E414 | This Week in European Tech with Dan Bowyer, Mads Jensen, Lomax Ward and Andrew J. Scott

0:00Well, this is Upside. it is now uh the week ending that will be the 16th feb we have andrew from 7 we've got lomax from outsides we've got mads and myself from supersede

0:16tear down this wall it's more than just an ally this is a union of values values let's start acting acting acting acting this show is not investment advice and the hosts of this episode may be invested in the funds and companies featured gents what has been happening this week on the work side had a really interesting call with an lp big lp um in asia has 3 000 managers individual managers in that portfolio venture within venture alone and what was shocking to me was he only had three in Europe. Three. And my takeaway from that is, let's just remember, we are part of a massive global ecosystem, right?

1:02And much as we are very European-centric. But secondly, there's a long way to go. Hopefully that three will become 10, will become 20. There's opportunity. Where are the others? Are they global? Are they US-centric? China, India, US, basically. And what was the insight onto the European or lack of European VC? It's just a blind spot for them. It's just like they're based in Southeast Asia. They started in Southeast Asia, then they go to the US. India and China are close, fast growing. There's a lot of capital in those ecosystems, right? There's just a perception that Europe is a little bit behind.

1:40The returns are not as interesting, which we know is not true. You educated him appropriately. Well, I mean, of course. Andrew, what's cooking with you? Well, actually, it's a great point. That's a perfect segue. Thank you, Lomax, because I guess what we're partly talking about is that the majority of money, arguably in Europe over the last sort of 10 years in this growth of venture capital, has gone into companies that only iterate. And they've sort of failed yet to create companies of global change and win markets. So when you look at the dominant tech stack today, most of these U.S. companies, there are some exceptions, Spotify being a fantastic one.

2:28And the epoch-defining technologies of the next 10, 20 years, currently, I would argue, in the lead are U.S. companies. Again, possibly some Chinese. So think about AI, health, robotics, space tech. There are some exceptions. And I think partly the reason is that European investors often lack the courage and ambition to back those genuinely next generation technologies early. And so I, along with actually Lomax on the board as well, but with Arpe, a Finnish entrepreneur, are launching Dragon Guild, which is designed to cultivate individuals to drive transformative change. So our basic goal is to foster the world's best investors who will back the step change technologies and business models To drive that economic growth in UK and Europe.

3:17So sometimes I'm accused of being a grumpy old bugger But this is something I'm probably doing. We're all with you darling. Yes to you to build those that that trillion dollar company be it protecting the environment improving societal health or quality of life protected democracy and and creating outsized returns. So we're raising the angel round at the moment for that new entity. It's separate from Dragon Chasers, but we'll sort of leverage the amazing people within Dragon Chasers. And it's called Dragon Guild. So you'll hear a bit more about that going forward. Mads, what's the latest with you this week?

3:55What's been happening with you? It's been a great week. We've just agreed terms on an investment into a wonderful young company in the insured tech space. Insurance sometimes gets accused of being boring. But yes, this business is very sexiful, using all the good stuff, the sensors, the AI, all the stuff to make lots of cool change into an old industry. Very excited about it. More to be announced shortly, I hope, but for now, just for you to get this one across the line. Cool, cool. This week, I've mostly been bothering BVCA students who are getting into VC. So one of the trainers there asked me to go and do a talk on best practice.

4:39And I said, James, I'm possibly not the best guy to talk about best practice, but I can certainly give some insights into what's been happening and how we grew and how we started and all the good, the bad. And he said, yes, I want that. I want all the kind of the slightly more controversial, contentious stuff. I said, okay, then I'm possibly a man. So that was my week. Well, not my whole week, but that was good fun. so let's cross over the channel over to Paris the France AI Summit now there are loads of ways to to kind of poke this bear but the I think the main vein that I took from from all of the coverage of the summit was JD Vance's speech just to backtrack for a second so the France the AI Summit was all about it was a statement on inclusive and sustainable artificial intelligence for people and planet that was the that was the directive that was the document to sign and there were a hundred countries in attendance at the summit of which 61 signed and the non-signatories quite interestingly were the uk and the us so looking from jd vance's perspective so he he was very much an american first as no great surprises american first if you're not in you're out we're in charge so i guess the broader question for the room is is this important that we didn't sign did we not sign for the right reasons um how did you think about how did you feel about the key aspects kind of covered there's accessibility ethical ai this innovation piece future of work sustainability collaboration it all reads a little bit like a dei contract which i obviously is not what America's all about and possibly not why the UK signed.

