frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

New way of making espresso with ultrasound

https://www.unsw.edu.au/newsroom/news/2026/06/New-way-making-espresso
29•darktoto•1h ago

Comments

thomasoffinga•1h ago
Someone get James Hoffmann on the line right now.
intended•47m ago
The American tech video is still surprising me.
willis936•2m ago
I will wait for his blind testing before homebrewing this one.

Might be a new meta for iced lattes.

mzitelli•1h ago
75% energy save to make a room temperature expresso. I am ok spending a bit more energy to have it warm.
alansaber•58m ago
Think of how cool you'll look with an oscilloscope on top of your coffee machine, though.
elil17•55m ago
It is intended for manufacturing canned iced lattes and similar, as stated in the article.
moodyScarf•49m ago
> The saving could be especially significant for companies who make coffee-based ready-to-drink products at industrial scale

It's not for you

weird-eye-issue•48m ago
Could be great for an iced Americano
klausa•39m ago
I'm _very much_ a coffee nerd.

There's a semi-famous, super hipster cafe near me in Tokyo, that I sometimes go to.

Once, they had a special on the menu, when they give you a flat white, and a double shot of espresso on the side, with a thermometer hovering over the shot, with suggestions of tasting notes that you can get out from sips at different temperatures.

Now, that's generally very much a thing — things definitely taste differently based on the temperature (or maybe _they_ don't, but we _perceive_ them differently? distinction without a difference, I guess.).

The suggested temperature ranges were 51-40C; 40-30C, and 30-20C.

51-40 was great. 40-30 was getting weird, but still _interesting_, because you definitely got different notes.

But the 30-20 was terrible. That is absolutely too cool to enjoy a shot of espresso. I'm all for experimentation and doing weird things, but that was no longer riding the line of "not great but interesting" and went straight into "why would you ever do this" territory.

thenthenthen•32m ago
Cool! I noticed most of my Chinese friends drink cold or iced coffee. Undrinkable imho but yeah, there is def. a market.
alansaber•58m ago
Not even the baristas are safe. Technology really is coming for everyone.
abujazar•58m ago
The crema looks like terrible, more like Nespresso, and having the coffee warm is kinda important.

But perhaps this can be used in the instant coffee industry or something.

blcknight•57m ago
I like my Nespresso thank you very much.
abujazar•54m ago
And this is why you can't use "100 regular coffee drinkers" to judge the quality of the brewing method. Most people can't even taste (or care about) the difference between arabica and robusta beans or whether it's red or white wine.
jemmyw•50m ago
Wine ruins the taste of coffee regardless of colour
kzrdude•49m ago
I like good quality coffee.

But I also need my coffee: I'll drink whatever quality coffee is being offered, as long as it's the best I can get that morning.

tjcvirage•48m ago
Sure you can. You can absolutely use those 100 participants if your aim is to develop and market a process that can then be used to make a product for those same types of people. Samples generalise if your sample is representative, and in this case, for large commercial coffee extraction companies, third wave coffee aficionados are not in their target audience.
uberex•56m ago
Awesome idea. I would love to try it. If that can also make my espresso routine easier I am up for it.
fabian2k•54m ago
But people want to drink coffee/espresso hot, not room-temperature. So you have to heat the water afterward anyway. I'm not seeing that much potential for energy savings here, unless you're comparing setups with large boilers inefficiently used for small amounts of coffee.
tjcvirage•52m ago
They explicitly mention large scale producers for mixed drinks as a massive target audience. E.g. iced coffee manufacturing will be heavily impacted, they would normally need to heat the liquid, extract the coffee and then cool it back down.
fabian2k•43m ago
If I read it correctly, they did espresso and filter taste tests at room temperature (thought they don't state the exact temperature, and how they managed to brew filter coffee with the described setup). Overall the press release is somewhat misleading in what the goal is until the part you mention. If the focus is industrial production of cold, mixed coffee drinks I'd have expected more quantitative measurements instead of taste tests. Testing how well your coffee is extracted is pretty trivial with the right equipment.
hultner•48m ago
According to the article they see the main use for industrial scale production of coffee/espresso based drinks, in that regard it does make sense. For home use not so much even if there could be some niche market for cold espresso drinks at home, using less ice would allow for less dilution and faster than prepping and then refrigerate the coffee. I sometimes put concentrated ice coffee into my whey/oat shakes, but this is indeed very niche use even for me.
ChrisMarshallNY
elAhmo•49m ago
When you thought that coffee snobs couldn't get any worse...
defrost•44m ago
...you discover frackuccino with cat processed beans: https://www.youtube.com/watch?v=rqw09I-BTvE
thenthenthen•33m ago
Civet coffee is the best! The production method not so much sadly :( Maybe they can create an artificial cat stomach?
erikgahner•11m ago
What exactly is the innovation here compared to the press release from two years ago?[1]

