mattthebaker

Worked in Japan for a year. The female janitors would clean the men's bathroom at any time, usually while many men were using urinals. At a Mountain hut on Mt Fuji, the men's urinal (more like trough) was next to a coed sink with only a 4 ft high divider. Many public bathroom had toilets or urinals easily visible when the doors were open, which were further often propped open. Most interesting experience, by far, was a urinal on the 4th or 5th floor of a high end shopping mall in Yokohama which was 1 ft from a floor to ceiling glass window overlooking the waterfront and a train station bustling with people.

I was pretty entertained by the cultural differences and thought immediately how none of these things would ever happen in the US.

Recently picked up one of these as well.

Regarding the fans, highly recommend using an overclocking tool to undervolt the CPU and iGPU -- reduced temp under load for me by nearly 15C, runs stable with cpu at -130mV. Stops the fans from turning on as often or running as aggressively.

I too was unlucky in receiving a Toshiba as the vendor of my 1TB SSD. There is a reddit thread out there about switching from Intel RST to AHCI with OCZ drivers -- this improved my random write performance significantly.

The most annoying part of the system to me is the "coil whine" coming from the internal DC/DC converters. Lots of different pitch and erratic high frequency noises come from the laptop. My colleagues can confirm that my ears are not particularly sensitive to high frequency noise but this laptop is very audible. Just do not feel like it should have this kind of defect on such a high end machine.

Prefer counting as well. At the end of the day weight loss is bound by thermodynamics. All diets are just hacks to help subconsciously tip the balance. May as well cut directly to the root and just count.

Health is a different story and more subject to what is actually consumed.

Not all motorcycles are that loud, especially the smaller ones more suitable for getting around IN the city. In proper urban areas, scooters are more convenient anyway -- smaller, more maneuverable, automatic transmissions, and more storage.

So Walmart gets reduced to things you need right now, i.e. groceries and random low cost convenience items.

Population has been shifting from rural to urban for decades, Walmart's advantage is in a shrinking market.

Drone delivery is marketing nonsense anyway.

Really Flash is more worse than memory, caches, standard cells, PLLs, etc -- it requires significant process additions to fabricate the floating gate, as well as modifications to support high voltages, circuits for high voltage (charge pumps).

That's ironic, EE usually has more math requirements than CS.

In logic, smaller is faster: less capacitance, shorter channels, and thinner oxides means faster switching.

In memory, smaller means less space to store your bits. I remember a quote from somewhere that with TLC at 20nm, you could lose only 1 ELECTRON PER YEAR from the floating gate to meet the memory retention spec.

Also, NAND had smaller feature sizes than logic for some time, before they really hit the wall. They innovated a lot of advanced techniques like quad patterning for memory cells. It helps in memory that you are optimizing for one and only one cell.

Yes, the process trade offs are not so good if you want embedded flash. You need a process that can withstand the high voltages for flash erase, which means thick oxides and slow transistors.

At 320Mhz, you should have icache and paged code. A process that can clock this high should have no problem with RAM densities to offer more than 16kB. Better to mirror what the entire industry does: low clock speed and embedded flash, or better process with faster clocks, more RAM and external flash.

Quad SPI NOR is still SPI, which requires serial timings. You need to serialize address and data with each transaction. Parallel NOR has parallel address and data, and an order of magnitude improved throughput and latency.

There is really no point to 320Mhz with such slow code memory.

Regarding point 1, this is a terrible design decision. Any low power or high performance uC that requires code in external SPI loads it into and executes from RAM. At 32Mhz it starts to make sense to have an ICACHE, with EMBEDDED Flash on a parallel bus. At 320Mhz you want to cache your RAM. With 320Mhz and external SPI (not even fast parallel NOR), you have to be insane. A single instruction cache miss will cost you 100s of clock cycles for a SPI read.

I really hope they just forgot to mention 100kB+ of RAM on that landing page, and the 16kB data is just a DCACHE.