6:25And how much teeth does this kind of directive really have? So, Andrew, what else do you know about it? What are your thoughts on the AI summit? First is it's Pope the Rooster, isn't it? French national animal on the bay. I think it was a smart move on the UK government's part. It's easy, diplomatic win with the US. I'm not sure how much this actually means to sign up to it. Things are moving so quickly anyway. Things could be very different within a year. So I don't think it matters either way, and I think we probably made the right decision. So aligning ourselves as independent and more kind of growth-centric, what kind of might take away?

7:12Yeah, I mean, there are so many unanswered questions, and the technology is moving so fast. I mean, I think there was the CEO of a Chinese company, i can't remember the name of it who presented a graph or saying that he expected ai consciousness by 2030 and ai awareness by 2027 i think that's aggressive um many greater experts probably including him than i so who knows maybe he's right but things are going to be so quickly i don't know how much a french directive yeah i mean yeah yeah but like a french directive what's that mean that What matters is delivering the technology, and that's going to outstrip.

7:51That technology is being developed and moving forward far faster than any legislator, I can't say the word, can write stuff into law. Sounds French to me, baby. Lomax, what do you think? I don't know whether we can swear, but I couldn't give two shits about this declaration. As in, first of all, it's not a directive. You used the word directive. It's not like any directive is law. This is just like... It's just a press release, right? It's a piece of paper. You know, I'm sort of envisaging, you know, they get back, you wave around, you know, you land in your home country, you wave it. It means literally nothing and governments will change and then ignore it or get put in a draw.

8:28I think the only significance here is that, you know, UK not signing it and sort of aligning itself more with the US like diplomatically. You know, there's probably tariffs in mind. there's probably you know um just looking looking to create a position that's slightly different from the eu and slightly away from the you know the eu um digital services act so i i think um oh the eu ai act i think i think that is that all directive should i say um i think it's a storm in a teacup really i think it's apart from this little kind of political sideshow on it i don't i don't really care about it i don't think it means a lot people will have we will not be talking about this in a year in months time yeah and also kind of over some of the other news from the eu about this kind of 50 billion euro investment boost so mads where where would you take this conversation yeah i'm even surprised we're talking about this press release now i think it was sort of you know headlined yesterday and forgotten yesterday and i think the bigger news the bigger context here is you know the u.s announced the 500 billion dollars for ai kind of the star Stargate initiative.

9:34And then Macron, he sort of matched it, right? He said, look, we got 109 billion euros in France, precisely matching the US bid on a per capita basis, which I thought is gutsy of him. I kind of like it. He's so far down in the popularity polls. I think he's possibly second or third off the bottom with an approval rating. So maybe he's trying to win some votes. I don't care. I don't care how popular he is. I want to see action from our leaders. Give him the money. Focus on priorities that matter. Listen, UK GDP growth was 0.1 % last quarter. Okay, I couldn't give a toss about a press release about some nice vague words.

10:15I want to see action about things that can actually make a difference to the economy. And with AI, we have the biggest opportunity in our generation, right? Let's focus on that. Not kind of all the softy-softy side of it, But how is it actually going to make meaningful difference to our economies in terms of growth, in terms of efficiency and productivity? I think that's what really matters. I thought it was really interesting to see the commission come out with a 50 billion euro kind of investment announcement, which goes beyond what the U.S. federal government has done. So more direct dollars, less deal making, perhaps with corporates, which is what you could miss a bit there.

10:51So in that sense, it's more Bidenomics maybe than Trumponomics. And the question is whether it will work, because we know that nothing happens without entrepreneurs and private capital. You can wave all the checkbooks around you want, but if there are no kind of privately driven projects, it's just all going to become white elephants. So the question here is, can the European Union, can Europe really convince founders and investors to go deep in terms of investment here? I think it was interesting to see France and see them leading the way. Now, in my book, the ideal approach would be something that combines the UK's very detailed plan for AI.

