Definitely: and just giving the LLM context before correcting (in this case the title and description of the video, often written by a person) creates much better transcripts.
My repo doesn't re process the audio track: instead it makes the raw ASR text transcript better by feeding it additional info (title and description) and asking the LLM to fix errors.
It is not perfect, it'd sometimes replace words with a synonym, but it is much faster and cheaper.
The low cost of Gemini 1.5 Flash-8B costs $1 per 500 hours of transcript.
I was playing with Gemini this weekend to solve that issue and your video transcript converted into malt.
Here’s the result:
Clean YouTube Transcripts
appblit.com
Created by @ldenoue , this experiment uses the Google Gemini gemini-1.5-flash-8b model to clean up YouTube transcripts, which lack punctuation and paragraphs.
This whiskey news report highlights a variety of new releases and announcements from various distilleries. Notable releases include a 200th-anniversary Mallen time-space collection, a Glenmorangie ice cream-inspired whiskey available in the Netherlands and Belgium, a share purchase of a whiskey distillery group by Kari, a Big Peat 15th-anniversary edition, Kingsbar Distillery's first peated whiskey, a Johnny Walker bottle with environmental benefits and reduced weight, an Octo Series release, and a redesigned Killhumman 100% Islay. Other announcements include new wood finishes from Tomatin Distillery, a redesigned Shivers Regal bottle, a 35-year-old Glenallachie, an independent Highland Park bottling from Duncan Taylor, a new Redbreast 18-year-old, a Jimmy Russell bourbon, Angels Envy triple cask release, Jack Daniel's Single Barrel special release, and a 25th-anniversary SL Distillery release. The report also promotes upcoming live tastings on whiskey.com and whiskey.de.
Hello and welcome to the whiskey.com news on September 23, 2024.
We have several interesting news items.
First, Mallen is celebrating its 200th anniversary with the Time Space Collection. This collection, with two or three bottlings, includes "Space," a blend of two single malts. The outer ring is from a 1940 vintage (84 years old), and the core is from a 2018 distillation at the new distillery. Only 200 sets are available. Pricing inquiries should be directed to Mallen.
Next is the Time Space Mastery, a non-age-statement bottling with 14 different cask finishes.
We also have news I've already covered, concerning the Glenmorangie "Tale of Ice Cream," a sweet, creamy whiskey with honey and notes of lavender. It's available at whiskey.com for consumers in the Netherlands and Belgium.
A follow-up on the Distel Group news: The group, which includes Buachaille, Habanero, Deanston, and Toomore distilleries, has formed a subgroup. Kari recently purchased a 14.6% stake for $92 million USD. The group is now part of Cape Vin, a South African holding company.
Further news: Big Peat 15th Anniversary Edition, a predominantly red wine cask finish with 50% ABV, is available at whiskey.com in the Netherlands and Belgium.
Kingsbar Distillery is launching its first peated whiskey, the Kingsbar Coaltown. This new core range whiskey is matured 100% in ex-peated casks, resulting in a peated flavor profile. The non-age-statement bottling has 46% ABV.
Johnny Walker is introducing an environmentally friendly, lightweight whiskey bottle. Weighing only 180g, this bottle will drastically reduce the environmental impact of glass production. It holds a new variant of the Blue Label. Only 888 bottles will be available in 2025. The company plans to share the lightweight bottle technology in the future.
The Octo Series 15 (151, 152, 153) is also a highlight. These highly-peated whiskies are expensive and, according to the report, some editions may be quickly sold out. Check whiskey.com for availability.
Killman is also launching a 100% Islay 2024 bottling. This 9-year-old (at least), non-chill-filtered, uncolored bottling with 50% ABV is a redesign of the 100% Islay series and includes an age statement.
Tomatin Distillery is releasing three new wood-finished bottlings: Madera, Amarone, and Tempano cask finishes. These 10 bottlings are available at 40% ABV and are also available at whiskey.com for consumers in the Netherlands and Belgium.
Shivers Regal 18-year-old is receiving a new, slimmer, and taller bottle design with reduced glass weight and intensified coloring. The release date is scheduled for the end of September.
Glenallachie 35-year-old, Billy Walker's oldest Glenallachie whisky, is a blend of three PX sherry hogsheads, one oloroso puncheon, and two virgin oak barrels with 48% ABV. No coloring or chill filtration. Available at whiskey.com.
