cool
180
and something that I think like average people getting into this whole ai stuff They'll think that this stuff is much cooler.
181
Then if we go into dots, the interface just is not good.
182
It feels rushed.
183
It feels very simple.
184
The one thing I think that it has nice, which Muse doesn't have, is that you can call it.
185
So you can call your dot, which is actually the thing that I showed earlier, which failed during the actual Dev Day launch.
186
But you can call it.
187
You can talk to it.
188
It will respond in real time.
189
It's pretty good voice.
190
It's pretty, like the latency isn't bad at all.
191
They both have their own computers, as you can see.
192
And this saves logins.
193
So very similar to GrokBot where you can save logins.
194
You can look at the actual sort of like terminal and you can also create and save files in here as well.
195
This is how you set up that Slack connection.
196
And then you can also review recent activity.
197
It's not very easy to manage your scheduled tasks inside of Dots.
198
As you can see, here's one, right?
199
Plan tomorrow.
200
And this is a repeat daily task at 8 p .m.
201
Central.
202
We have some advanced settings as well.
203
But I was asking Dots, I was like, how do I actually manage this stuff?
204
And it was like, oh, click on my profile and go to scheduled.
205
There's no scheduled in here.
206
I guess the way you do it is you go to your scheduled tasks as
207
if you were looking at like your codex scheduled tasks or your ChatGPT work scheduled tasks, and you can see them in here.
208
But why would you not be able to manage your scheduled tasks inside of the dots interface?
209
I thought that was just a little bit bizarre.
210
And it also was telling me that it was really good at managing my codex threads, which was cool, but it really doesn't do it.
211
Here you can see I was like, oh, okay, so you're kicking off new threads and it's just appearing in my recents.
212
It's not actually putting it inside my Herc 2.
213
So I obviously like to manage my threads right in here in my Herc 2 project, but it would kick them off down here.
214
So right here, you can see here's a thread that it kicked off.
215
Actually, no, not that one.
216
This one, you can see that it was sent by Herc.
217
So my dot is basically kicking off these tasks, but it's not putting them inside of my Herc 2 project.
218
So I'm like losing the organization.
219
So that's leading me to this point, which is like Dots doesn't even know how it works.
220
I love when I can talk to Claude Code or Codex
221
or GrokBot about how it works so I can get a better understanding of how I should utilize it.
222
But Dots will just tell me things that's like, okay, that's not actually how you work.
223
So how do I, you know, I start to lose trust over what kind of knowledge it actually has.
224
But hey, at least this little guy is cute and customizable because ultimately people seem to care way more about that.
225
Okay, so ultimately where do we sit now?
226
Well, I think that I'm going to continue to play around with both of them.
227
But if I had to choose one, for me, I would be choosing Codex.
228
I would be choosing Dots simply because I live inside of Codex and Cloud Code.
229
I don't care as much about this super polished
230
and super nice interface because I already have ideas
231
and I already have goals set and I already have a lot of the stuff set up already.
232
But if I didn't, I would 100 % want this because Muse's interface is so much better.
233
Their advantage there is meaningful.
234
But what I think is interesting is that Dots
235
or OpenAI is probably getting so much feedback around Dots and they've obviously been looking at Muse and taking inspiration from it, which is why I'm so confused.
236
I was expecting Dots to come out and have a very similar style interface.
237
but it's just like lacking so much.
238
But what they're gonna do probably is they're gonna go to Muse and be like, okay, cool.
239
I like this, you know, ideas tab.
240
Let's put that inside of Dots.
241
I like the way that these artifacts are set up.
242
And, you know, OpenAI also did announce their spaces thing, which is gonna be cool.
243
But they're just gonna take features of what people really like about Muse
244
or what people really like about GrockBot and just put it in there as well.
245
So what I think is gonna happen is like Muse, GrockBot, Dots, they're probably gonna all sort of blend into one sort of tool.
246
meaning they're all going to look functionally very, very similar.
247
The differences will be some of the native connections and the capabilities of integrating with what you already have.
248
And obviously the underlying model that powers the thing and then pricing, you know, like the cheaper ones might be perceived as better to certain audiences.
249
And that was kind of my next point, which is I think Astra is just obviously a much better model than Muse Spark.
250
So that is one advantage that it has right now.
251
I also think that GrokBot right now is a different organizing model
252
because it has a bunch of crockbots that you use together rather than just kind of working with one.
253
thread.
254
And I think that there's pros and cons to each.
255
I think
256
that the average user who is not deep into the world of AI will probably like the interface of Muse
257
and Dots more because it just seems more simple.