11:28Lomax, you often ask for the plan. I actually thought the UK did some great work developing a detailed plan. But what we then need is we need money and action and targets and goals. And I think the EU's fiscal injection of capital to prime the pump is great. And then I think the U.S. approach to deal-making and really strike deals with private business to help attract the private capital and get the entrepreneurs going is essential. So something that combines those three approaches in my book would be the ideal. And I'm sort of waiting for somebody to come with that. But can it come from Europe?

12:02That would be great. And maybe France can lead the way here. Well, let's stay in France. Let's go Le Chat from Mistral. I'm sure I've said that wrong. Très bon. Can't use the wrong tense. So Mads, this doesn't look like a massive game-changing device. The kind of the... What I've taken away is this is based on the fact that it's a faster way of getting results from an LM. Is it a bit too late? They've based a lot of their content from the Jurno AFP, so the partnership they have with the AFP. It's GDPR limited, perhaps, because it's in the EU. It lacks that brand value and reach. And many of the users are complaining that it's quite boring to use as an LLM.

12:52So, Mads, maybe you can kick off with how you see Le Chat as this new product from Miss Joanne in France? Dan's just alienated all our French users. I know. Let me just upset everyone. It was with love and relish, I know. So look, the truth is there were some recriminations after the DeepSea launch. I think many people asked, sort of said, look, we've been looking at the U.S. for, kind of with Ennui for years, saying they have all this money, all this capex, they have all these chips, they can build all these amazing models. No chance we can catch up. We are, we're doomed. And then lo and behold, the Chinese with cut down chips that are, you know, handicapped due to export controls, et cetera, et cetera.

13:35They're a pretty darn good model. and we're all looking at our friends at Mistral and saying, well, why not you? Why could you not have done that? So now it's the French AI week and Mistral is back with a vengeance. They had this huge announcement of Le Chat and I thought the marketing was beautiful and it was all done very well. And kind of the headline piece is this super cool demo where they say, look, our model is efficient and you can ask the model to write the snake game in Python and it will do it much faster than the Claude from Anthropik or the GPT model from OpenAI. And I must say, I thought the demo was beautiful.

14:17I did my own personal test and I was unable to reproduce that performance. So whether some of it was sort of made specifically for the demo, I don't know. I love to see that there is still some ambition in Europe to play in the Super League. I think our big chance here is an open source. I think the DeepSeek team really showed the way on open source and showed that there is still a way to create incredible performance with open source models. We have superb academic talent in Europe. In France, you've got incredible mathematicians in the UK, in Cambridge, in Oxford, Imperial. There's amazing talent.

14:58And the fact that we now have more focus on open source means more sharing of knowledge, means it's easier to develop models outside of the hyperscalers. And so I think to some extent, DeepSeek might be a bigger catalyst for European AI than Mistral. I think Mistral is still important. It's important because European founders need to know that there is a way to start, build and raise and scale here. And so if there's any takeaway, I'd say I'd love that they're in the game. I love that they're giving a good kick and hopefully they can make some hair here while the sun shines. The open source angle, the whole deep seek push into open source, I think Mads, as you were saying, made it feel to me like this has put Europe back in the game.

15:42But Andrew, does this matter? Should we be focusing on foundationals in Europe? What's your take? I asked ChatGPT what it thought of Le Chat, and it thought it was a French cartoon. So then I educated it what it was. And then it came back and said, well, speed isn't everything. And it said, if you need nuanced explanations and research or complex problem solving, that I should speak to it. So that's the American perspective. As in, you should speak to ChatGPT, not Le Chat. Is that what it was saying? Yeah, it didn't really care for it. very much. It's all a bit iterative, isn't it? It sort of reminds me of the...

16:26It reminds me of that Internet Explorer meme. Do you remember the browser lineup? And then you had, it was all, we're Internet Explorers. And then, sorry, we're search engines. And then Internet Explorer, I'm a search engine. It reminds me of that meme that kind of the European is the one on the far right. And who knows where I live. Mr. R's got, I don't know what they're doing in their r &d lab who knows um so i'm speaking from a position of ignorance but i do think that if you're going to announce something you know make sure it's impressive and not interesting i you know again like my point i suppose my point is none of this is bad this is all good stuff you know it's just like i wish you know just it would be nice just having thought about it it's like it would be nice if we had more going on here you know like more money more projects more people working on it i'm sure this is this is good i mean two weeks ago we thought mistrial was a failed project and was was going to be sold for you know for parts so this is great mads i wanted to imagine i have designed and created this beautiful sting intro to ai corner and i would love for you because we talk about ai a lot so i think we should do a segment of the kind of the broader AI stuff that's cooking in the world.