Independent bottler Duncan Taylor is releasing the Accolade Highland Park 1970. This 54-year-old Highland Park is available in 158 bottles. The cask has been in Duncan Taylor's possession since its inception. It boasts 42.9% ABV and sherry cask maturation.
From Ireland, Red Breast 18-year-old, an Irish single pot still whiskey, is now a permanent addition to the core range, expanding from the core range four years prior. It's available at 46% ABV and matured in Cream Sherry casks.
Turning to the United States, Wild Turkey is partnering with bourbon legend Jimmy Russell. A special 8-year-old bourbon, the Jimmy Russell, will be released internationally this fall. It has a 50.5% ABV.
Angels Envy is expanding its Signature Series with a new triple cask finish (Hungarian Oak, Spanish Campin Oak, and French Oak). The release is scheduled for October. This 92 proof (46% ABV) bottling will be available in the US and, hopefully, internationally.
Jack Daniels is releasing a Single Barrel special release called Koi Hill. Bottled at cask strength, the whisky is between 61 and 68.75% ABV depending on the warehouse level. It's approximately 11 years old.
Finally, SL Distillery in Bavaria, Germany, is celebrating its 25th anniversary with a Distillers' Choice Edition. Using five undisclosed casks, this bottling boasts 48.8% ABV with no coloring or chill filtration. Available at whiskey.de and whiskey.com.
Upcoming Live Tastings: Live tastings will be held on whiskey.com and whiskey.de. The first, featuring Patrick, will be on October 31st at 7:00 CEST (whiskey.com) and 8:00 CEST (whiskey.de). Live tasting kits are available for purchase on whiskey.com.
Working on restoring punctuation marks in raw YouTube transcripts.
Bert model trained as a token classifier works but it doesn’t scale well with more languages. See https://www.appblit.com/scribe for the current solution
Now looking into fine tuning a small LLM like Gemma 2b-it to this task.
Any advice on other LLMs that would work appreciated.
Same here: I work on speech for a big company and working from the office is terrible. I had to squat the restroom 3 times yesterday so I could work (talk to my phone!)
At home I am so much more productive and zero commute.
Because of the badging policy in place, I end up scheduling non productive days at the office (doing email, reading other docs, meetings which are always with remote folks anyway but at least they see a genuine meeting room or phone booth behind my pretty face, so I guess that counts? ;)
I suggest you dispel of this notion that the players in this game have any sense of rationality or logic about this topic. There are massive (many trillion) commercial real estate interests at play here that totally trump all productivity, health, and even the holiest of holy climate change benefits. I don’t think people quite understand the real order of priorities and issues related to remote working. The big corporations are under both pulling and pushing pressures from the government and interests that control it to “put butts in seats” in big commercial real estate. They don’t care about anything else and they are throwing around money and fear to whip everyone into shape.
Yes, there will be some pressure pushing back and dread by corporations stuck between the ol’ rock of competition and the government-favor hard place, but it is unlikely to win out at the corporate level unless some real independent competition rises that is putting on massive pressure by not having commercial real estate capital expenditures.
A small office in a city can easily cost $1M per year, sure there are tax benefits, but they’re not benefits if you are competing with someone that does not have those expenses at all or far fewer, even after paying for team meetups that also fund a family vacation.
I suppose I can imagine that there's C level pressure from their relationship with a mayor of, idk, Palo Alto for a headquarters, or a new build in Austin, but couldn't that also be offset by pressure from smaller governments in areas that are looking to do "digital nomad" revitalization?
Various aspects of the Taiwan government, including the national level, are engaging in big digital nomad pushes. Small counties like Taidong in particular have been actively exploring revitalization through incentive programs, exploratory "hacker house" type events, and engaging with NGOs in the field to advise on how they can get more digital nomads into their county.
Would you mind explaining what these commercial real estate interests are? How does a company or the government benefit from having "butts in seats", all else e.g. productivity, equal?
One is near retail and footfall, coffeeshops, restaurants etc. But it's the government that cares about that and I don't think there are many place where gov has meaningfully intervened in private company policy
I don't see how real estate companies can influence companies/tenants, they don't hold much power here since everyone is shrinking or cancelling their leases
Where I work, RTO is partially Finance driven, bean counters just don't like seeing seats they paid for go empty
I work in Finance and we recently decided to sell our building and just rent office space catered toward meeting rooms since hybrid has worked so well for us. I think it just depends on the company's culture and the pragmatism of its leadership.