Yeah, these are correct thoughts for a serial entrepreneur. With two, assumingly successful, exits he now likely has a net worth somewhere between 2-100M. He could live off his investments/net worth. If his next exit is not a very big one, he will not really gain much. i.e. If he exits for 10M, which could be good for the early employees, he has only made a small fractional increase in net worth, for presumably a ton of work -- he wasted his time. So of course he will want to take big risks. This is probably a big part of why VCs prefer serial entrepreneurs, besides other obvious benefits like experience, they are prepared to shoot for the stars which is how VCs get paid.

The potential of being a competitor in an as of yet unproven market wouldn't be a good reason to acquire them. If that market becomes very very big AND Nvidia wins it, Intel could buy them in cash if they wanted, they certainly have it on hand.

Why they haven't acquired them in the past, Intel clearly doesn't think they are worth it and preferred to compete against them directly on graphics.

This is considering IoT from a semiconductor company's perspective; There will not be much revenue from the chips for this market. How much can they make per smart light bulb? You also have to note that Intel doesn't have semiconductor IP for nearly any radio protocol (BT, BTLE, Zigbee, they have wifi but not low end) necessary to build SoCs to win those markets.

IoT will be a big market, but the profits won't be going primarily to the semiconductor companies via chip sales. We won't likely see an "Intel Cloud Platform" or Intel selling IoT deployment services.

The FPGA tech is interesting, but mostly irrelevant to IoT. FPGAs can really benefit from Intels leading edge manufacturing, so that is a smart acquisition (even before considering IA integration possibilities).

Intel does not have much of a presence in networking, and there are a lot of competitors there who are more invested, and further, design their own silicon (Cisco, Juniper, Huawei, brocade maybe). Maybe if they are able to productize the silicon (as opposed to more exotic materials) optical transceivers they have spent a ton researching they would have something huge here.

It was, and to appease investors that they are doing something about new markets after missing mobile. No one in their right mind would choose Intel as the platform for a wearable. The microcontroller companies: ST, NXP, Freescale, etc have been deeply embedded there for decades.

If that market does play out and grow as much as predicted, Nvidia will be the company most likely to give Intel a run for their money.

Intel has never been a company focused on chip volume, they chose to exit the DRAM business 20+ years ago. They focus on high end high margin products, which is now primarily servers with the contraction of the PC market. This is how they justify greater than $10B yearly expenditures on fabs and R+D. They have maintained, and still do maintain a 2 generation lead on process technology.

Alternate architectures are not going to do much to dent the two generation lead in process.

IoT and Wearables mean nothing to Intel, there is not enough money there. Mostly a marketing effort to signal to their investors they aren't going to miss the next big opportunity like they did mobile. Of course that ignores the fact that Intel was never going to be much of a player in mobile. All the money in mobile is in the base stations, not in the handsets, and Qualcomm (3rd largest semi conductor company behind Intel and Samsung) completely dominates the IP in that market. What matters there is radio technology, not microprocessors.

"We" in the way used was intended to mean an American or Americans as opposed to the foreign entity. That it goes directly to the top is a separate issue, and suggestive of increased taxation, which Trump is generally against. Obviously a tariff is an option here, but it won't automatically bring jobs back.

Pollution and other externalities accounting is interesting, of course Trump has denied climate change so I would not expect much progress here.

Have to consider how trading is working with most of these countries as well. We're not selling entirely foreign products, especially when it comes to the ones offshoring American labor. The American companies go in to make use of the labor, and then "they", the American company is importing those products for sale on the American and global markets. "We" again meaning Americans in aggregate, are mostly reaping the benefits.

To have an effective trade strategy which builds in externalities you would have to have a widespread agreement between many countries. Otherwise, if say, an American company operating in China exporting to Africa had to pay those taxes, but a Polish company for example did not, we would be losing.

Ultimately, trying to bring those jobs back is trying to force a sub optimal solution. As it is, Americans are winning, in aggregate. It would be better to focus on redistributing our increased winnings.

Hmm, yeah. I'm struggling to understand the difference between the WWC and the poor outside of identity. Is it ideological in the sense that the WWC are people who want to work but can't find work, and they perceive the poor as people unwilling or unable to work.

Because it is more efficient to do so. Either those crops require more labor, different climate, or are lower yielding, etc. than what we grow locally. If it paid more to grow those crops, we would grow them.