17:42And I know you think about this a lot. So what is, what's some of the broader picture? What's some of the other aspects that are going on in AI right now? Yeah, just a bit of news from the other side of the Atlantic, some of the frontier labs. You know, whether we think they are, they're important or not, I certainly think they are, and I follow it quite closely. And Sam Altman has been out this week saying that there's news around GPT 4.5. It's on the way, and we're expecting it within the next weeks to months, so kind of call it four to eight weeks. He's saying it's going to be the last non-chain-of-thought model, and everything will be reasoning, which is interesting.

18:19So huge paradigm shift coming at GPT-5, which will be the model after that, which he expects to be a true omni-model that can do it all. And I'm not surprised, because if anybody is using JantGPT recently, you'll know that there is this delightful model picker today, which allows you to pick 4.0, 0.1, 0.3 mini, 0.3 mini high, the pro version, the older versions. With tasks, with reasoning. I don't know what's working. Whoever came up with that product conventioning is, I think, just needs some basic brand architecture training because it's absolutely confusing. And so he's saying they're going to try and put a stop to all of that.

18:59They want a true Omni model that can do it all. In parallel, Elon Musk has been out disclosing that they are very close to releasing Grok 3. And now people will remember the XAI Colossus Data Center, kind of this 100 ,000 GPU cluster that was built at record time last year. And since then, the XAI team has been busy developing Grok 3. And it promises to be a really exceptional model. And I'd say personally, I can't wait to get my hands on it and test it out when it's released here in a few weeks' time. Now, in the context of DeepSeek, I'd say next level reasoning is more important than ever, because this was one of the areas where DeepSeek made some meaningful enhancements.

19:43And this means a lot more inference. And when there's more inference, there's more need for processing and far, far more capability. So, you know, from my perspective, look, we're only a month and a bit into 2025, but it sure promises to be a great year for innovation. I'm actually really, really excited by some of the things we're seeing. I want to very quickly discuss the very small 25 bps rate cut that the Bank of England did this week with a view to asking you lovely people the question, is there going to be any effect on investing, any effect on startups or any slightly, I guess, medium or longer term impact this rate cut?

20:27Now, from what I've read, we're hopefully expecting another couple of rate cuts, but we have this stagflation potential challenge. Unemployment looks like it's possibly going to go up. Inflation is going up. Growth predicted in the UK this year has been halved to 0.75%. Fixed rate mortgage holders are paying more, even though the base rate has been dropped. So there are definitely aspects here that I think are going to affect the startup and investment community. Mads, where would you take this one? I would say that the announcement that the UK economy grew by 0.1 % in Q4 tells us everything we need to know when we know that that's celebrated as a good achievement.

21:21It wasn't a recession, baby. It wasn't backwards. Right. We've got four entwined challenges. We've got inflation that's sticky. We've got a fiscal deficit that means no room to maneuver for the government. We've got low growth and we've got creaking public services. And it's quite dire when you think about that. You sort of talk about stagflation, which is not something anybody wants to return to. But so I'd say on the flip side of that, the question is, is there a way to address all of these things with one unified strategy? And I think that the starting point of that is the UK's new AI strategy, because with AI, we have the most powerful tool for automation and creating efficiency we have had in our lifetime.

22:03And it's right there in our fingertips, at our feet, ready to use. And the thing we've seen is the government now be ready to, I think Andrew made the point earlier, not just talk about and look at AI as something to give grants to, but actually look at the government as an early customer of these technologies, having a clear program to adopt them, to use them, to scale them. I think big credit to Matt Clifford for developing that as part of the AI strategy. The UK government is now focused on doing that. The thing I'm still missing is some real numbers around the strategy. So if we talk about using AI to create better public services and more efficiency, what do the numbers look like?

22:46Where are the efficiency gains going to come? Where are we going to save money? How is, we saw this week, earlier this week, Bloomberg reporting that the NHS, the productivity of the NHS has declined from 2019, has declined 20%. Okay, so on one hand, we have this incredibly powerful technology available. And at the same time, productivity is not only stagnating, it's going backwards. And it's bonkers. And it means we have these public services that clearly are creaking. We have, as we know, people working there that are busting their guts to try and deliver. But because we have allowed things to atrophy with inefficient technologies and inefficient processes and systems, we are just not delivering the type of support and service we need.