Even if it's set up with the "right" incentives, internal cost allocation can cause bad unintended side effects at the firm level. I worked at at large bank during the financial crisis. Individual teams got charged an implied rent for the space they occupied to their P&L - sensible because it means you can get a better idea of how profitable they really are. However, they went further and the rent was also higher for "nicer" parts of the building (which the firm owned and did not sublet to anyone). So when lots of space became available as large numbers of people were let go, teams moved themselves to empty space which was "less desirable" so had a lower internal rent. The space they left was left unoccupied, so this didn't save the firm any money. Worse, an external contractor was used to move the stuff between desks so it actually cost the firm money while "saving" the team money.
> it actually cost the firm money while "saving" the team money.
Only in the short term. It left the desirable parts of the building empty. It might be possible to rent out this space. That's what happened when the company I worked for downsized the factory; they just partitioned the building and rented out the empty space.
> Where I work, RTO is partially Finance driven, bean counters just don't like seeing seats they paid for go empty
Same here. Leaders up the chain get a "use it or leave it" mail for office space and suddenly everyone is asked to keep those seats warm by coming in x days a week.
Hmm, that could explain what's happening, but this doesn't sound "Finance driven" to me. What those "bean counters" are doing is actually quite rational.
If leadership wants people to RTO instead of just giving up seats, then it's 100% on them.
If you buy 2 machines but only use them at 50% capacity, the only you might as well just buy one machine.
Problem is office space doesn't work like that when the entire team is in on 'office' days to collaborate, you need the all seats. If you right size down to 60% seats (3 days office, 2 home) and have people rotate, you lose the 3 in-office collaborative days because everyday it's likely 30% of people are dialling in from home. You save 40% rent but it's closer to full remote in terms of collaboration.
Because the real estates as well as all of the other business services that manage, maintain and cater to the office spaces and the commuters are owned by the same mega corporations that own a large stake in your business.
You buy land, buildings, out have a contract for these types of things. When there nobody at these locations it makes it appear that these were bad decisions rather than the fact that there was a black swan event that caused a paradigm shift. Some walnut will try to say that the writing was on the walls about covid or about the benefits of remote work but that's mostly contingent on post hoc analysis rather than is situ. And as they say: hindsight is 20/20...
Tldr: fear of looking bad because metrics are more important than actual results
Some responses approached the reason, but they are essentially the behemoth financial interests of the various pension funds of essentially all the state governments and corporations, high net worth family funds, institutional money like university endowments and REITs/IRA/401ks, even many foreign sovereign and pension funds like the Scandinavians, Dutch, Germans, etc.
As you may be surmising, this not only carries rather major domestic risks if pension and other domestic funds start crumbling, but it also has massive implications for foreign countries’ domestic financiers and social stability, but it also has geopolitical implications from it.
During the post housing fraud period, a rather understated change was implemented to encourage accounting to not mark real estate to market value, i.e., record what the market is willing to pay, but rather keep real estate on the books for whatever value one would like to keep it at by various methods and practices.
What that essentially affected was a cooking of the books to prevent on book from showing losses. It is essentially still going on, but especially in commercial real estate since the COVID happenings.
You now still have massive buildings essentially still totally empty, all still valued at full occupation valuation even though they are, e.g., only taking in barely enough to cover operating costs in a freeze state, i.e., minimal services.
This is where things like property taxes come in, as the properties are still assessed at fabricated values, and property taxes are used to fund the local governments, everyone with financial interests in commercial real estate (many, because it was considered very safe) are now crying for mom. It gets a bit off topic here, but I think you get the gist.
To keep property prices high across the board. If one part breaks, the whole front collapses. Meaning that the banks who own the government and own the population will loose their grip on power.
> There are massive (many trillion) commercial real estate interests at play here that t
That theory is bullshit though. Yes, there are companies that stand to lose if office buildings clear out. But they're not the same companies that make the RTO decisions. The companies making those decisions could actually gain if they ditched the office buildings... facility cost is some absurdly large line item on the ledger for most businesses.
Without a clear connection between the two, I have to chalk this up to irrationality. Companies are still run by humans, and humans are irrational more often than rational. Especially with something like this, where there's no clear precedent to steer by.
Totally. A lot of corps are ruled by management and sales people. Those often really enjoy talking and connecting, and it is a form of control for them. Of lot of these people think they can't do their job well if the quiet people (IT, devs) disappear into their homes. And they often genuinely think the workforce needs to have meetings and show up to be accountable. They don't really think about what IT people actually need, or they do sometimes but it won't be a decisive factor in the end.