At the end of the day, it is about subsidizing an inefficient labor market. For the tariff to bring low value jobs back, it will have to be significant. Offshoring will still be more resource efficient, so you are better off paying back the tariff fees as welfare than bringing the jobs back. It would be much better to use the funds to re-educate this part of the population to work higher value jobs.

It is also disregarding the externalities of enacting a tariff. So we put something in place to bring electronic manufacturing back from China. They still control most of the raw materials, and can put their own tariff in place on export of those materials, further reducing our competitiveness. International politics are not so simple.

Taxing for regulation differential would only make a small change. Labor and associated costs, health insurance, taxes, etc will still dominate.

The views fail to recognize that US is not the only world market. More than half of Apple's revenue for example, is oversees. Exports to elsewhere in the world will be even more difficult to bring back, and harm the worldwide competitiveness of our businesses. His plan to encourage repatriation of overseas revenue for taxation is smart though.

It also overlooks the fact that we have outsourced the labor to those countries, but the US still captures most of the value. i.e. For every iPhone sold, how much money goes to Apple and how much to their suppliers? This applies to most consumer industries. The investments and trade deals mostly allow US businesses to go in and extract value from other countries, and most of the profits still come back to the top. We really don't need to capture the last mile of value.

We also can not overlook the fact that the market is already voting against US production with their wallets. The market is free to pay a premium for US made products, but clearly prefers lower cost goods.

His statements are really just feel good for a group of workers that has not adapted to the new economy.

It is a generalized statement. Very few SoCs have internal switch mode controllers. It is going to either be an LDO, or take a fixed operating voltage. Requirements for such will be specified in current, which is more useful than power.

Module level specs would be irrelevant in engineering because there is little incentive to use something like this, the radio isn't even certified. On that note, get your popcorn ready for when they ship to German backers and customs proceeds to trash them.

Generally, don't touch anything with mediatek or allwinner if you want it to work at the end of the day. Their driver support is very bad. Funny personal story, I turned down a contract almost four years ago that would have involved getting 3D running on an Allwinner based board. Four years later and CHIP is demoing an "Alpha" of 3D on Allwinner.

Meanwhile even hardware accelerated 3D and video decoding works on the Broadcom SoCs in the Pi, and has for years.

Mediatek and Allwinner are cheap, but useless.

Current is much more useful in engineering, because it is more constant for this type of device than power. Power varies with supply voltage while current normally does not.

Other than doing temperature calculations, power is not a very useful metric. And for such calculations you would use a max power rather than average power metric.

If you were optimizing a system for "power" use, it is much easier to work in units of energy than power anyway.

Reddit is owned by a mass media conglomerate. A significant portion of everything on the site is native advertising. Self promotion would interfere with what they are really pushing and undermine their influence. The smaller communities exist as goodwill to make the high traffic portions seem more authentic.

Using sponsored headlines makes you look even worse than posting unfortunately. You have to be an insider to get the native placement I guess.

Just watch for the next big hollywood movie to come out and random disguised content x to magically bubble up and plaster the front page.

Amazing how ignorant people are to what is actually in an inkjet cartridge. It is not the cost of the ink. The cartridges have precision MEMS components to dispense the micro scale drops of ink. As with anything mechanical, they have limited cycling, can clog with impurities in the ink, etc.

Or they just don't want to be distributing gigabytes of porn... most of which is probably under copyright.

Making the data set available and whether you can tinker with or retrain it are very different things.

Intel only recently introduced this feature, and it surely required a lot of R&D. At the time they were most likely the only one, and may still be.

Field firmware updates for hardware are to greatly reduce the risk a product needs to be recalled -- this is a worst case scenario with great cost for the manufacturer. To some extent it also reduces the pre-ship software verification costs -- potential damage from a bug slipped is greatly reduced.

Security risks are harder to quantify for the bottom line.

- j contributed something, anything to the company

- j was never compensated

- j was never forced to legally release rights to anything when he left

- legally, if j contributed anything of significance, even ideas, cruise is about to learn a very expensive lesson