23:34So I think there is a unique opportunity here to implement next-gen technologies, to use AI, to drive efficiency. And what I would love the government to do is to put real numbers of that. Where are the efficiencies going to be? How are we going to save money and where and over what period of time? And then use that to do a number of things. First of all, to improve the level of service because we need that, but also to create fiscal room for tax cuts. We talked about the tax burden we have. It's too high. Things I'd advocate right away to reverse the national insurance for employers increase that we know is impacting jobs and to reduce VAT because that could help address inflation.

24:17Those two initiatives would cost money, but I think we can more than make up for that through investment in efficiency in public services. So that would be my plea to the powers that be, is to take this very seriously and not just look at it as some things we're going to experiment with, but actually put real numbers and real targets around it. Well, Rachel listens, so we can speak directly to Mrs. Reeves. So get a handle on cost savings, cut taxes, invest and grow. I fear for the time and the inertia baked in there. But Lomax, where would you take this? well no i i think you know we just just to say you know we talked about 0.1 percent um gdp growth in the uk and q4 but also interesting below that level right is is the drip down to gdp per capita right in the individual level that is actually going backwards it's gone backwards now two quarters in a row so you can talk about the gdp and the macro at the high level but if you talk about per capita basis it's actually going backwards right so it makes it even more incumbent on on on the things that mads was outlining we've talked about before for these initiatives to take hold and for people to start feeling this you know change at the individual level this is also i think why you know you talk to a little there's a general moroseness about people when you chat to them about the uk at the individual level because literally we are going backwards the data now shows that the rate cuts it did you know i mean you know we've now gone like three you know quarter um not point you know two five percent cuts or 25 basis point cuts um i think that's great what does that mean for us in tech i mean i think the trickle through takes takes time right you know um the the less the less money you earn from interest the more this triggers a kind of slow rotation potentially into equity into the equities rather than debt which um you know would include private markets which would include us but that that takes a long times trickle through so there's no real immediate trickle through on the kind of investment in investment lens um one thing i i totally agree with mads i think mads is being a bit optimistic on the impact of ai on on government and and the total government spend you know i you know you look at a lot of that government budget and you know maybe this is something we could talk about all the time it's like a lot of that you know i don't know 200 million 200 billion sorry is going on pensions and benefits and things like that right the government's spending over just over a billion a year like that it's like that's a policy decision that's not sorry mads a trillion what's a trillion sorry you know yeah government spend just over a trillion a year but they're spending a quarter of that on on benefits and pensions right so it's like we're not we're not we're not automate those payments you know we can maybe automate the way you deliver those payments but that's just it's gonna you know there's a there's some absolute things in in the budget that you can't just automate, you know?

27:11So I feel like Mads is right. This is great. You know, my question is the government, is it, you know, the civil servant, is it a great adopter of, you know, AI? Well, that would take time. And B is the impact might actually, sadly, be smaller than you would hope. One other point that I've just been reflecting on, and this is maybe me getting out of my wheelhouse, but is such in the context of inflationary, it's like the government has announced a lot of infrastructure spending and if you go through all of the infrastructure the government want to build you know the new houses the reservoirs the runways the data centers etc i was doing some math and i was getting chat gpt to help me out this is like 500 billion of spending i mean not now but like in the next few years right if all these projects want to like and actually that could be hugely inflationary in itself right if there's not enough slack in the economy that's a very interesting kind of thing to think about which is like actually i mean there's all it's all very great investing but we need to worry about the sort of second order consequences of that and get a handle on on on that maybe that's me jumping at shadows and and you know we should just crack all this infrastructure investment but um yeah you don't want to see inflation going up and then rates going up again to counter that's where we need to create efficiency to free up resources well there you go there you go back circular baby circular yeah get the robot you know get the robot we need to get the robots out you know for those projects i mean they're not going to be built with software call elon he's on it although i saw that i saw yesterday that apple are looking at building humanoid robots did you see that and you love it it's the biggest i listen i i i am a massive apple fan where's the waiting list yeah but i think as soon as you put an accountant in charge i knew this as soon as tim cook take us oh no we need somebody who's just got that kind of inspiration design thinking slightly more maverick taste rather than the accountant running the ship but i think they i think maybe that that is now on the turn because they realize uh that things need to change and they need to you know at least be looking like they're innovating but that's what you have with the government isn't it the chancellor really runs the ship and because if you don't have the money you can't spend it and she sets policies which encourage growth or discourage growth.