I've worked in places where sales people were seated next to programmers, and the sales people were shouting through their phones continuously. The programmers complained endlessly about all the noise - without effect. First lockdown we had showed an increase of at least 300% productivity - hard and reliable numbers because all output was tracked voluntarily by the team (management never asked for this). Number of builds, commits, releases...everything was way up. It was quite shocking.
As soon as lockdowns were lifted managers began talking about being in the office fulltime, because it was so good to talk to each other and align your work. I remember working in a team that did 1 day a week at the office, that day we couldn't get anything done because everybody was just chit chatting all the time. Even if you wanted to - it was just impossible to focus.
Our security officer (CISO) remarked how the lockdown enabled him to think seriously about a security issue for the first time in almost two years. Isn't that tragic?
Companies are as rational as consumer behavior. You can't make this stuff up. Never attribute to malice that which is adequately explained by stupidity.
I'd be very interested to hear the perspective of your sales and management teams. Your view is obviously very biased and only tracks "productivity" from the programmers. "Number of builds, commits, releases" off-handedly does not sound like straight progress, typically fast direction changing (adapting to the business environment) from management and sales is what drives profit. Code is not an asset, it is a liability.
The loss of programmer productivity can easily be overshadowed by gains from other parts of the business. I know it's not always the case, and nobody wants to hear that their suffering is better for the company as a whole since it devalues your work, but I would be super curious to hear why the decision was made.
I'm not claiming what sales & management do is worthless, it's just lopsided (in this case). If you force programmers to sit in an open floor plan with people who are loud on the telephone, not once but every day, and not only expect them to perform but also ignore their complaints - it is incompetence and/or lack of empathy. In the end 80% of senior devs walked out and got another job btw, maybe that is telling you I'm not just making stuff up from my own tunnel vision.
Productivity is not progress, for sure. You can be very productive building the wrong thing. I've been there, wasting a year on some crap that was canned. But you need to be able to deliver and not actively frustrate your devs that want to get shit done. Otherwise the 'adapted to the market' is just a scam at best.
btw, not every dev is the same. Some actually do like being in the office, even putting on music and have lots of small talk. That's also fine, if it works for them. And I also see benefit of going to the office myself (once a week or so).
I'm pointing out the pattern of sales & management incompetently projecting their own needs and biases onto the whole company and treating devs as grunt workers, forcing them to comply with their rules and fulfill their needs regardless of how it impacts their ability to concentrate and deliver. Software development is not grunt work.
I see fulltime RTO orders as a reflection of this. Smarter management will understand what devs actually need and take that seriously. Usually this kind of management has a technical background. Office can be a part of that, sure, but wfh invariably is too, I'm convinced.
I've worked for dozens and dozens of managers in my career. Some will go out of their way to buy you the best laptop, even make you coffee and give up the best seat in the office just so that you can work in peace. Because they know that such investments in their staff will pay off. Then, others will loudly interrupt you, deny any small request out of petty resentment, and blame any and all problems on your laziness - that is if they actually show up to notice there is a problem.
The one thing I’ve seen are where companies have tax incentives tied to butts in seats. Usually like 0 property tax, with the government assumption that they’ll make it up in sales tax (lunch, gas, etc.) and taxes from employees that move to the town.
But honestly, I think a lot of companies are just doing this instead of layoffs or in addition to small “don’t raise eyebrows” layoffs. Raise the pain to get attrition.
Companies like Amazon literally own billions in office space. Making that space valuable again with RTO policies, especially if you create an RTO trend, a way of enabling them to sell at their previous values.
Do they actually "own" it? I work for a large company who spent millions building a new office and then immediately sold it to a different company so they could lease it instead of owning it (apparently this looks better on the books!)
Or, as we have seen hundreds of times, it's sold at market price to a "third party" company that is actually owned by one of the board members or executives, which then rent it back to the company for a slight premium.
We do see this occasionally. You'll have some private equity group buy a restaurant chain like Red Lobster. They make RL lease the real estate from them after having RL sell the real estate to them. Sometimes it's not restaurants, I think this is what murdered Toys'R'Us (correct me on that if I'm wrong).
But this isn't the norm, and it's not happening to well-managed businesses. It's something a vulture does after the company has been struggling for years. If that happened with a Microsoft or an Amazon, or any of the companies we work for. It's silly to suggest that is the cause of widespread RTO mandates.
Amazon used to lease several buildings from Paul Allen for years but ended up buying them outright, and then going on a construction binge beyond that.