258
Oh, I have one personal assistant that I can talk to
259
and delegate work to rather than trying to figure out how many bots I should make inside of CrockBot.
260
And then finally, just a little bit of disappointment here that Dots feels so rushed.
261
I mean, I was really excited when I heard that OpenAI was working on their own sort of always -on agents.
262
I thought it was going to be like crockbot i thought
263
that it was going to be really really crisp and slick
264
but it just feels very buggy i actually tweeted this last night
265
and i said dots should have been called bugs
266
because i was having so many issues i was going to make a video yesterday
267
but i was just getting so frustrated because nothing was working
268
and i was like i couldn't get it on my phone
269
and my desktop app wasn't working
270
and i was just a little frustrated it felt very buggy
271
and it still does but it does feel better today
272
but it's just interesting how limited dots feels at the moment especially
273
when you had a very clear example like Muse to pull from.
274
So anyways, like I talked about earlier, I think it's really about understanding what are the things that matter most to you?
275
Are you already kind of ingrained in the ChatGBT codex ecosystem, or are you more ingrained in like Instagram and Facebook?
276
And that kind of stuff might be more useful for you right now.
277
Would you rather be able to talk to this thing on WhatsApp, or would you rather be able to have this in Slack with your team?
278
I mean, I think the ability for my team to, in Slack, be able to talk to my, essentially my codex, I think that that's a pretty cool like functionality,
279
or for me to be able to assign my codex things from Slack and have it be sent right back into Slack.
280
I think that is super cool.
281
And I'd like to see that come to other apps as well, but that is going to do it for today.
282
So if you guys enjoyed, you learned something new, please give it a like.
283
It helps me out a ton.
284
And as always, I appreciate you guys making it to the end of the video.
285
I'll see you on the next one.
286
Thanks everyone.
Télécharger l'application
Tout pour parler couramment
Prononciation IAChaque phrase notée
Phonétique (IPA)Maîtrisez chaque son
VocabulaireEnrichissez votre lexique
Jeu de vocabulaireApprenez en jouant
Même chaîne
FOR YOU
Vidéo recommandée
B11
41 phrases
009 3032 MarshallHaas NAM Acq EN UAC 1080x1920 82s MUSE M10 2015
Daily ConversationWorking Adults
B22
188 phrases
Dot Plots: How to Actually See What Your Users Are Doing
Daily ConversationWorking Adults
B2694 phrases
Which AI is The Best Chef?
Daily ConversationWorking Adults
B273 phrases
ChatGPT Tutorial | Chat GPT Explained | What is Chat GPT ? | Edureka
Daily ConversationWorking Adults
B21
191 phrases
I Rebuilt My Entire Website in Astro… Was It Worth It?
Daily ConversationWorking Adults
B21
185 phrases
15 Minutes to Fix Your AI Dev Workflow with OpenCode
Daily ConversationWorking Adults
B283 phrases
Ray-Ban Meta Glasses Tips and Tricks You Need To Know!
Daily ConversationWorking Adults
B223 phrases
AI Made the Most Passive Side Hustle Even Easier
Daily ConversationWorking Adults
B2268 phrases
Disturbing Robot Thinks It's Human..
Daily ConversationWorking Adults
B2292 phrases
Apple Freeform - Apple’s Most Underrated App!?
Daily ConversationWorking Adults
B21
157 phrases
Claude Design just got MASSIVE Upgrades (2.0 Update)
Daily ConversationWorking Adults
B22
174 phrases
One Woman’s Search for Meaning Led Her to AI Delusion
Daily ConversationWorking Adults
Vocabulaire et conseils d’expression pour cette leçon
Cette leçon d’expression orale de niveau B2 s’appuie sur la vidéo « I Tested OpenAI's Dots vs. Meta's Muse. What You Need to Know ». Les mots qui reviennent le plus souvent : dot, muse, app, inside. Cette vidéo contient 286 phrases et 4078 mots à répéter en shadowing. La partie parlée dure 16:29. Le locuteur parle vite, environ 247 mots par minute : attendez-vous à des liaisons et à des sons réduits. 84 % des mots font partie des 3 000 mots les plus courants en anglais ; le reste mérite d’être vu avant de commencer.