29:33And when we talk about interest rates going up and down, you can get really down on the wheat on this, and there are plenty of economists who get paid lots of money to spout very eloquently, far more than I can, on the topic. But ultimately, sentiment drives the market overall. The economy is so interwoven and complex. obviously the relieves that we can pull of automation but if there is poor sentiment and there is lack of confidence in a nation then the nation will be charged more for its debt people won't invest they won't risk their money they'll save their money it won't spend their money companies won't choose to base themselves in that country and that's exactly what you see currently in the UK and many European countries and conversely you see the opposite despite all the arguments that there are fundamental issues with the American economy of enormous unprecedented debt.

30:31But there is optimism and there is confidence and there is hope. And unfortunately, the UK government and many of the European governments don't understand how important sentiment and the psychology of that is. And so when you keep, you can say growth, growth, growth as many times as you want. But then if your policies don't match your rhetoric and actually worse, they contradict, nobody believes you. And that's why I think you're seeing interest rates going up. That's why you're seeing the cost of borrowing going up. That's why you're seeing companies not invest. It's actually a very simple thing to change.

31:10And that is exactly right. And that's why I think it's so important and would be so powerful to say, let's set some real objectives around efficiency. Lomax, to your point, we've had the NHS go backwards. Over six years, productivity has declined 20%. It's not beyond the weight of man that you could imagine using tech to get productivity just back to where it was five, six years ago. And literally, with that alone, you would free up tens of billions of pounds that either could be reinvested elsewhere or freed up for the kind of tax cuts we need. If you came out with a bold and ambitious budget to say, look, efficiency here, free up resource, use that to invest in some tax cuts over there.

31:53I think you could use that to trigger exactly the animal spirits, Andrew, you were talking about to get the economy back and going. And it's not, you've got a majority of 170, 80 people. I mean, what are they for if not to be bold and do some ambitious things? So I just, I'd love to see that. Now is the time to be bold, but again, call me a pessimist, but I don't think you're going find more than 10 billion of savings from this ai driving government more than 10 billion 10 billion what the hell is that is that your pledge on the table 10 billion but again we could argue that why are we looking at ai to save make the savings in government actually shouldn't we look at the fundamental cost base of government the number of people employed yeah what the way she's doing spending that's lower hanging i mean i have a lot that's why i have a lot of sympathy for the whole department of government efficiency you know elon's wheel and the axe i mean he it's a messy process for sure and he's making it he's making it far messier than possibly needs to be but the the goal and the outcome he's trying to achieve is the right one because ai and those sort of things should be deployed not to let's save 10 million in government expenditure it should be actually how can we accelerate the economy how can we do things better not just the the scarcity mindset of like oh we can only use this to save money now i think i i fundamentally disagree with you why can't it be both why i think i think it can be both but i just of course i don't think i don't think ai you can't ai on its own without changes in how the civil service operate it to lomax's point is not going to move the needle of course not because people can't do any jobs like look you're reducing the expenditure on the department or you're not because you're using AI.

33:33So what you're really saying is reduce headcount. And that's not what I'm hearing. I'm hearing, let's use AI to make things more efficient. But you're not actually talking about fundamentally cutting the budget. No, no, you're applying selective hearing, because that is exactly what I'm saying. I'm saying, let's get a proper budget. Not you, the government. Yeah. And that's what I'm saying. That's the one thing we're missing from all the plans, because I think they've talked a lot about the right things. But the one thing they haven't talked about is where is that efficiency going to materialize in real numbers, in real pounds saved?

34:07Because without that, it's all a bit meaningless. I want to talk about something very non-topical, but I thought it was quite interesting. So a report came out this week from Microsoft and Carnegie Mellon, which was called The Impact of Gen AI on Critical Thinking. And it really caught my eye. And I thought it was a really interesting research report. Now, what they've done is they've spoken to 319 knowledge workers, over 936 different tasks that they were doing within their businesses or whatever they were doing in their knowledge work using AI. And the inference was that we're losing critical thinking skills.