If I have to choose between "the owning class colludes" and "the owning class is acting irrationally, and just coincidentally happen to act in accordance with the interests of other members of the owning class", Occam's razor points me toward the first.
The first is a complex solution that would require many elements to work. The second requires nothing at all; it couldn't be simpler. Why would Occam's Razor point you to the first?
The simplicity of "it's all just a coincidence" is about the same as "it's all just magic". And you would be right about the collusion explanation's need for some elements to work, were it not for the fact that we see those elements working together in broad daylight. They attend the same universities, join the same clubs, they even have a town in Switzerland that's become synonymous with the owning class meeting up (Davos).
The people we're talking to don't really understand those concepts. Or possibly they just reject them implicitly (which would be an even stranger explanation).
Yes, evil rich people hiding in the woodpile, snickering while silently watching through their Monopoly Guy monocles... that's the only sane explanation.
The "owning class" consists of a bunch of companies managed by a revolving door of middle management and an even faster revolving door of upper management. And they don't own a damned thing.
Not coincidentally, it is because of their bias, not some hidden conspiracy.
Occams razor requires you to drop assumptions that aren't required to explain a thing. A conspiracy involves a lot of extra stuff, furthermore it is often hard to find evidence for - usually because they just don't exist. There are so many problems with them, it should really be a matter of last resort.
Conspiracies do exist of course, but I feel you should only use time to explain the world if you have sound evidence, and still be open to falsification.
It is not bullshit, though it does not apply universally either. In many cases it may just be managers who need to justify having spent all that money on their office spaces.
That being said:
Deflating a real estate bubble is painful, just ask China. By artificially keeping up demand for office spaces through unnecessary RTO, banks and governments avoid having to go through that, at least for the time being.
And yes, with (partially government-guaranteed) mortgage-backed securities and all that, large-scale devaluation of commercial real estate is going to be a huge pain.
In NYC where the vacancy situation is particularly dire, politicians have been banging the RTO drum for a while.
> In many cases it may just be managers who need to justify having spent all that money on their office spaces.
I find this argument uncompelling as everyone can obviously see that things changed since February 2020, and sunk costs do not justify throwing more good money after bad. It could easily have been the correct decision to acquire more office space years back and it could just as easily now be the correct decision to divest that office space if more workers are remote, and it's not anyone's fault for not having seen into the future that there'd be a global pandemic.
Certainly nobody will blame managers for real estate decisions which were made prior to 2020. But a large number of companies revised expectations that workers would return to office as late as Q1 2023 and some even later.
> sunk costs do not justify throwing more good money after bad
Sure. In a rational world, the office space costs are sunk and everybody just moves on. But seeing that RTO just decreases competitiveness for hiring without hard benefits to show for, this indicates that the decisions are not rational. The sunk cost fallacy is easy to fall for, and that is even before considering how business leadership roles attract narcissists who have a hard time taking blame for anything.
‘Be careful what you wish for’ - if remote work is the norm, there is all the sudden massive downward competitive pressure on salaries, as now people can work from much lower cost of living areas. Or even lower cost of living countries.
Either officially, or unofficially.
What globalization did to manufacturing, remote work is going to do to knowledge work.
And in most cases, if the bosses give up on RTO, they’re going to be explicit about this. Because why not?
I don't think that is an issue. Yes, companies which insist on RTO have to pay more, but as extra compensation in order to make up for the difference in time and money spent commuting.
But I'd argue what could be outsourced profitably has been outsourced already. And countries which welcome digital nomads who "work from home" while visiting on tourist visas are getting fewer also. In fact the Thai Government launched DTV which can be seen as precursor to measure and later capture (read: tax) some of the benefits of that for themselves.
Remote work means ‘outsourcing’ for a FAANG includes Ohio (prev. Sunnyvale), not just India.
And if you think this isn’t having an effect on offered labor rates, available positions in certain countries, etc. then you’re not paying attention to the current market.
There are massive shifts to offshore work at most FAANGs actively going on right now. But it takes time.
Are there attempts to crack down/extract extra taxes? Yeah. But most are ineffective and/or don’t meaningfully change the economics. There is a lot of fat that can be trimmed/extracted from the super high peak of Silicon Valley comp (for one example).
A lot of folks in the US are just waiting to see what happens before they move, or are stuck with high mortgages.
The pension funds owning the org also own the office it rents. Could pay rent in shares or scrib..
You dont want to explain to a retiring generation of "scream it into relity"-babyboomers that there retirement plan is wonky, especially after investing fiscally conversative.