Vocabulaire clé de cette vidéo
Les 15 mots les plus avancés de la vidéo, avec leur prononciation et leur sens :
| Mot | Prononciation | Sens |
|---|---|---|
| muse verbe | /mjuːz/ | songer |
| codex nom | /ˈkoʊdɛks/ | codex |
| slack adjectif | /slæk/ | lâche |
| interface nom | /ˈɪn(t)ɚˌfeɪs/ | interface |
| ecosystem nom | /ˈikoʊˌsɪstəm/ | écosystème |
| artifact nom | /ˈɑɹtɪfækt/ | artefact, artéfact |
| automation nom | /ˌɔ.təˈmeɪ.ʃən/ | automatisation |
| spark nom | /ˈspɑɹk/ | étincelle, flammèche |
| dashboard nom | /ˈdæʃˌboɹd/ | tableau de bord |
| bot nom | /bɑt/ | robot, bot |
| subscription nom | /səbˈskɹɪpʃən/ | abonnement |
| podcast nom | /ˈpɑdˌkæst/ | podcast, balado |
| connector nom | /kəˈnɛktɚ/ | connecteur |
| integrate verbe | /ˈɪn.tɪ.ɡɹeɪt/ | intégrer |
| buggy nom | /ˈbʌ.ɡi/ | calèche |
Les verbes à particule que vous entendrez
| Mot | Prononciation | Sens |
|---|---|---|
| figure out verbe | cerner | |
| set up verbe | /ˌsɛt ˈʌp/ | mettre en place, installer |
| end up verbe | se terminer, se finir | |
| come out verbe | /ˌkʌm ˈaʊt/ | sortir |
| get up to verbe | commettre | |
| go ahead verbe | /ˌɡoʊ əˈhɛd/ | vas-y, allez-y |
| open up verbe | ouvrir | |
| start over verbe | recommencer |
Des phrases à répéter
Des phrases courtes et complètes de la vidéo, à réutiliser dans la conversation de tous les jours :
- Let's get back to the video.
- It said, hey, I'm here.
- Let's chat about this, whatever.
- As you can see, here's one, right?
- I'll see you on the next one.
La grammaire de cette vidéo
Les structures que le locuteur utilise le plus, avec les mots exacts de la vidéo :
| Structure | Dans la vidéo |
|---|---|
| Phrases conditionnelles if + proposition, will/would + verbe — une condition et son résultat | if I didn't, I would |
| Voix passive be + participe passé — l’accent est mis sur ce qui arrive, pas sur qui le fait | are headed · was connected · is named |
| Propositions relatives who / which + proposition — une précision sur une personne ou une chose | Slack, which is · connector, which is · tab, which is |
| Present perfect have/has + participe passé — une action passée qui compte encore maintenant | have built · has been proven · you've actually received |
Prononciation à surveiller
Le locuteur utilise 75 contractions et formes réduites, comme I'm, gonna, they're. Prononcez-les sous leur forme courte, telles que vous les entendez.
- Les sons « sh » et « zh »: automation /ˌɔ.təˈmeɪ.ʃən/, dashboard /ˈdæʃˌboɹd/, subscription /səbˈskɹɪpʃən/, foundational /faʊnˈdeɪ.ʃə.nəl/, benchmark /ˈbɛn(t)ʃmɑːk/
- Mots longs — placez bien l’accent: ecosystem /ˈikoʊˌsɪstəm/, automation /ˌɔ.təˈmeɪ.ʃən/, foundational /faʊnˈdeɪ.ʃə.nəl/, intimidate /ɪnˈtɪmɪdeɪt/, proactively /pɹoʊˈæk.tɪv.li/
Les sons difficiles pour les francophones :
- /r/ anglais — langue recourbée, sans frotter la gorge: artifact /ˈɑɹtɪfækt/, spark /ˈspɑɹk/, dashboard /ˈdæʃˌboɹd/, subscription /səbˈskɹɪpʃən/, integrate /ˈɪn.tɪ.ɡɹeɪt/
Comment s’entraîner avec cette vidéo
- Écoutez la vidéo en entier une fois sans parler et notez les mots que vous ne connaissez pas.
- Commencez à la vitesse 0,75×, répétez phrase par phrase, puis revenez à la vitesse normale quand cela devient facile.
- Enregistrez-vous et comparez avec l’original, en faisant attention à des mots comme muse, codex, slack.
Qu'est-ce que la technique du Shadowing ?
Le Shadowing est une technique d'apprentissage des langues fondée sur la science, développée à l'origine pour la formation des interprètes professionnels. Le principe est simple mais puissant : vous écoutez de l'anglais natif et le répétez immédiatement à voix haute — comme une ombre suivant le locuteur avec un décalage de 1 à 2 secondes. Les recherches montrent une amélioration significative de la précision de la prononciation, de l'intonation, du rythme, des liaisons, de la compréhension orale et de la fluidité.