[1]: https://www.unsw.edu.au/newsroom/news/2024/05/Ultrasonic_col...

swiftcoder•11m ago
Does this work with cold water? Because if so, my iced espresso drinks are crying out for it

Where's James Hoffmann when you need him?

jurf•7m ago
> Both espresso samples were served at 22 °C to ensure a fair like-for-like comparison […]

> It is noted that espresso is normally consumed hot and has transient sensory attributes that are temperature- and time-dependent. Hence, serving espresso at 22 °C will alter its sensory characteristics.

This is a weird test, coffee get’s so much worse when cold. So people can’t distinguish between two bad coffees.

abujazar•16m ago
The product is moving into a different territory, though. I doubt these guys even know how to pull a proper espresso shot as a baseline for the test. I'm sure you can grind espresso coffee beans to a powder and just shake it until it achieves a flavor similar to espresso, but that doesn't mean the resulting drink can really compare to something that was brewed at the right temperature and pressure.

Technically you can also buy a bottle of grape juice from the grocery store, let it sit on the kitchen sink with a yeast lock for a few weeks and call it wine, and technically it even is, but it's also going to taste quite shitty.

hultner•43m ago
I jumped on this at first as well. But to be honest if their target demographic is industrial production of coffee like beverages (like those Star Bucks soda can ”coffees”), well then it might not be so bad. I was thinking that a lot of flavour compounds of espresso breaks down quite rapidly while the drink cools, so the method of cooling all drinks to equal temperatures could be enough to skew the results regardless, but again for commercial coffee based soft drinks this is already the case. Headline is a bit misleading though.
hultner•47m ago
Considering the vibrations from those capsule machines it might not be so far off
elil17•56m ago
>But perhaps this can be used in the instant coffee industry or something.

As stated in the article, the whole point is for use in ready-to-drink coffee manufacturing.

uberex•54m ago
It is the GPT2 of espresso. Looks bad but idea is sound. There will be a GPT5.5 later on.
rstuart4133•51m ago
> having the coffee warm is kinda important.

Cold drip coffee is a thing, done well a very nice thing.

nomilk•47m ago
Could be useful for mass production of espresso-based drinks (like the ones sold at convenience stores), and possibly various foods like Tiramisu.

An average coffee shop's espresso machine might use $200/month of electricity, so even though the percent saving (75%) is high, it's off a base that's small relative to other costs; possibly too small to be enticing.

klausa•46m ago
For certain styles of coffee, crema in espresso is not necessarily desired — it typically has a high concentration of more bitter flavor compounds.

If you're drinking light, floral and acidic coffees, it's been relatively "trendy" recently to skim the crema off before drinking it.

I don't bother with that, but pulling two shots and removing the crema from one of them and trying them side by side is an interesting sensory experience — I'd encourage you to try, at least once!

abujazar•22m ago
In that case I'd say espresso is not the right brewing method : )
klausa•18m ago
Fully disagree, but I understand that perspective.
•
23m ago
I have friend with a Jura coffeemaker.

Makes nice coffee, but I don’t think it’s worth the cost (but he has a lot of money, so it’s not a big deal for him).

I envision some fairly high-end kit, coming from this.

ma2kx•28m ago
I usually forget about my coffee anyway until it's cold.
exitb•23m ago
To be fair, precedent has already been made with cold brew.

Alibaba unveils AI models for robots, amid shift from chatbots to agents

https://www.reuters.com/world/asia-pacific/alibaba-unveils-ai-models-robots-amid-shift-chatbots-a...
2•giuliomagnifico•1m ago•0 comments

Show HN: Brainfuck but with Turtle Graphics

https://czterycztery.pl/programy/turtlefreeze/
1•fourgreen•1m ago•0 comments

Show HN: A prompt generator to help AI agents implement my new email API

https://emailsdone.dev/#prompt-generator
1•mikeapple•3m ago•0 comments

A Claude Code plugin that makes coding agents measurably cheaper over time

https://github.com/vukkt/token-warden
1•vukkt•5m ago•0 comments

Show HN: Build a timeline of your experiences and generate tailored job pitches

https://www.pitchlikethis.com
1•pranavm27•5m ago•0 comments

Attention Backpropagation: Step by step derivation

https://liyuan24.github.io/writings/attention_backprop.html
1•deadf00d•8m ago•0 comments

Ratchet: An AI Delivery Loop That Can Only Move Forward

https://praveenvijayan.substack.com/p/ratchet-an-ai-delivery-loop-that
1•praveenvijayan•8m ago•0 comments

Skills Are Where Your Judgment Goes

https://geoffstearns.com/blog/skills-are-where-your-judgment-goes/
1•tensafefrogs•8m ago•0 comments

IPO boom mints thousands of new millionaires and Silicon Valley angst

https://www.washingtonpost.com/technology/2026/06/13/ipo-boom-mints-thousands-new-millionaires-si...
1•bookofjoe•11m ago•1 comments

Building a Deep Research Agent That Survives

https://steel.dev/blog/durable-researcher
1•nkko•12m ago•0 comments

Haskell for Elm developers: giving names to stuff (Part 8 – IO)

https://flaviocorpa.com/haskell-for-elm-developers-giving-names-to-stuff-part-8-io.html
1•cekrem•15m ago•0 comments

Show HN: Rediscover the Small Web with an AI-Ranked RSS Reader

https://bubblewire.net/
3•pchm•16m ago•0 comments

Fox Buys Roku, the Problem with Fox's Smart Strategy, Streaming That Works

https://stratechery.com/2026/fox-buys-roku-the-problem-with-foxs-smart-strategy-streaming-that-wo...
1•swolpers•16m ago•0 comments

AI Instruction Use 78.7% of repos configure 1 tool but devs use avg 2-4

https://tomevault.io/standards/state-reports/2026
2•olijboyd•17m ago•0 comments

I'm building a local-first TypeScript guard for runaway AI-agent costs

https://github.com/salimassili62-afk/ai-costguard
1•salim2006•19m ago•0 comments

Show HN: CoreMCP – MCP Server for On-Prem DBs

https://github.com/corebasehq/coremcp
2•y11t0•19m ago•1 comments

Show HN: O(1) Recursive Moving Linear Regression

1•Pierdimi•23m ago•1 comments

Show HN: I built a dev diary that writes itself

https://mydev.day/
1•lucasmartinic•28m ago•0 comments

The Robots Work at Night

https://www.givedirectly.org/the-robots-work-at-night
1•cash_transfers•29m ago•0 comments

MistralAI's Le Chaton Fat Tops Web Dev Benchmark

3•nsoonhui•30m ago•0 comments

SearchLeak: We Turned M365 Copilot into a One-Click Data Exfiltration Weapon

https://www.varonis.com/blog/searchleak
4•croes•30m ago•0 comments

From hieroglyphics to APL (1991) [pdf]

https://sigapl.org/Articles/Language%20as%20an%20intellectual%20tool-From%20hieroglyphics%20to%20...
1•tosh•31m ago•0 comments

Correlated randomness in Slay the Spire 2

https://tck.mn/blog/correlated-randomness-sts2/
2•rdmuser•37m ago•0 comments

Brick: SOTA LLM Routing

https://arxiv.org/abs/2606.13241
2•FrancescoMassa•39m ago•0 comments

A Chinese rocket breaks apart dangerously close to the Starlink constellation

https://arstechnica.com/space/2026/06/a-chinese-rocket-breaks-apart-dangerously-close-to-the-star...
1•rbanffy•39m ago•0 comments

Florida lawsuit accuses TikTok of violating state's child social media ban

https://www.theguardian.com/us-news/2026/jun/15/florida-sues-tiktok-teen-social-media-access-law
2•beardyw•39m ago•0 comments

Papers from the Lost Culture of Array Languages (2011)

https://prog21.dadgum.com/114.html
1•tosh•40m ago•0 comments

France's sovereignty push can't escape the Microsoft gravity well

https://www.theregister.com/on-prem/2026/06/16/frances-digital-sovereignty-push-is-struggling-to-...
2•Timofeibu•41m ago•1 comments

Notes from Tired Egyptian Whose Job Is Explaining That Humans Built the Pyramids

https://www.mcsweeneys.net/articles/notes-from-a-tired-egyptian-guy-whose-job-is-explaining-that-...
3•Geekette•47m ago•1 comments

Show HN: Hyperbox- $40/month Mac mini rentals

https://hyperbox.sh
1•Areibman•50m ago•0 comments